2022-04-24 16:59:22 +00:00
|
|
|
/*-*- mode:c;indent-tabs-mode:nil;c-basic-offset:2;tab-width:8;coding:utf-8 -*-│
|
|
|
|
│ vi: set et ft=c ts=2 sts=2 sw=2 fenc=utf-8 :vi │
|
|
|
|
╚──────────────────────────────────────────────────────────────────────────────╝
|
|
|
|
│ │
|
|
|
|
│ Lua │
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
│ Copyright © 2004-2023 Lua.org, PUC-Rio. │
|
2022-04-24 16:59:22 +00:00
|
|
|
│ │
|
|
|
|
│ Permission is hereby granted, free of charge, to any person obtaining │
|
|
|
|
│ a copy of this software and associated documentation files (the │
|
|
|
|
│ "Software"), to deal in the Software without restriction, including │
|
|
|
|
│ without limitation the rights to use, copy, modify, merge, publish, │
|
|
|
|
│ distribute, sublicense, and/or sell copies of the Software, and to │
|
|
|
|
│ permit persons to whom the Software is furnished to do so, subject to │
|
|
|
|
│ the following conditions: │
|
|
|
|
│ │
|
|
|
|
│ The above copyright notice and this permission notice shall be │
|
|
|
|
│ included in all copies or substantial portions of the Software. │
|
|
|
|
│ │
|
|
|
|
│ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, │
|
|
|
|
│ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF │
|
|
|
|
│ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. │
|
|
|
|
│ IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY │
|
|
|
|
│ CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, │
|
|
|
|
│ TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE │
|
|
|
|
│ SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. │
|
|
|
|
│ │
|
|
|
|
╚─────────────────────────────────────────────────────────────────────────────*/
|
2021-03-02 13:51:10 +00:00
|
|
|
#define lvm_c
|
|
|
|
#define LUA_CORE
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
|
2022-08-11 07:15:29 +00:00
|
|
|
#include "libc/str/str.h"
|
2021-03-07 21:26:57 +00:00
|
|
|
#include "third_party/lua/ldebug.h"
|
|
|
|
#include "third_party/lua/ldo.h"
|
|
|
|
#include "third_party/lua/lfunc.h"
|
|
|
|
#include "third_party/lua/lgc.h"
|
|
|
|
#include "third_party/lua/lobject.h"
|
|
|
|
#include "third_party/lua/lopcodes.h"
|
|
|
|
#include "third_party/lua/lprefix.h"
|
|
|
|
#include "third_party/lua/lstate.h"
|
|
|
|
#include "third_party/lua/lstring.h"
|
|
|
|
#include "third_party/lua/ltable.h"
|
|
|
|
#include "third_party/lua/ltm.h"
|
|
|
|
#include "third_party/lua/lua.h"
|
|
|
|
#include "third_party/lua/lvm.h"
|
Release Cosmopolitan v3.3
This change upgrades to GCC 12.3 and GNU binutils 2.42. The GNU linker
appears to have changed things so that only a single de-duplicated str
table is present in the binary, and it gets placed wherever the linker
wants, regardless of what the linker script says. To cope with that we
need to stop using .ident to embed licenses. As such, this change does
significant work to revamp how third party licenses are defined in the
codebase, using `.section .notice,"aR",@progbits`.
This new GCC 12.3 toolchain has support for GNU indirect functions. It
lets us support __target_clones__ for the first time. This is used for
optimizing the performance of libc string functions such as strlen and
friends so far on x86, by ensuring AVX systems favor a second codepath
that uses VEX encoding. It shaves some latency off certain operations.
It's a useful feature to have for scientific computing for the reasons
explained by the test/libcxx/openmp_test.cc example which compiles for
fifteen different microarchitectures. Thanks to the upgrades, it's now
also possible to use newer instruction sets, such as AVX512FP16, VNNI.
Cosmo now uses the %gs register on x86 by default for TLS. Doing it is
helpful for any program that links `cosmo_dlopen()`. Such programs had
to recompile their binaries at startup to change the TLS instructions.
That's not great, since it means every page in the executable needs to
be faulted. The work of rewriting TLS-related x86 opcodes, is moved to
fixupobj.com instead. This is great news for MacOS x86 users, since we
previously needed to morph the binary every time for that platform but
now that's no longer necessary. The only platforms where we need fixup
of TLS x86 opcodes at runtime are now Windows, OpenBSD, and NetBSD. On
Windows we morph TLS to point deeper into the TIB, based on a TlsAlloc
assignment, and on OpenBSD/NetBSD we morph %gs back into %fs since the
kernels do not allow us to specify a value for the %gs register.
OpenBSD users are now required to use APE Loader to run Cosmo binaries
and assimilation is no longer possible. OpenBSD kernel needs to change
to allow programs to specify a value for the %gs register, or it needs
to stop marking executable pages loaded by the kernel as mimmutable().
This release fixes __constructor__, .ctor, .init_array, and lastly the
.preinit_array so they behave the exact same way as glibc.
We no longer use hex constants to define math.h symbols like M_PI.
2024-02-20 19:12:09 +00:00
|
|
|
__static_yoink("lua_notice");
|
2021-03-07 21:26:57 +00:00
|
|
|
|
2021-03-02 13:51:10 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
** By default, use jump tables in the main interpreter loop on gcc
|
|
|
|
** and compatible compilers.
|
|
|
|
*/
|
|
|
|
#if !defined(LUA_USE_JUMPTABLE)
|
|
|
|
#if defined(__GNUC__)
|
|
|
|
#define LUA_USE_JUMPTABLE 1
|
|
|
|
#else
|
|
|
|
#define LUA_USE_JUMPTABLE 0
|
|
|
|
#endif
|
|
|
|
#endif
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/* limit for table tag-method chains (to avoid infinite loops) */
|
|
|
|
#define MAXTAGLOOP 2000
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** 'l_intfitsf' checks whether a given integer is in the range that
|
|
|
|
** can be converted to a float without rounding. Used in comparisons.
|
|
|
|
*/
|
|
|
|
|
|
|
|
/* number of bits in the mantissa of a float */
|
|
|
|
#define NBM (l_floatatt(MANT_DIG))
|
|
|
|
|
|
|
|
/*
|
|
|
|
** Check whether some integers may not fit in a float, testing whether
|
|
|
|
** (maxinteger >> NBM) > 0. (That implies (1 << NBM) <= maxinteger.)
|
|
|
|
** (The shifts are done in parts, to avoid shifting by more than the size
|
|
|
|
** of an integer. In a worst case, NBM == 113 for long double and
|
|
|
|
** sizeof(long) == 32.)
|
|
|
|
*/
|
|
|
|
#if ((((LUA_MAXINTEGER >> (NBM / 4)) >> (NBM / 4)) >> (NBM / 4)) \
|
|
|
|
>> (NBM - (3 * (NBM / 4)))) > 0
|
|
|
|
|
|
|
|
/* limit for integers that fit in a float */
|
|
|
|
#define MAXINTFITSF ((lua_Unsigned)1 << NBM)
|
|
|
|
|
|
|
|
/* check whether 'i' is in the interval [-MAXINTFITSF, MAXINTFITSF] */
|
|
|
|
#define l_intfitsf(i) ((MAXINTFITSF + l_castS2U(i)) <= (2 * MAXINTFITSF))
|
|
|
|
|
|
|
|
#else /* all integers fit in a float precisely */
|
|
|
|
|
|
|
|
#define l_intfitsf(i) 1
|
|
|
|
|
|
|
|
#endif
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** Try to convert a value from string to a number value.
|
|
|
|
** If the value is not a string or is a string not representing
|
|
|
|
** a valid numeral (or if coercions from strings to numbers
|
|
|
|
** are disabled via macro 'cvt2num'), do not modify 'result'
|
|
|
|
** and return 0.
|
|
|
|
*/
|
|
|
|
static int l_strton (const TValue *obj, TValue *result) {
|
|
|
|
lua_assert(obj != result);
|
|
|
|
if (!cvt2num(obj)) /* is object not a string? */
|
|
|
|
return 0;
|
|
|
|
else
|
|
|
|
return (luaO_str2num(svalue(obj), result) == vslen(obj) + 1);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** Try to convert a value to a float. The float case is already handled
|
|
|
|
** by the macro 'tonumber'.
|
|
|
|
*/
|
|
|
|
int luaV_tonumber_ (const TValue *obj, lua_Number *n) {
|
|
|
|
TValue v;
|
|
|
|
if (ttisinteger(obj)) {
|
|
|
|
*n = cast_num(ivalue(obj));
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
else if (l_strton(obj, &v)) { /* string coercible to number? */
|
|
|
|
*n = nvalue(&v); /* convert result of 'luaO_str2num' to a float */
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
else
|
|
|
|
return 0; /* conversion failed */
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** try to convert a float to an integer, rounding according to 'mode'.
|
|
|
|
*/
|
|
|
|
int luaV_flttointeger (lua_Number n, lua_Integer *p, F2Imod mode) {
|
|
|
|
lua_Number f = l_floor(n);
|
|
|
|
if (n != f) { /* not an integral value? */
|
|
|
|
if (mode == F2Ieq) return 0; /* fails if mode demands integral value */
|
|
|
|
else if (mode == F2Iceil) /* needs ceil? */
|
|
|
|
f += 1; /* convert floor to ceil (remember: n != f) */
|
|
|
|
}
|
|
|
|
return lua_numbertointeger(f, p);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** try to convert a value to an integer, rounding according to 'mode',
|
|
|
|
** without string coercion.
|
|
|
|
** ("Fast track" handled by macro 'tointegerns'.)
|
|
|
|
*/
|
|
|
|
int luaV_tointegerns (const TValue *obj, lua_Integer *p, F2Imod mode) {
|
|
|
|
if (ttisfloat(obj))
|
|
|
|
return luaV_flttointeger(fltvalue(obj), p, mode);
|
|
|
|
else if (ttisinteger(obj)) {
|
|
|
|
*p = ivalue(obj);
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
else
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** try to convert a value to an integer.
|
|
|
|
*/
|
|
|
|
int luaV_tointeger (const TValue *obj, lua_Integer *p, F2Imod mode) {
|
|
|
|
TValue v;
|
|
|
|
if (l_strton(obj, &v)) /* does 'obj' point to a numerical string? */
|
|
|
|
obj = &v; /* change it to point to its corresponding number */
|
|
|
|
return luaV_tointegerns(obj, p, mode);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** Try to convert a 'for' limit to an integer, preserving the semantics
|
|
|
|
** of the loop. Return true if the loop must not run; otherwise, '*p'
|
|
|
|
** gets the integer limit.
|
|
|
|
** (The following explanation assumes a positive step; it is valid for
|
|
|
|
** negative steps mutatis mutandis.)
|
|
|
|
** If the limit is an integer or can be converted to an integer,
|
|
|
|
** rounding down, that is the limit.
|
|
|
|
** Otherwise, check whether the limit can be converted to a float. If
|
|
|
|
** the float is too large, clip it to LUA_MAXINTEGER. If the float
|
|
|
|
** is too negative, the loop should not run, because any initial
|
|
|
|
** integer value is greater than such limit; so, the function returns
|
|
|
|
** true to signal that. (For this latter case, no integer limit would be
|
|
|
|
** correct; even a limit of LUA_MININTEGER would run the loop once for
|
|
|
|
** an initial value equal to LUA_MININTEGER.)
|
|
|
|
*/
|
|
|
|
static int forlimit (lua_State *L, lua_Integer init, const TValue *lim,
|
|
|
|
lua_Integer *p, lua_Integer step) {
|
|
|
|
if (!luaV_tointeger(lim, p, (step < 0 ? F2Iceil : F2Ifloor))) {
|
|
|
|
/* not coercible to in integer */
|
|
|
|
lua_Number flim; /* try to convert to float */
|
|
|
|
if (!tonumber(lim, &flim)) /* cannot convert to float? */
|
|
|
|
luaG_forerror(L, lim, "limit");
|
|
|
|
/* else 'flim' is a float out of integer bounds */
|
|
|
|
if (luai_numlt(0, flim)) { /* if it is positive, it is too large */
|
|
|
|
if (step < 0) return 1; /* initial value must be less than it */
|
|
|
|
*p = LUA_MAXINTEGER; /* truncate */
|
|
|
|
}
|
|
|
|
else { /* it is less than min integer */
|
|
|
|
if (step > 0) return 1; /* initial value must be greater than it */
|
|
|
|
*p = LUA_MININTEGER; /* truncate */
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return (step > 0 ? init > *p : init < *p); /* not to run? */
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** Prepare a numerical for loop (opcode OP_FORPREP).
|
|
|
|
** Return true to skip the loop. Otherwise,
|
|
|
|
** after preparation, stack will be as follows:
|
|
|
|
** ra : internal index (safe copy of the control variable)
|
|
|
|
** ra + 1 : loop counter (integer loops) or limit (float loops)
|
|
|
|
** ra + 2 : step
|
|
|
|
** ra + 3 : control variable
|
|
|
|
*/
|
|
|
|
static int forprep (lua_State *L, StkId ra) {
|
|
|
|
TValue *pinit = s2v(ra);
|
|
|
|
TValue *plimit = s2v(ra + 1);
|
|
|
|
TValue *pstep = s2v(ra + 2);
|
|
|
|
if (ttisinteger(pinit) && ttisinteger(pstep)) { /* integer loop? */
|
|
|
|
lua_Integer init = ivalue(pinit);
|
|
|
|
lua_Integer step = ivalue(pstep);
|
|
|
|
lua_Integer limit;
|
|
|
|
if (step == 0)
|
|
|
|
luaG_runerror(L, "'for' step is zero");
|
|
|
|
setivalue(s2v(ra + 3), init); /* control variable */
|
|
|
|
if (forlimit(L, init, plimit, &limit, step))
|
|
|
|
return 1; /* skip the loop */
|
|
|
|
else { /* prepare loop counter */
|
|
|
|
lua_Unsigned count;
|
|
|
|
if (step > 0) { /* ascending loop? */
|
|
|
|
count = l_castS2U(limit) - l_castS2U(init);
|
|
|
|
if (step != 1) /* avoid division in the too common case */
|
|
|
|
count /= l_castS2U(step);
|
|
|
|
}
|
|
|
|
else { /* step < 0; descending loop */
|
|
|
|
count = l_castS2U(init) - l_castS2U(limit);
|
|
|
|
/* 'step+1' avoids negating 'mininteger' */
|
|
|
|
count /= l_castS2U(-(step + 1)) + 1u;
|
|
|
|
}
|
|
|
|
/* store the counter in place of the limit (which won't be
|
|
|
|
needed anymore) */
|
|
|
|
setivalue(plimit, l_castU2S(count));
|
|
|
|
}
|
|
|
|
}
|
|
|
|
else { /* try making all values floats */
|
|
|
|
lua_Number init; lua_Number limit; lua_Number step;
|
2021-07-28 16:26:35 +00:00
|
|
|
if (l_unlikely(!tonumber(plimit, &limit)))
|
2021-03-02 13:51:10 +00:00
|
|
|
luaG_forerror(L, plimit, "limit");
|
2021-07-28 16:26:35 +00:00
|
|
|
if (l_unlikely(!tonumber(pstep, &step)))
|
2021-03-02 13:51:10 +00:00
|
|
|
luaG_forerror(L, pstep, "step");
|
2021-07-28 16:26:35 +00:00
|
|
|
if (l_unlikely(!tonumber(pinit, &init)))
|
2021-03-02 13:51:10 +00:00
|
|
|
luaG_forerror(L, pinit, "initial value");
|
|
|
|
if (step == 0)
|
|
|
|
luaG_runerror(L, "'for' step is zero");
|
|
|
|
if (luai_numlt(0, step) ? luai_numlt(limit, init)
|
|
|
|
: luai_numlt(init, limit))
|
|
|
|
return 1; /* skip the loop */
|
|
|
|
else {
|
|
|
|
/* make sure internal values are all floats */
|
|
|
|
setfltvalue(plimit, limit);
|
|
|
|
setfltvalue(pstep, step);
|
|
|
|
setfltvalue(s2v(ra), init); /* internal index */
|
|
|
|
setfltvalue(s2v(ra + 3), init); /* control variable */
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** Execute a step of a float numerical for loop, returning
|
|
|
|
** true iff the loop must continue. (The integer case is
|
|
|
|
** written online with opcode OP_FORLOOP, for performance.)
|
|
|
|
*/
|
|
|
|
static int floatforloop (StkId ra) {
|
|
|
|
lua_Number step = fltvalue(s2v(ra + 2));
|
|
|
|
lua_Number limit = fltvalue(s2v(ra + 1));
|
|
|
|
lua_Number idx = fltvalue(s2v(ra)); /* internal index */
|
|
|
|
idx = luai_numadd(L, idx, step); /* increment index */
|
|
|
|
if (luai_numlt(0, step) ? luai_numle(idx, limit)
|
|
|
|
: luai_numle(limit, idx)) {
|
|
|
|
chgfltvalue(s2v(ra), idx); /* update internal index */
|
|
|
|
setfltvalue(s2v(ra + 3), idx); /* and control variable */
|
|
|
|
return 1; /* jump back */
|
|
|
|
}
|
|
|
|
else
|
|
|
|
return 0; /* finish the loop */
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** Finish the table access 'val = t[key]'.
|
|
|
|
** if 'slot' is NULL, 't' is not a table; otherwise, 'slot' points to
|
|
|
|
** t[k] entry (which must be empty).
|
|
|
|
*/
|
|
|
|
void luaV_finishget (lua_State *L, const TValue *t, TValue *key, StkId val,
|
|
|
|
const TValue *slot) {
|
|
|
|
int loop; /* counter to avoid infinite loops */
|
|
|
|
const TValue *tm; /* metamethod */
|
|
|
|
for (loop = 0; loop < MAXTAGLOOP; loop++) {
|
|
|
|
if (slot == NULL) { /* 't' is not a table? */
|
|
|
|
lua_assert(!ttistable(t));
|
|
|
|
tm = luaT_gettmbyobj(L, t, TM_INDEX);
|
2021-07-28 16:26:35 +00:00
|
|
|
if (l_unlikely(notm(tm)))
|
2021-03-02 13:51:10 +00:00
|
|
|
luaG_typeerror(L, t, "index"); /* no metamethod */
|
|
|
|
/* else will try the metamethod */
|
|
|
|
}
|
|
|
|
else { /* 't' is a table */
|
|
|
|
lua_assert(isempty(slot));
|
|
|
|
tm = fasttm(L, hvalue(t)->metatable, TM_INDEX); /* table's metamethod */
|
|
|
|
if (tm == NULL) { /* no metamethod? */
|
|
|
|
setnilvalue(s2v(val)); /* result is nil */
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
/* else will try the metamethod */
|
|
|
|
}
|
|
|
|
if (ttisfunction(tm)) { /* is metamethod a function? */
|
|
|
|
luaT_callTMres(L, tm, t, key, val); /* call it */
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
t = tm; /* else try to access 'tm[key]' */
|
|
|
|
if (luaV_fastget(L, t, key, slot, luaH_get)) { /* fast track? */
|
|
|
|
setobj2s(L, val, slot); /* done */
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
/* else repeat (tail call 'luaV_finishget') */
|
|
|
|
}
|
|
|
|
luaG_runerror(L, "'__index' chain too long; possible loop");
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** Finish a table assignment 't[key] = val'.
|
|
|
|
** If 'slot' is NULL, 't' is not a table. Otherwise, 'slot' points
|
|
|
|
** to the entry 't[key]', or to a value with an absent key if there
|
|
|
|
** is no such entry. (The value at 'slot' must be empty, otherwise
|
|
|
|
** 'luaV_fastget' would have done the job.)
|
|
|
|
*/
|
|
|
|
void luaV_finishset (lua_State *L, const TValue *t, TValue *key,
|
|
|
|
TValue *val, const TValue *slot) {
|
|
|
|
int loop; /* counter to avoid infinite loops */
|
|
|
|
for (loop = 0; loop < MAXTAGLOOP; loop++) {
|
|
|
|
const TValue *tm; /* '__newindex' metamethod */
|
|
|
|
if (slot != NULL) { /* is 't' a table? */
|
|
|
|
Table *h = hvalue(t); /* save 't' table */
|
|
|
|
lua_assert(isempty(slot)); /* slot must be empty */
|
|
|
|
tm = fasttm(L, h->metatable, TM_NEWINDEX); /* get metamethod */
|
|
|
|
if (tm == NULL) { /* no metamethod? */
|
|
|
|
luaH_finishset(L, h, key, slot, val); /* set new value */
|
|
|
|
invalidateTMcache(h);
|
|
|
|
luaC_barrierback(L, obj2gco(h), val);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
/* else will try the metamethod */
|
|
|
|
}
|
|
|
|
else { /* not a table; check metamethod */
|
|
|
|
tm = luaT_gettmbyobj(L, t, TM_NEWINDEX);
|
2021-07-28 16:26:35 +00:00
|
|
|
if (l_unlikely(notm(tm)))
|
2021-03-02 13:51:10 +00:00
|
|
|
luaG_typeerror(L, t, "index");
|
|
|
|
}
|
|
|
|
/* try the metamethod */
|
|
|
|
if (ttisfunction(tm)) {
|
|
|
|
luaT_callTM(L, tm, t, key, val);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
t = tm; /* else repeat assignment over 'tm' */
|
|
|
|
if (luaV_fastget(L, t, key, slot, luaH_get)) {
|
|
|
|
luaV_finishfastset(L, t, slot, val);
|
|
|
|
return; /* done */
|
|
|
|
}
|
|
|
|
/* else 'return luaV_finishset(L, t, key, val, slot)' (loop) */
|
|
|
|
}
|
|
|
|
luaG_runerror(L, "'__newindex' chain too long; possible loop");
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** Compare two strings 'ls' x 'rs', returning an integer less-equal-
|
|
|
|
** -greater than zero if 'ls' is less-equal-greater than 'rs'.
|
|
|
|
** The code is a little tricky because it allows '\0' in the strings
|
|
|
|
** and it uses 'strcoll' (to respect locales) for each segments
|
|
|
|
** of the strings.
|
|
|
|
*/
|
|
|
|
static int l_strcmp (const TString *ls, const TString *rs) {
|
|
|
|
const char *l = getstr(ls);
|
|
|
|
size_t ll = tsslen(ls);
|
|
|
|
const char *r = getstr(rs);
|
|
|
|
size_t lr = tsslen(rs);
|
|
|
|
for (;;) { /* for each segment */
|
|
|
|
int temp = strcoll(l, r);
|
|
|
|
if (temp != 0) /* not equal? */
|
|
|
|
return temp; /* done */
|
|
|
|
else { /* strings are equal up to a '\0' */
|
|
|
|
size_t len = strlen(l); /* index of first '\0' in both strings */
|
|
|
|
if (len == lr) /* 'rs' is finished? */
|
|
|
|
return (len == ll) ? 0 : 1; /* check 'ls' */
|
|
|
|
else if (len == ll) /* 'ls' is finished? */
|
|
|
|
return -1; /* 'ls' is less than 'rs' ('rs' is not finished) */
|
|
|
|
/* both strings longer than 'len'; go on comparing after the '\0' */
|
|
|
|
len++;
|
|
|
|
l += len; ll -= len; r += len; lr -= len;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** Check whether integer 'i' is less than float 'f'. If 'i' has an
|
|
|
|
** exact representation as a float ('l_intfitsf'), compare numbers as
|
|
|
|
** floats. Otherwise, use the equivalence 'i < f <=> i < ceil(f)'.
|
|
|
|
** If 'ceil(f)' is out of integer range, either 'f' is greater than
|
|
|
|
** all integers or less than all integers.
|
|
|
|
** (The test with 'l_intfitsf' is only for performance; the else
|
|
|
|
** case is correct for all values, but it is slow due to the conversion
|
|
|
|
** from float to int.)
|
|
|
|
** When 'f' is NaN, comparisons must result in false.
|
|
|
|
*/
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
l_sinline int LTintfloat (lua_Integer i, lua_Number f) {
|
2021-03-02 13:51:10 +00:00
|
|
|
if (l_intfitsf(i))
|
|
|
|
return luai_numlt(cast_num(i), f); /* compare them as floats */
|
|
|
|
else { /* i < f <=> i < ceil(f) */
|
|
|
|
lua_Integer fi;
|
|
|
|
if (luaV_flttointeger(f, &fi, F2Iceil)) /* fi = ceil(f) */
|
|
|
|
return i < fi; /* compare them as integers */
|
|
|
|
else /* 'f' is either greater or less than all integers */
|
|
|
|
return f > 0; /* greater? */
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** Check whether integer 'i' is less than or equal to float 'f'.
|
|
|
|
** See comments on previous function.
|
|
|
|
*/
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
l_sinline int LEintfloat (lua_Integer i, lua_Number f) {
|
2021-03-02 13:51:10 +00:00
|
|
|
if (l_intfitsf(i))
|
|
|
|
return luai_numle(cast_num(i), f); /* compare them as floats */
|
|
|
|
else { /* i <= f <=> i <= floor(f) */
|
|
|
|
lua_Integer fi;
|
|
|
|
if (luaV_flttointeger(f, &fi, F2Ifloor)) /* fi = floor(f) */
|
|
|
|
return i <= fi; /* compare them as integers */
|
|
|
|
else /* 'f' is either greater or less than all integers */
|
|
|
|
return f > 0; /* greater? */
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** Check whether float 'f' is less than integer 'i'.
|
|
|
|
** See comments on previous function.
|
|
|
|
*/
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
l_sinline int LTfloatint (lua_Number f, lua_Integer i) {
|
2021-03-02 13:51:10 +00:00
|
|
|
if (l_intfitsf(i))
|
|
|
|
return luai_numlt(f, cast_num(i)); /* compare them as floats */
|
|
|
|
else { /* f < i <=> floor(f) < i */
|
|
|
|
lua_Integer fi;
|
|
|
|
if (luaV_flttointeger(f, &fi, F2Ifloor)) /* fi = floor(f) */
|
|
|
|
return fi < i; /* compare them as integers */
|
|
|
|
else /* 'f' is either greater or less than all integers */
|
|
|
|
return f < 0; /* less? */
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** Check whether float 'f' is less than or equal to integer 'i'.
|
|
|
|
** See comments on previous function.
|
|
|
|
*/
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
l_sinline int LEfloatint (lua_Number f, lua_Integer i) {
|
2021-03-02 13:51:10 +00:00
|
|
|
if (l_intfitsf(i))
|
|
|
|
return luai_numle(f, cast_num(i)); /* compare them as floats */
|
|
|
|
else { /* f <= i <=> ceil(f) <= i */
|
|
|
|
lua_Integer fi;
|
|
|
|
if (luaV_flttointeger(f, &fi, F2Iceil)) /* fi = ceil(f) */
|
|
|
|
return fi <= i; /* compare them as integers */
|
|
|
|
else /* 'f' is either greater or less than all integers */
|
|
|
|
return f < 0; /* less? */
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** Return 'l < r', for numbers.
|
|
|
|
*/
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
l_sinline int LTnum (const TValue *l, const TValue *r) {
|
2021-03-02 13:51:10 +00:00
|
|
|
lua_assert(ttisnumber(l) && ttisnumber(r));
|
|
|
|
if (ttisinteger(l)) {
|
|
|
|
lua_Integer li = ivalue(l);
|
|
|
|
if (ttisinteger(r))
|
|
|
|
return li < ivalue(r); /* both are integers */
|
|
|
|
else /* 'l' is int and 'r' is float */
|
|
|
|
return LTintfloat(li, fltvalue(r)); /* l < r ? */
|
|
|
|
}
|
|
|
|
else {
|
|
|
|
lua_Number lf = fltvalue(l); /* 'l' must be float */
|
|
|
|
if (ttisfloat(r))
|
|
|
|
return luai_numlt(lf, fltvalue(r)); /* both are float */
|
|
|
|
else /* 'l' is float and 'r' is int */
|
|
|
|
return LTfloatint(lf, ivalue(r));
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** Return 'l <= r', for numbers.
|
|
|
|
*/
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
l_sinline int LEnum (const TValue *l, const TValue *r) {
|
2021-03-02 13:51:10 +00:00
|
|
|
lua_assert(ttisnumber(l) && ttisnumber(r));
|
|
|
|
if (ttisinteger(l)) {
|
|
|
|
lua_Integer li = ivalue(l);
|
|
|
|
if (ttisinteger(r))
|
|
|
|
return li <= ivalue(r); /* both are integers */
|
|
|
|
else /* 'l' is int and 'r' is float */
|
|
|
|
return LEintfloat(li, fltvalue(r)); /* l <= r ? */
|
|
|
|
}
|
|
|
|
else {
|
|
|
|
lua_Number lf = fltvalue(l); /* 'l' must be float */
|
|
|
|
if (ttisfloat(r))
|
|
|
|
return luai_numle(lf, fltvalue(r)); /* both are float */
|
|
|
|
else /* 'l' is float and 'r' is int */
|
|
|
|
return LEfloatint(lf, ivalue(r));
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** return 'l < r' for non-numbers.
|
|
|
|
*/
|
|
|
|
static int lessthanothers (lua_State *L, const TValue *l, const TValue *r) {
|
|
|
|
lua_assert(!ttisnumber(l) || !ttisnumber(r));
|
|
|
|
if (ttisstring(l) && ttisstring(r)) /* both are strings? */
|
|
|
|
return l_strcmp(tsvalue(l), tsvalue(r)) < 0;
|
|
|
|
else
|
|
|
|
return luaT_callorderTM(L, l, r, TM_LT);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** Main operation less than; return 'l < r'.
|
|
|
|
*/
|
|
|
|
int luaV_lessthan (lua_State *L, const TValue *l, const TValue *r) {
|
|
|
|
if (ttisnumber(l) && ttisnumber(r)) /* both operands are numbers? */
|
|
|
|
return LTnum(l, r);
|
|
|
|
else return lessthanothers(L, l, r);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** return 'l <= r' for non-numbers.
|
|
|
|
*/
|
|
|
|
static int lessequalothers (lua_State *L, const TValue *l, const TValue *r) {
|
|
|
|
lua_assert(!ttisnumber(l) || !ttisnumber(r));
|
|
|
|
if (ttisstring(l) && ttisstring(r)) /* both are strings? */
|
|
|
|
return l_strcmp(tsvalue(l), tsvalue(r)) <= 0;
|
|
|
|
else
|
|
|
|
return luaT_callorderTM(L, l, r, TM_LE);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** Main operation less than or equal to; return 'l <= r'.
|
|
|
|
*/
|
|
|
|
int luaV_lessequal (lua_State *L, const TValue *l, const TValue *r) {
|
|
|
|
if (ttisnumber(l) && ttisnumber(r)) /* both operands are numbers? */
|
|
|
|
return LEnum(l, r);
|
|
|
|
else return lessequalothers(L, l, r);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** Main operation for equality of Lua values; return 't1 == t2'.
|
|
|
|
** L == NULL means raw equality (no metamethods)
|
|
|
|
*/
|
|
|
|
int luaV_equalobj (lua_State *L, const TValue *t1, const TValue *t2) {
|
|
|
|
const TValue *tm;
|
|
|
|
if (ttypetag(t1) != ttypetag(t2)) { /* not the same variant? */
|
|
|
|
if (ttype(t1) != ttype(t2) || ttype(t1) != LUA_TNUMBER)
|
|
|
|
return 0; /* only numbers can be equal with different variants */
|
|
|
|
else { /* two numbers with different variants */
|
2021-07-28 16:26:35 +00:00
|
|
|
/* One of them is an integer. If the other does not have an
|
|
|
|
integer value, they cannot be equal; otherwise, compare their
|
|
|
|
integer values. */
|
|
|
|
lua_Integer i1, i2;
|
|
|
|
return (luaV_tointegerns(t1, &i1, F2Ieq) &&
|
|
|
|
luaV_tointegerns(t2, &i2, F2Ieq) &&
|
|
|
|
i1 == i2);
|
2021-03-02 13:51:10 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
/* values have same type and same variant */
|
|
|
|
switch (ttypetag(t1)) {
|
|
|
|
case LUA_VNIL: case LUA_VFALSE: case LUA_VTRUE: return 1;
|
|
|
|
case LUA_VNUMINT: return (ivalue(t1) == ivalue(t2));
|
|
|
|
case LUA_VNUMFLT: return luai_numeq(fltvalue(t1), fltvalue(t2));
|
|
|
|
case LUA_VLIGHTUSERDATA: return pvalue(t1) == pvalue(t2);
|
|
|
|
case LUA_VLCF: return fvalue(t1) == fvalue(t2);
|
|
|
|
case LUA_VSHRSTR: return eqshrstr(tsvalue(t1), tsvalue(t2));
|
|
|
|
case LUA_VLNGSTR: return luaS_eqlngstr(tsvalue(t1), tsvalue(t2));
|
|
|
|
case LUA_VUSERDATA: {
|
|
|
|
if (uvalue(t1) == uvalue(t2)) return 1;
|
|
|
|
else if (L == NULL) return 0;
|
|
|
|
tm = fasttm(L, uvalue(t1)->metatable, TM_EQ);
|
|
|
|
if (tm == NULL)
|
|
|
|
tm = fasttm(L, uvalue(t2)->metatable, TM_EQ);
|
|
|
|
break; /* will try TM */
|
|
|
|
}
|
|
|
|
case LUA_VTABLE: {
|
|
|
|
if (hvalue(t1) == hvalue(t2)) return 1;
|
|
|
|
else if (L == NULL) return 0;
|
|
|
|
tm = fasttm(L, hvalue(t1)->metatable, TM_EQ);
|
|
|
|
if (tm == NULL)
|
|
|
|
tm = fasttm(L, hvalue(t2)->metatable, TM_EQ);
|
|
|
|
break; /* will try TM */
|
|
|
|
}
|
|
|
|
default:
|
|
|
|
return gcvalue(t1) == gcvalue(t2);
|
|
|
|
}
|
|
|
|
if (tm == NULL) /* no TM? */
|
|
|
|
return 0; /* objects are different */
|
|
|
|
else {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
luaT_callTMres(L, tm, t1, t2, L->top.p); /* call TM */
|
|
|
|
return !l_isfalse(s2v(L->top.p));
|
2021-03-02 13:51:10 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/* macro used by 'luaV_concat' to ensure that element at 'o' is a string */
|
|
|
|
#define tostring(L,o) \
|
|
|
|
(ttisstring(o) || (cvt2str(o) && (luaO_tostring(L, o), 1)))
|
|
|
|
|
|
|
|
#define isemptystr(o) (ttisshrstring(o) && tsvalue(o)->shrlen == 0)
|
|
|
|
|
|
|
|
/* copy strings in stack from top - n up to top - 1 to buffer */
|
|
|
|
static void copy2buff (StkId top, int n, char *buff) {
|
|
|
|
size_t tl = 0; /* size already copied */
|
|
|
|
do {
|
|
|
|
size_t l = vslen(s2v(top - n)); /* length of string being copied */
|
|
|
|
memcpy(buff + tl, svalue(s2v(top - n)), l * sizeof(char));
|
|
|
|
tl += l;
|
|
|
|
} while (--n > 0);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** Main operation for concatenation: concat 'total' values in the stack,
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
** from 'L->top.p - total' up to 'L->top.p - 1'.
|
2021-03-02 13:51:10 +00:00
|
|
|
*/
|
|
|
|
void luaV_concat (lua_State *L, int total) {
|
|
|
|
if (total == 1)
|
|
|
|
return; /* "all" values already concatenated */
|
|
|
|
do {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId top = L->top.p;
|
2021-03-02 13:51:10 +00:00
|
|
|
int n = 2; /* number of elements handled in this pass (at least 2) */
|
|
|
|
if (!(ttisstring(s2v(top - 2)) || cvt2str(s2v(top - 2))) ||
|
|
|
|
!tostring(L, s2v(top - 1)))
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
luaT_tryconcatTM(L); /* may invalidate 'top' */
|
2021-03-02 13:51:10 +00:00
|
|
|
else if (isemptystr(s2v(top - 1))) /* second operand is empty? */
|
|
|
|
cast_void(tostring(L, s2v(top - 2))); /* result is first operand */
|
|
|
|
else if (isemptystr(s2v(top - 2))) { /* first operand is empty string? */
|
|
|
|
setobjs2s(L, top - 2, top - 1); /* result is second op. */
|
|
|
|
}
|
|
|
|
else {
|
|
|
|
/* at least two non-empty string values; get as many as possible */
|
|
|
|
size_t tl = vslen(s2v(top - 1));
|
|
|
|
TString *ts;
|
|
|
|
/* collect total length and number of strings */
|
|
|
|
for (n = 1; n < total && tostring(L, s2v(top - n - 1)); n++) {
|
|
|
|
size_t l = vslen(s2v(top - n - 1));
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
if (l_unlikely(l >= (MAX_SIZE/sizeof(char)) - tl)) {
|
|
|
|
L->top.p = top - total; /* pop strings to avoid wasting stack */
|
2021-03-02 13:51:10 +00:00
|
|
|
luaG_runerror(L, "string length overflow");
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
}
|
2021-03-02 13:51:10 +00:00
|
|
|
tl += l;
|
|
|
|
}
|
|
|
|
if (tl <= LUAI_MAXSHORTLEN) { /* is result a short string? */
|
|
|
|
char buff[LUAI_MAXSHORTLEN];
|
|
|
|
copy2buff(top, n, buff); /* copy strings to buffer */
|
|
|
|
ts = luaS_newlstr(L, buff, tl);
|
|
|
|
}
|
|
|
|
else { /* long string; copy strings directly to final result */
|
|
|
|
ts = luaS_createlngstrobj(L, tl);
|
|
|
|
copy2buff(top, n, getstr(ts));
|
|
|
|
}
|
|
|
|
setsvalue2s(L, top - n, ts); /* create result */
|
|
|
|
}
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
total -= n - 1; /* got 'n' strings to create one new */
|
|
|
|
L->top.p -= n - 1; /* popped 'n' strings and pushed one */
|
2021-03-02 13:51:10 +00:00
|
|
|
} while (total > 1); /* repeat until only 1 result left */
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** Main operation 'ra = #rb'.
|
|
|
|
*/
|
|
|
|
void luaV_objlen (lua_State *L, StkId ra, const TValue *rb) {
|
|
|
|
const TValue *tm;
|
|
|
|
switch (ttypetag(rb)) {
|
|
|
|
case LUA_VTABLE: {
|
|
|
|
Table *h = hvalue(rb);
|
|
|
|
tm = fasttm(L, h->metatable, TM_LEN);
|
|
|
|
if (tm) break; /* metamethod? break switch to call it */
|
|
|
|
setivalue(s2v(ra), luaH_getn(h)); /* else primitive len */
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
case LUA_VSHRSTR: {
|
|
|
|
setivalue(s2v(ra), tsvalue(rb)->shrlen);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
case LUA_VLNGSTR: {
|
|
|
|
setivalue(s2v(ra), tsvalue(rb)->u.lnglen);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
default: { /* try metamethod */
|
|
|
|
tm = luaT_gettmbyobj(L, rb, TM_LEN);
|
2021-07-28 16:26:35 +00:00
|
|
|
if (l_unlikely(notm(tm))) /* no metamethod? */
|
2021-03-02 13:51:10 +00:00
|
|
|
luaG_typeerror(L, rb, "get length of");
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
luaT_callTMres(L, tm, rb, rb, ra);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** Integer division; return 'm // n', that is, floor(m/n).
|
|
|
|
** C division truncates its result (rounds towards zero).
|
|
|
|
** 'floor(q) == trunc(q)' when 'q >= 0' or when 'q' is integer,
|
|
|
|
** otherwise 'floor(q) == trunc(q) - 1'.
|
|
|
|
*/
|
|
|
|
lua_Integer luaV_idiv (lua_State *L, lua_Integer m, lua_Integer n) {
|
2021-07-28 16:26:35 +00:00
|
|
|
if (l_unlikely(l_castS2U(n) + 1u <= 1u)) { /* special cases: -1 or 0 */
|
2021-03-02 13:51:10 +00:00
|
|
|
if (n == 0)
|
|
|
|
luaG_runerror(L, "attempt to divide by zero");
|
|
|
|
return intop(-, 0, m); /* n==-1; avoid overflow with 0x80000...//-1 */
|
|
|
|
}
|
|
|
|
else {
|
|
|
|
lua_Integer q = m / n; /* perform C division */
|
|
|
|
if ((m ^ n) < 0 && m % n != 0) /* 'm/n' would be negative non-integer? */
|
|
|
|
q -= 1; /* correct result for different rounding */
|
|
|
|
return q;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** Integer modulus; return 'm % n'. (Assume that C '%' with
|
|
|
|
** negative operands follows C99 behavior. See previous comment
|
|
|
|
** about luaV_idiv.)
|
|
|
|
*/
|
|
|
|
lua_Integer luaV_mod (lua_State *L, lua_Integer m, lua_Integer n) {
|
2021-07-28 16:26:35 +00:00
|
|
|
if (l_unlikely(l_castS2U(n) + 1u <= 1u)) { /* special cases: -1 or 0 */
|
2021-03-02 13:51:10 +00:00
|
|
|
if (n == 0)
|
|
|
|
luaG_runerror(L, "attempt to perform 'n%%0'");
|
|
|
|
return 0; /* m % -1 == 0; avoid overflow with 0x80000...%-1 */
|
|
|
|
}
|
|
|
|
else {
|
|
|
|
lua_Integer r = m % n;
|
|
|
|
if (r != 0 && (r ^ n) < 0) /* 'm/n' would be non-integer negative? */
|
|
|
|
r += n; /* correct result for different rounding */
|
|
|
|
return r;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** Float modulus
|
|
|
|
*/
|
|
|
|
lua_Number luaV_modf (lua_State *L, lua_Number m, lua_Number n) {
|
|
|
|
lua_Number r;
|
|
|
|
luai_nummod(L, m, n, r);
|
|
|
|
return r;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/* number of bits in an integer */
|
|
|
|
#define NBITS cast_int(sizeof(lua_Integer) * CHAR_BIT)
|
|
|
|
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
|
2021-03-02 13:51:10 +00:00
|
|
|
/*
|
|
|
|
** Shift left operation. (Shift right just negates 'y'.)
|
|
|
|
*/
|
|
|
|
lua_Integer luaV_shiftl (lua_Integer x, lua_Integer y) {
|
|
|
|
if (y < 0) { /* shift right? */
|
|
|
|
if (y <= -NBITS) return 0;
|
|
|
|
else return intop(>>, x, -y);
|
|
|
|
}
|
|
|
|
else { /* shift left */
|
|
|
|
if (y >= NBITS) return 0;
|
|
|
|
else return intop(<<, x, y);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** create a new Lua closure, push it in the stack, and initialize
|
|
|
|
** its upvalues.
|
|
|
|
*/
|
|
|
|
static void pushclosure (lua_State *L, Proto *p, UpVal **encup, StkId base,
|
|
|
|
StkId ra) {
|
|
|
|
int nup = p->sizeupvalues;
|
|
|
|
Upvaldesc *uv = p->upvalues;
|
|
|
|
int i;
|
|
|
|
LClosure *ncl = luaF_newLclosure(L, nup);
|
|
|
|
ncl->p = p;
|
|
|
|
setclLvalue2s(L, ra, ncl); /* anchor new closure in stack */
|
|
|
|
for (i = 0; i < nup; i++) { /* fill in its upvalues */
|
|
|
|
if (uv[i].instack) /* upvalue refers to local variable? */
|
|
|
|
ncl->upvals[i] = luaF_findupval(L, base + uv[i].idx);
|
|
|
|
else /* get upvalue from enclosing function */
|
|
|
|
ncl->upvals[i] = encup[uv[i].idx];
|
|
|
|
luaC_objbarrier(L, ncl, ncl->upvals[i]);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** finish execution of an opcode interrupted by a yield
|
|
|
|
*/
|
|
|
|
void luaV_finishOp (lua_State *L) {
|
|
|
|
CallInfo *ci = L->ci;
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId base = ci->func.p + 1;
|
2021-03-02 13:51:10 +00:00
|
|
|
Instruction inst = *(ci->u.l.savedpc - 1); /* interrupted instruction */
|
|
|
|
OpCode op = GET_OPCODE(inst);
|
|
|
|
switch (op) { /* finish its execution */
|
|
|
|
case OP_MMBIN: case OP_MMBINI: case OP_MMBINK: {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
setobjs2s(L, base + GETARG_A(*(ci->u.l.savedpc - 2)), --L->top.p);
|
2021-03-02 13:51:10 +00:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
case OP_UNM: case OP_BNOT: case OP_LEN:
|
|
|
|
case OP_GETTABUP: case OP_GETTABLE: case OP_GETI:
|
|
|
|
case OP_GETFIELD: case OP_SELF: {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
setobjs2s(L, base + GETARG_A(inst), --L->top.p);
|
2021-03-02 13:51:10 +00:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
case OP_LT: case OP_LE:
|
|
|
|
case OP_LTI: case OP_LEI:
|
|
|
|
case OP_GTI: case OP_GEI:
|
|
|
|
case OP_EQ: { /* note that 'OP_EQI'/'OP_EQK' cannot yield */
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
int res = !l_isfalse(s2v(L->top.p - 1));
|
|
|
|
L->top.p--;
|
2021-03-02 13:51:10 +00:00
|
|
|
#if defined(LUA_COMPAT_LT_LE)
|
|
|
|
if (ci->callstatus & CIST_LEQ) { /* "<=" using "<" instead? */
|
|
|
|
ci->callstatus ^= CIST_LEQ; /* clear mark */
|
|
|
|
res = !res; /* negate result */
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
lua_assert(GET_OPCODE(*ci->u.l.savedpc) == OP_JMP);
|
|
|
|
if (res != GETARG_k(inst)) /* condition failed? */
|
|
|
|
ci->u.l.savedpc++; /* skip jump instruction */
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
case OP_CONCAT: {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId top = L->top.p - 1; /* top when 'luaT_tryconcatTM' was called */
|
2021-03-02 13:51:10 +00:00
|
|
|
int a = GETARG_A(inst); /* first element to concatenate */
|
|
|
|
int total = cast_int(top - 1 - (base + a)); /* yet to concatenate */
|
|
|
|
setobjs2s(L, top - 2, top); /* put TM result in proper position */
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
L->top.p = top - 1; /* top is one after last element (at top-2) */
|
2021-03-02 13:51:10 +00:00
|
|
|
luaV_concat(L, total); /* concat them (may yield again) */
|
|
|
|
break;
|
|
|
|
}
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
case OP_CLOSE: { /* yielded closing variables */
|
2021-03-02 13:51:10 +00:00
|
|
|
ci->u.l.savedpc--; /* repeat instruction to close other vars. */
|
|
|
|
break;
|
|
|
|
}
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
case OP_RETURN: { /* yielded closing variables */
|
|
|
|
StkId ra = base + GETARG_A(inst);
|
|
|
|
/* adjust top to signal correct number of returns, in case the
|
|
|
|
return is "up to top" ('isIT') */
|
|
|
|
L->top.p = ra + ci->u2.nres;
|
|
|
|
/* repeat instruction to close other vars. and complete the return */
|
|
|
|
ci->u.l.savedpc--;
|
|
|
|
break;
|
|
|
|
}
|
2021-03-02 13:51:10 +00:00
|
|
|
default: {
|
|
|
|
/* only these other opcodes can yield */
|
|
|
|
lua_assert(op == OP_TFORCALL || op == OP_CALL ||
|
|
|
|
op == OP_TAILCALL || op == OP_SETTABUP || op == OP_SETTABLE ||
|
|
|
|
op == OP_SETI || op == OP_SETFIELD);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** {==================================================================
|
|
|
|
** Macros for arithmetic/bitwise/comparison opcodes in 'luaV_execute'
|
|
|
|
** ===================================================================
|
|
|
|
*/
|
|
|
|
|
|
|
|
#define l_addi(L,a,b) intop(+, a, b)
|
|
|
|
#define l_subi(L,a,b) intop(-, a, b)
|
|
|
|
#define l_muli(L,a,b) intop(*, a, b)
|
|
|
|
#define l_band(a,b) intop(&, a, b)
|
|
|
|
#define l_bor(a,b) intop(|, a, b)
|
|
|
|
#define l_bxor(a,b) intop(^, a, b)
|
|
|
|
|
|
|
|
#define l_lti(a,b) (a < b)
|
|
|
|
#define l_lei(a,b) (a <= b)
|
|
|
|
#define l_gti(a,b) (a > b)
|
|
|
|
#define l_gei(a,b) (a >= b)
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** Arithmetic operations with immediate operands. 'iop' is the integer
|
|
|
|
** operation, 'fop' is the float operation.
|
|
|
|
*/
|
|
|
|
#define op_arithI(L,iop,fop) { \
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i); \
|
2021-03-02 13:51:10 +00:00
|
|
|
TValue *v1 = vRB(i); \
|
|
|
|
int imm = GETARG_sC(i); \
|
|
|
|
if (ttisinteger(v1)) { \
|
|
|
|
lua_Integer iv1 = ivalue(v1); \
|
|
|
|
pc++; setivalue(s2v(ra), iop(L, iv1, imm)); \
|
|
|
|
} \
|
|
|
|
else if (ttisfloat(v1)) { \
|
|
|
|
lua_Number nb = fltvalue(v1); \
|
|
|
|
lua_Number fimm = cast_num(imm); \
|
|
|
|
pc++; setfltvalue(s2v(ra), fop(L, nb, fimm)); \
|
|
|
|
}}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** Auxiliary function for arithmetic operations over floats and others
|
|
|
|
** with two register operands.
|
|
|
|
*/
|
|
|
|
#define op_arithf_aux(L,v1,v2,fop) { \
|
|
|
|
lua_Number n1; lua_Number n2; \
|
|
|
|
if (tonumberns(v1, n1) && tonumberns(v2, n2)) { \
|
|
|
|
pc++; setfltvalue(s2v(ra), fop(L, n1, n2)); \
|
|
|
|
}}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** Arithmetic operations over floats and others with register operands.
|
|
|
|
*/
|
|
|
|
#define op_arithf(L,fop) { \
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i); \
|
2021-03-02 13:51:10 +00:00
|
|
|
TValue *v1 = vRB(i); \
|
|
|
|
TValue *v2 = vRC(i); \
|
|
|
|
op_arithf_aux(L, v1, v2, fop); }
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** Arithmetic operations with K operands for floats.
|
|
|
|
*/
|
|
|
|
#define op_arithfK(L,fop) { \
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i); \
|
2021-03-02 13:51:10 +00:00
|
|
|
TValue *v1 = vRB(i); \
|
2021-07-28 16:26:35 +00:00
|
|
|
TValue *v2 = KC(i); lua_assert(ttisnumber(v2)); \
|
2021-03-02 13:51:10 +00:00
|
|
|
op_arithf_aux(L, v1, v2, fop); }
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** Arithmetic operations over integers and floats.
|
|
|
|
*/
|
|
|
|
#define op_arith_aux(L,v1,v2,iop,fop) { \
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i); \
|
2021-03-02 13:51:10 +00:00
|
|
|
if (ttisinteger(v1) && ttisinteger(v2)) { \
|
|
|
|
lua_Integer i1 = ivalue(v1); lua_Integer i2 = ivalue(v2); \
|
|
|
|
pc++; setivalue(s2v(ra), iop(L, i1, i2)); \
|
|
|
|
} \
|
|
|
|
else op_arithf_aux(L, v1, v2, fop); }
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** Arithmetic operations with register operands.
|
|
|
|
*/
|
|
|
|
#define op_arith(L,iop,fop) { \
|
|
|
|
TValue *v1 = vRB(i); \
|
|
|
|
TValue *v2 = vRC(i); \
|
|
|
|
op_arith_aux(L, v1, v2, iop, fop); }
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** Arithmetic operations with K operands.
|
|
|
|
*/
|
|
|
|
#define op_arithK(L,iop,fop) { \
|
|
|
|
TValue *v1 = vRB(i); \
|
2021-07-28 16:26:35 +00:00
|
|
|
TValue *v2 = KC(i); lua_assert(ttisnumber(v2)); \
|
2021-03-02 13:51:10 +00:00
|
|
|
op_arith_aux(L, v1, v2, iop, fop); }
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** Bitwise operations with constant operand.
|
|
|
|
*/
|
|
|
|
#define op_bitwiseK(L,op) { \
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i); \
|
2021-03-02 13:51:10 +00:00
|
|
|
TValue *v1 = vRB(i); \
|
|
|
|
TValue *v2 = KC(i); \
|
|
|
|
lua_Integer i1; \
|
|
|
|
lua_Integer i2 = ivalue(v2); \
|
|
|
|
if (tointegerns(v1, &i1)) { \
|
|
|
|
pc++; setivalue(s2v(ra), op(i1, i2)); \
|
|
|
|
}}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** Bitwise operations with register operands.
|
|
|
|
*/
|
|
|
|
#define op_bitwise(L,op) { \
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i); \
|
2021-03-02 13:51:10 +00:00
|
|
|
TValue *v1 = vRB(i); \
|
|
|
|
TValue *v2 = vRC(i); \
|
|
|
|
lua_Integer i1; lua_Integer i2; \
|
|
|
|
if (tointegerns(v1, &i1) && tointegerns(v2, &i2)) { \
|
|
|
|
pc++; setivalue(s2v(ra), op(i1, i2)); \
|
|
|
|
}}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** Order operations with register operands. 'opn' actually works
|
|
|
|
** for all numbers, but the fast track improves performance for
|
|
|
|
** integers.
|
|
|
|
*/
|
|
|
|
#define op_order(L,opi,opn,other) { \
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i); \
|
2021-03-02 13:51:10 +00:00
|
|
|
int cond; \
|
|
|
|
TValue *rb = vRB(i); \
|
|
|
|
if (ttisinteger(s2v(ra)) && ttisinteger(rb)) { \
|
|
|
|
lua_Integer ia = ivalue(s2v(ra)); \
|
|
|
|
lua_Integer ib = ivalue(rb); \
|
|
|
|
cond = opi(ia, ib); \
|
|
|
|
} \
|
|
|
|
else if (ttisnumber(s2v(ra)) && ttisnumber(rb)) \
|
|
|
|
cond = opn(s2v(ra), rb); \
|
|
|
|
else \
|
|
|
|
Protect(cond = other(L, s2v(ra), rb)); \
|
|
|
|
docondjump(); }
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** Order operations with immediate operand. (Immediate operand is
|
|
|
|
** always small enough to have an exact representation as a float.)
|
|
|
|
*/
|
|
|
|
#define op_orderI(L,opi,opf,inv,tm) { \
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i); \
|
2021-03-02 13:51:10 +00:00
|
|
|
int cond; \
|
|
|
|
int im = GETARG_sB(i); \
|
|
|
|
if (ttisinteger(s2v(ra))) \
|
|
|
|
cond = opi(ivalue(s2v(ra)), im); \
|
|
|
|
else if (ttisfloat(s2v(ra))) { \
|
|
|
|
lua_Number fa = fltvalue(s2v(ra)); \
|
|
|
|
lua_Number fim = cast_num(im); \
|
|
|
|
cond = opf(fa, fim); \
|
|
|
|
} \
|
|
|
|
else { \
|
|
|
|
int isf = GETARG_C(i); \
|
|
|
|
Protect(cond = luaT_callorderiTM(L, s2v(ra), im, inv, isf, tm)); \
|
|
|
|
} \
|
|
|
|
docondjump(); }
|
|
|
|
|
|
|
|
/* }================================================================== */
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** {==================================================================
|
|
|
|
** Function 'luaV_execute': main interpreter loop
|
|
|
|
** ===================================================================
|
|
|
|
*/
|
|
|
|
|
|
|
|
/*
|
|
|
|
** some macros for common tasks in 'luaV_execute'
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
|
|
#define RA(i) (base+GETARG_A(i))
|
|
|
|
#define RB(i) (base+GETARG_B(i))
|
|
|
|
#define vRB(i) s2v(RB(i))
|
|
|
|
#define KB(i) (k+GETARG_B(i))
|
|
|
|
#define RC(i) (base+GETARG_C(i))
|
|
|
|
#define vRC(i) s2v(RC(i))
|
|
|
|
#define KC(i) (k+GETARG_C(i))
|
|
|
|
#define RKC(i) ((TESTARG_k(i)) ? k + GETARG_C(i) : s2v(base + GETARG_C(i)))
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
#define updatetrap(ci) (trap = ci->u.l.trap)
|
|
|
|
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
#define updatebase(ci) (base = ci->func.p + 1)
|
2021-03-02 13:51:10 +00:00
|
|
|
|
|
|
|
|
2021-07-28 16:26:35 +00:00
|
|
|
#define updatestack(ci) \
|
|
|
|
{ if (l_unlikely(trap)) { updatebase(ci); ra = RA(i); } }
|
2021-03-02 13:51:10 +00:00
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** Execute a jump instruction. The 'updatetrap' allows signals to stop
|
|
|
|
** tight loops. (Without it, the local copy of 'trap' could never change.)
|
|
|
|
*/
|
|
|
|
#define dojump(ci,i,e) { pc += GETARG_sJ(i) + e; updatetrap(ci); }
|
|
|
|
|
|
|
|
|
|
|
|
/* for test instructions, execute the jump instruction that follows it */
|
|
|
|
#define donextjump(ci) { Instruction ni = *pc; dojump(ci, ni, 1); }
|
|
|
|
|
|
|
|
/*
|
|
|
|
** do a conditional jump: skip next instruction if 'cond' is not what
|
|
|
|
** was expected (parameter 'k'), else do next instruction, which must
|
|
|
|
** be a jump.
|
|
|
|
*/
|
|
|
|
#define docondjump() if (cond != GETARG_k(i)) pc++; else donextjump(ci);
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** Correct global 'pc'.
|
|
|
|
*/
|
|
|
|
#define savepc(L) (ci->u.l.savedpc = pc)
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** Whenever code can raise errors, the global 'pc' and the global
|
|
|
|
** 'top' must be correct to report occasional errors.
|
|
|
|
*/
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
#define savestate(L,ci) (savepc(L), L->top.p = ci->top.p)
|
2021-03-02 13:51:10 +00:00
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
** Protect code that, in general, can raise errors, reallocate the
|
|
|
|
** stack, and change the hooks.
|
|
|
|
*/
|
|
|
|
#define Protect(exp) (savestate(L,ci), (exp), updatetrap(ci))
|
|
|
|
|
|
|
|
/* special version that does not change the top */
|
|
|
|
#define ProtectNT(exp) (savepc(L), (exp), updatetrap(ci))
|
|
|
|
|
|
|
|
/*
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
** Protect code that can only raise errors. (That is, it cannot change
|
2021-03-02 13:51:10 +00:00
|
|
|
** the stack or hooks.)
|
|
|
|
*/
|
|
|
|
#define halfProtect(exp) (savestate(L,ci), (exp))
|
|
|
|
|
|
|
|
/* 'c' is the limit of live values in the stack */
|
|
|
|
#define checkGC(L,c) \
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
{ luaC_condGC(L, (savepc(L), L->top.p = (c)), \
|
2021-03-02 13:51:10 +00:00
|
|
|
updatetrap(ci)); \
|
|
|
|
luai_threadyield(L); }
|
|
|
|
|
|
|
|
|
|
|
|
/* fetch an instruction and prepare its execution */
|
|
|
|
#define vmfetch() { \
|
2021-07-28 16:26:35 +00:00
|
|
|
if (l_unlikely(trap)) { /* stack reallocation or hooks? */ \
|
2021-03-02 13:51:10 +00:00
|
|
|
trap = luaG_traceexec(L, pc); /* handle hooks */ \
|
|
|
|
updatebase(ci); /* correct stack */ \
|
|
|
|
} \
|
|
|
|
i = *(pc++); \
|
|
|
|
}
|
|
|
|
|
|
|
|
#define vmdispatch(o) switch(o)
|
|
|
|
#define vmcase(l) case l:
|
|
|
|
#define vmbreak break
|
|
|
|
|
|
|
|
|
|
|
|
void luaV_execute (lua_State *L, CallInfo *ci) {
|
|
|
|
LClosure *cl;
|
|
|
|
TValue *k;
|
|
|
|
StkId base;
|
|
|
|
const Instruction *pc;
|
|
|
|
int trap;
|
|
|
|
#if LUA_USE_JUMPTABLE
|
2021-03-07 21:26:57 +00:00
|
|
|
#include "third_party/lua/ljumptab.inc"
|
2021-03-02 13:51:10 +00:00
|
|
|
#endif
|
|
|
|
startfunc:
|
|
|
|
trap = L->hookmask;
|
|
|
|
returning: /* trap already set */
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
cl = clLvalue(s2v(ci->func.p));
|
2021-03-02 13:51:10 +00:00
|
|
|
k = cl->p->k;
|
|
|
|
pc = ci->u.l.savedpc;
|
2021-07-28 16:26:35 +00:00
|
|
|
if (l_unlikely(trap)) {
|
2021-03-02 13:51:10 +00:00
|
|
|
if (pc == cl->p->code) { /* first instruction (not resuming)? */
|
|
|
|
if (cl->p->is_vararg)
|
|
|
|
trap = 0; /* hooks will start after VARARGPREP instruction */
|
|
|
|
else /* check 'call' hook */
|
|
|
|
luaD_hookcall(L, ci);
|
|
|
|
}
|
|
|
|
ci->u.l.trap = 1; /* assume trap is on, for now */
|
|
|
|
}
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
base = ci->func.p + 1;
|
2021-03-02 13:51:10 +00:00
|
|
|
/* main loop of interpreter */
|
|
|
|
for (;;) {
|
|
|
|
Instruction i; /* instruction being executed */
|
|
|
|
vmfetch();
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
#if 0
|
|
|
|
/* low-level line tracing for debugging Lua */
|
|
|
|
printf("line: %d\n", luaG_getfuncline(cl->p, pcRel(pc, cl->p)));
|
|
|
|
#endif
|
|
|
|
lua_assert(base == ci->func.p + 1);
|
|
|
|
lua_assert(base <= L->top.p && L->top.p <= L->stack_last.p);
|
2021-03-02 13:51:10 +00:00
|
|
|
/* invalidate top for instructions not expecting it */
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
lua_assert(isIT(i) || (cast_void(L->top.p = base), 1));
|
2021-03-02 13:51:10 +00:00
|
|
|
vmdispatch (GET_OPCODE(i)) {
|
|
|
|
vmcase(OP_MOVE) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
setobjs2s(L, ra, RB(i));
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_LOADI) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
lua_Integer b = GETARG_sBx(i);
|
|
|
|
setivalue(s2v(ra), b);
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_LOADF) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
int b = GETARG_sBx(i);
|
|
|
|
setfltvalue(s2v(ra), cast_num(b));
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_LOADK) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
TValue *rb = k + GETARG_Bx(i);
|
|
|
|
setobj2s(L, ra, rb);
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_LOADKX) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
TValue *rb;
|
|
|
|
rb = k + GETARG_Ax(*pc); pc++;
|
|
|
|
setobj2s(L, ra, rb);
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_LOADFALSE) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
setbfvalue(s2v(ra));
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_LFALSESKIP) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
setbfvalue(s2v(ra));
|
|
|
|
pc++; /* skip next instruction */
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_LOADTRUE) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
setbtvalue(s2v(ra));
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_LOADNIL) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
int b = GETARG_B(i);
|
|
|
|
do {
|
|
|
|
setnilvalue(s2v(ra++));
|
|
|
|
} while (b--);
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_GETUPVAL) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
int b = GETARG_B(i);
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
setobj2s(L, ra, cl->upvals[b]->v.p);
|
2021-03-02 13:51:10 +00:00
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_SETUPVAL) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
UpVal *uv = cl->upvals[GETARG_B(i)];
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
setobj(L, uv->v.p, s2v(ra));
|
2021-03-02 13:51:10 +00:00
|
|
|
luaC_barrier(L, uv, s2v(ra));
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_GETTABUP) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
const TValue *slot;
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
TValue *upval = cl->upvals[GETARG_B(i)]->v.p;
|
2021-03-02 13:51:10 +00:00
|
|
|
TValue *rc = KC(i);
|
|
|
|
TString *key = tsvalue(rc); /* key must be a string */
|
|
|
|
if (luaV_fastget(L, upval, key, slot, luaH_getshortstr)) {
|
|
|
|
setobj2s(L, ra, slot);
|
|
|
|
}
|
|
|
|
else
|
|
|
|
Protect(luaV_finishget(L, upval, rc, ra, slot));
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_GETTABLE) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
const TValue *slot;
|
|
|
|
TValue *rb = vRB(i);
|
|
|
|
TValue *rc = vRC(i);
|
|
|
|
lua_Unsigned n;
|
|
|
|
if (ttisinteger(rc) /* fast track for integers? */
|
|
|
|
? (cast_void(n = ivalue(rc)), luaV_fastgeti(L, rb, n, slot))
|
|
|
|
: luaV_fastget(L, rb, rc, slot, luaH_get)) {
|
|
|
|
setobj2s(L, ra, slot);
|
|
|
|
}
|
|
|
|
else
|
|
|
|
Protect(luaV_finishget(L, rb, rc, ra, slot));
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_GETI) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
const TValue *slot;
|
|
|
|
TValue *rb = vRB(i);
|
|
|
|
int c = GETARG_C(i);
|
2021-08-07 22:20:33 +00:00
|
|
|
if (luaV_fastgeti(L, rb, c, slot)) {
|
2021-03-02 13:51:10 +00:00
|
|
|
setobj2s(L, ra, slot);
|
|
|
|
}
|
|
|
|
else {
|
|
|
|
TValue key;
|
|
|
|
setivalue(&key, c);
|
|
|
|
Protect(luaV_finishget(L, rb, &key, ra, slot));
|
|
|
|
}
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_GETFIELD) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
const TValue *slot;
|
|
|
|
TValue *rb = vRB(i);
|
|
|
|
TValue *rc = KC(i);
|
|
|
|
TString *key = tsvalue(rc); /* key must be a string */
|
|
|
|
if (luaV_fastget(L, rb, key, slot, luaH_getshortstr)) {
|
|
|
|
setobj2s(L, ra, slot);
|
|
|
|
}
|
|
|
|
else
|
|
|
|
Protect(luaV_finishget(L, rb, rc, ra, slot));
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_SETTABUP) {
|
|
|
|
const TValue *slot;
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
TValue *upval = cl->upvals[GETARG_A(i)]->v.p;
|
2021-03-02 13:51:10 +00:00
|
|
|
TValue *rb = KB(i);
|
|
|
|
TValue *rc = RKC(i);
|
|
|
|
TString *key = tsvalue(rb); /* key must be a string */
|
|
|
|
if (luaV_fastget(L, upval, key, slot, luaH_getshortstr)) {
|
|
|
|
luaV_finishfastset(L, upval, slot, rc);
|
|
|
|
}
|
|
|
|
else
|
|
|
|
Protect(luaV_finishset(L, upval, rb, rc, slot));
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_SETTABLE) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
const TValue *slot;
|
|
|
|
TValue *rb = vRB(i); /* key (table is in 'ra') */
|
|
|
|
TValue *rc = RKC(i); /* value */
|
|
|
|
lua_Unsigned n;
|
|
|
|
if (ttisinteger(rb) /* fast track for integers? */
|
|
|
|
? (cast_void(n = ivalue(rb)), luaV_fastgeti(L, s2v(ra), n, slot))
|
|
|
|
: luaV_fastget(L, s2v(ra), rb, slot, luaH_get)) {
|
|
|
|
luaV_finishfastset(L, s2v(ra), slot, rc);
|
|
|
|
}
|
|
|
|
else
|
|
|
|
Protect(luaV_finishset(L, s2v(ra), rb, rc, slot));
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_SETI) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
const TValue *slot;
|
|
|
|
int c = GETARG_B(i);
|
|
|
|
TValue *rc = RKC(i);
|
2021-08-07 22:20:33 +00:00
|
|
|
if (luaV_fastgeti(L, s2v(ra), c, slot)) {
|
2021-03-02 13:51:10 +00:00
|
|
|
luaV_finishfastset(L, s2v(ra), slot, rc);
|
|
|
|
}
|
|
|
|
else {
|
|
|
|
TValue key;
|
|
|
|
setivalue(&key, c);
|
|
|
|
Protect(luaV_finishset(L, s2v(ra), &key, rc, slot));
|
|
|
|
}
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_SETFIELD) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
const TValue *slot;
|
|
|
|
TValue *rb = KB(i);
|
|
|
|
TValue *rc = RKC(i);
|
|
|
|
TString *key = tsvalue(rb); /* key must be a string */
|
|
|
|
if (luaV_fastget(L, s2v(ra), key, slot, luaH_getshortstr)) {
|
|
|
|
luaV_finishfastset(L, s2v(ra), slot, rc);
|
|
|
|
}
|
|
|
|
else
|
|
|
|
Protect(luaV_finishset(L, s2v(ra), rb, rc, slot));
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_NEWTABLE) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
int b = GETARG_B(i); /* log2(hash size) + 1 */
|
|
|
|
int c = GETARG_C(i); /* array size */
|
|
|
|
Table *t;
|
|
|
|
if (b > 0)
|
|
|
|
b = 1 << (b - 1); /* size is 2^(b - 1) */
|
|
|
|
lua_assert((!TESTARG_k(i)) == (GETARG_Ax(*pc) == 0));
|
|
|
|
if (TESTARG_k(i)) /* non-zero extra argument? */
|
|
|
|
c += GETARG_Ax(*pc) * (MAXARG_C + 1); /* add it to size */
|
|
|
|
pc++; /* skip extra argument */
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
L->top.p = ra + 1; /* correct top in case of emergency GC */
|
2021-03-02 13:51:10 +00:00
|
|
|
t = luaH_new(L); /* memory allocation */
|
|
|
|
sethvalue2s(L, ra, t);
|
2021-08-07 22:20:33 +00:00
|
|
|
if (b != 0 || c != 0)
|
|
|
|
luaH_resize(L, t, c, b); /* idem */
|
2021-03-02 13:51:10 +00:00
|
|
|
checkGC(L, ra + 1);
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_SELF) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
const TValue *slot;
|
|
|
|
TValue *rb = vRB(i);
|
|
|
|
TValue *rc = RKC(i);
|
|
|
|
TString *key = tsvalue(rc); /* key must be a string */
|
|
|
|
setobj2s(L, ra + 1, rb);
|
|
|
|
if (luaV_fastget(L, rb, key, slot, luaH_getstr)) {
|
|
|
|
setobj2s(L, ra, slot);
|
|
|
|
}
|
|
|
|
else
|
|
|
|
Protect(luaV_finishget(L, rb, rc, ra, slot));
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_ADDI) {
|
|
|
|
op_arithI(L, l_addi, luai_numadd);
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_ADDK) {
|
|
|
|
op_arithK(L, l_addi, luai_numadd);
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_SUBK) {
|
|
|
|
op_arithK(L, l_subi, luai_numsub);
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_MULK) {
|
|
|
|
op_arithK(L, l_muli, luai_nummul);
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_MODK) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
savestate(L, ci); /* in case of division by 0 */
|
2021-03-02 13:51:10 +00:00
|
|
|
op_arithK(L, luaV_mod, luaV_modf);
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_POWK) {
|
|
|
|
op_arithfK(L, luai_numpow);
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_DIVK) {
|
|
|
|
op_arithfK(L, luai_numdiv);
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_IDIVK) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
savestate(L, ci); /* in case of division by 0 */
|
2021-03-02 13:51:10 +00:00
|
|
|
op_arithK(L, luaV_idiv, luai_numidiv);
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_BANDK) {
|
|
|
|
op_bitwiseK(L, l_band);
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_BORK) {
|
|
|
|
op_bitwiseK(L, l_bor);
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_BXORK) {
|
|
|
|
op_bitwiseK(L, l_bxor);
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_SHRI) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
TValue *rb = vRB(i);
|
|
|
|
int ic = GETARG_sC(i);
|
|
|
|
lua_Integer ib;
|
|
|
|
if (tointegerns(rb, &ib)) {
|
|
|
|
pc++; setivalue(s2v(ra), luaV_shiftl(ib, -ic));
|
|
|
|
}
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_SHLI) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
TValue *rb = vRB(i);
|
|
|
|
int ic = GETARG_sC(i);
|
|
|
|
lua_Integer ib;
|
|
|
|
if (tointegerns(rb, &ib)) {
|
|
|
|
pc++; setivalue(s2v(ra), luaV_shiftl(ic, ib));
|
|
|
|
}
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_ADD) {
|
|
|
|
op_arith(L, l_addi, luai_numadd);
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_SUB) {
|
|
|
|
op_arith(L, l_subi, luai_numsub);
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_MUL) {
|
|
|
|
op_arith(L, l_muli, luai_nummul);
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_MOD) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
savestate(L, ci); /* in case of division by 0 */
|
2021-03-02 13:51:10 +00:00
|
|
|
op_arith(L, luaV_mod, luaV_modf);
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_POW) {
|
|
|
|
op_arithf(L, luai_numpow);
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_DIV) { /* float division (always with floats) */
|
|
|
|
op_arithf(L, luai_numdiv);
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_IDIV) { /* floor division */
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
savestate(L, ci); /* in case of division by 0 */
|
2021-03-02 13:51:10 +00:00
|
|
|
op_arith(L, luaV_idiv, luai_numidiv);
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_BAND) {
|
|
|
|
op_bitwise(L, l_band);
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_BOR) {
|
|
|
|
op_bitwise(L, l_bor);
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_BXOR) {
|
|
|
|
op_bitwise(L, l_bxor);
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_SHR) {
|
|
|
|
op_bitwise(L, luaV_shiftr);
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_SHL) {
|
|
|
|
op_bitwise(L, luaV_shiftl);
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_MMBIN) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
Instruction pi = *(pc - 2); /* original arith. expression */
|
|
|
|
TValue *rb = vRB(i);
|
|
|
|
TMS tm = (TMS)GETARG_C(i);
|
|
|
|
StkId result = RA(pi);
|
|
|
|
lua_assert(OP_ADD <= GET_OPCODE(pi) && GET_OPCODE(pi) <= OP_SHR);
|
|
|
|
Protect(luaT_trybinTM(L, s2v(ra), rb, result, tm));
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_MMBINI) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
Instruction pi = *(pc - 2); /* original arith. expression */
|
|
|
|
int imm = GETARG_sB(i);
|
|
|
|
TMS tm = (TMS)GETARG_C(i);
|
|
|
|
int flip = GETARG_k(i);
|
|
|
|
StkId result = RA(pi);
|
|
|
|
Protect(luaT_trybiniTM(L, s2v(ra), imm, flip, result, tm));
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_MMBINK) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
Instruction pi = *(pc - 2); /* original arith. expression */
|
|
|
|
TValue *imm = KB(i);
|
|
|
|
TMS tm = (TMS)GETARG_C(i);
|
|
|
|
int flip = GETARG_k(i);
|
|
|
|
StkId result = RA(pi);
|
|
|
|
Protect(luaT_trybinassocTM(L, s2v(ra), imm, flip, result, tm));
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_UNM) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
TValue *rb = vRB(i);
|
|
|
|
lua_Number nb;
|
|
|
|
if (ttisinteger(rb)) {
|
|
|
|
lua_Integer ib = ivalue(rb);
|
|
|
|
setivalue(s2v(ra), intop(-, 0, ib));
|
|
|
|
}
|
|
|
|
else if (tonumberns(rb, nb)) {
|
|
|
|
setfltvalue(s2v(ra), luai_numunm(L, nb));
|
|
|
|
}
|
|
|
|
else
|
|
|
|
Protect(luaT_trybinTM(L, rb, rb, ra, TM_UNM));
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_BNOT) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
TValue *rb = vRB(i);
|
|
|
|
lua_Integer ib;
|
|
|
|
if (tointegerns(rb, &ib)) {
|
|
|
|
setivalue(s2v(ra), intop(^, ~l_castS2U(0), ib));
|
|
|
|
}
|
|
|
|
else
|
|
|
|
Protect(luaT_trybinTM(L, rb, rb, ra, TM_BNOT));
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_NOT) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
TValue *rb = vRB(i);
|
|
|
|
if (l_isfalse(rb))
|
|
|
|
setbtvalue(s2v(ra));
|
|
|
|
else
|
|
|
|
setbfvalue(s2v(ra));
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_LEN) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
Protect(luaV_objlen(L, ra, vRB(i)));
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_CONCAT) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
int n = GETARG_B(i); /* number of elements to concatenate */
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
L->top.p = ra + n; /* mark the end of concat operands */
|
2021-03-02 13:51:10 +00:00
|
|
|
ProtectNT(luaV_concat(L, n));
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
checkGC(L, L->top.p); /* 'luaV_concat' ensures correct top */
|
2021-03-02 13:51:10 +00:00
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_CLOSE) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
Protect(luaF_close(L, ra, LUA_OK, 1));
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_TBC) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
/* create new to-be-closed upvalue */
|
|
|
|
halfProtect(luaF_newtbcupval(L, ra));
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_JMP) {
|
|
|
|
dojump(ci, i, 0);
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_EQ) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
int cond;
|
|
|
|
TValue *rb = vRB(i);
|
|
|
|
Protect(cond = luaV_equalobj(L, s2v(ra), rb));
|
|
|
|
docondjump();
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_LT) {
|
|
|
|
op_order(L, l_lti, LTnum, lessthanothers);
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_LE) {
|
|
|
|
op_order(L, l_lei, LEnum, lessequalothers);
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_EQK) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
TValue *rb = KB(i);
|
|
|
|
/* basic types do not use '__eq'; we can use raw equality */
|
|
|
|
int cond = luaV_rawequalobj(s2v(ra), rb);
|
|
|
|
docondjump();
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_EQI) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
int cond;
|
|
|
|
int im = GETARG_sB(i);
|
|
|
|
if (ttisinteger(s2v(ra)))
|
|
|
|
cond = (ivalue(s2v(ra)) == im);
|
|
|
|
else if (ttisfloat(s2v(ra)))
|
|
|
|
cond = luai_numeq(fltvalue(s2v(ra)), cast_num(im));
|
|
|
|
else
|
|
|
|
cond = 0; /* other types cannot be equal to a number */
|
|
|
|
docondjump();
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_LTI) {
|
|
|
|
op_orderI(L, l_lti, luai_numlt, 0, TM_LT);
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_LEI) {
|
|
|
|
op_orderI(L, l_lei, luai_numle, 0, TM_LE);
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_GTI) {
|
|
|
|
op_orderI(L, l_gti, luai_numgt, 1, TM_LT);
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_GEI) {
|
|
|
|
op_orderI(L, l_gei, luai_numge, 1, TM_LE);
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_TEST) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
int cond = !l_isfalse(s2v(ra));
|
|
|
|
docondjump();
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_TESTSET) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
TValue *rb = vRB(i);
|
|
|
|
if (l_isfalse(rb) == GETARG_k(i))
|
|
|
|
pc++;
|
|
|
|
else {
|
|
|
|
setobj2s(L, ra, rb);
|
|
|
|
donextjump(ci);
|
|
|
|
}
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_CALL) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
CallInfo *newci;
|
|
|
|
int b = GETARG_B(i);
|
|
|
|
int nresults = GETARG_C(i) - 1;
|
|
|
|
if (b != 0) /* fixed number of arguments? */
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
L->top.p = ra + b; /* top signals number of arguments */
|
2021-03-02 13:51:10 +00:00
|
|
|
/* else previous instruction set top */
|
|
|
|
savepc(L); /* in case of errors */
|
|
|
|
if ((newci = luaD_precall(L, ra, nresults)) == NULL)
|
|
|
|
updatetrap(ci); /* C call; nothing else to be done */
|
|
|
|
else { /* Lua call: run function in this same C frame */
|
|
|
|
ci = newci;
|
|
|
|
goto startfunc;
|
|
|
|
}
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_TAILCALL) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
int b = GETARG_B(i); /* number of arguments + 1 (function) */
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
int n; /* number of results when calling a C function */
|
2021-03-02 13:51:10 +00:00
|
|
|
int nparams1 = GETARG_C(i);
|
|
|
|
/* delta is virtual 'func' - real 'func' (vararg functions) */
|
|
|
|
int delta = (nparams1) ? ci->u.l.nextraargs + nparams1 : 0;
|
|
|
|
if (b != 0)
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
L->top.p = ra + b;
|
2021-03-02 13:51:10 +00:00
|
|
|
else /* previous instruction set top */
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
b = cast_int(L->top.p - ra);
|
2021-03-02 13:51:10 +00:00
|
|
|
savepc(ci); /* several calls here can raise errors */
|
|
|
|
if (TESTARG_k(i)) {
|
|
|
|
luaF_closeupval(L, base); /* close upvalues from current call */
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
lua_assert(L->tbclist.p < base); /* no pending tbc variables */
|
|
|
|
lua_assert(base == ci->func.p + 1);
|
2021-03-02 13:51:10 +00:00
|
|
|
}
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
if ((n = luaD_pretailcall(L, ci, ra, b, delta)) < 0) /* Lua function? */
|
|
|
|
goto startfunc; /* execute the callee */
|
|
|
|
else { /* C function? */
|
|
|
|
ci->func.p -= delta; /* restore 'func' (if vararg) */
|
|
|
|
luaD_poscall(L, ci, n); /* finish caller */
|
2021-03-02 13:51:10 +00:00
|
|
|
updatetrap(ci); /* 'luaD_poscall' can change hooks */
|
|
|
|
goto ret; /* caller returns after the tail call */
|
|
|
|
}
|
|
|
|
}
|
|
|
|
vmcase(OP_RETURN) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
int n = GETARG_B(i) - 1; /* number of results */
|
|
|
|
int nparams1 = GETARG_C(i);
|
|
|
|
if (n < 0) /* not fixed? */
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
n = cast_int(L->top.p - ra); /* get what is available */
|
2021-03-02 13:51:10 +00:00
|
|
|
savepc(ci);
|
|
|
|
if (TESTARG_k(i)) { /* may there be open upvalues? */
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
ci->u2.nres = n; /* save number of returns */
|
|
|
|
if (L->top.p < ci->top.p)
|
|
|
|
L->top.p = ci->top.p;
|
2021-03-02 13:51:10 +00:00
|
|
|
luaF_close(L, base, CLOSEKTOP, 1);
|
|
|
|
updatetrap(ci);
|
|
|
|
updatestack(ci);
|
|
|
|
}
|
|
|
|
if (nparams1) /* vararg function? */
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
ci->func.p -= ci->u.l.nextraargs + nparams1;
|
|
|
|
L->top.p = ra + n; /* set call for 'luaD_poscall' */
|
2021-03-02 13:51:10 +00:00
|
|
|
luaD_poscall(L, ci, n);
|
|
|
|
updatetrap(ci); /* 'luaD_poscall' can change hooks */
|
|
|
|
goto ret;
|
|
|
|
}
|
|
|
|
vmcase(OP_RETURN0) {
|
2021-07-28 16:26:35 +00:00
|
|
|
if (l_unlikely(L->hookmask)) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
|
|
|
L->top.p = ra;
|
2021-03-02 13:51:10 +00:00
|
|
|
savepc(ci);
|
|
|
|
luaD_poscall(L, ci, 0); /* no hurry... */
|
|
|
|
trap = 1;
|
|
|
|
}
|
|
|
|
else { /* do the 'poscall' here */
|
2021-07-28 16:26:35 +00:00
|
|
|
int nres;
|
2021-03-02 13:51:10 +00:00
|
|
|
L->ci = ci->previous; /* back to caller */
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
L->top.p = base - 1;
|
2021-07-28 16:26:35 +00:00
|
|
|
for (nres = ci->nresults; l_unlikely(nres > 0); nres--)
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
setnilvalue(s2v(L->top.p++)); /* all results are nil */
|
2021-03-02 13:51:10 +00:00
|
|
|
}
|
|
|
|
goto ret;
|
|
|
|
}
|
|
|
|
vmcase(OP_RETURN1) {
|
2021-07-28 16:26:35 +00:00
|
|
|
if (l_unlikely(L->hookmask)) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
|
|
|
L->top.p = ra + 1;
|
2021-03-02 13:51:10 +00:00
|
|
|
savepc(ci);
|
|
|
|
luaD_poscall(L, ci, 1); /* no hurry... */
|
|
|
|
trap = 1;
|
|
|
|
}
|
|
|
|
else { /* do the 'poscall' here */
|
|
|
|
int nres = ci->nresults;
|
|
|
|
L->ci = ci->previous; /* back to caller */
|
|
|
|
if (nres == 0)
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
L->top.p = base - 1; /* asked for no results */
|
2021-03-02 13:51:10 +00:00
|
|
|
else {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
setobjs2s(L, base - 1, ra); /* at least this result */
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
L->top.p = base;
|
2021-07-28 16:26:35 +00:00
|
|
|
for (; l_unlikely(nres > 1); nres--)
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
setnilvalue(s2v(L->top.p++)); /* complete missing results */
|
2021-03-02 13:51:10 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
ret: /* return from a Lua function */
|
|
|
|
if (ci->callstatus & CIST_FRESH)
|
|
|
|
return; /* end this frame */
|
|
|
|
else {
|
|
|
|
ci = ci->previous;
|
|
|
|
goto returning; /* continue running caller in this frame */
|
|
|
|
}
|
|
|
|
}
|
|
|
|
vmcase(OP_FORLOOP) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
if (ttisinteger(s2v(ra + 2))) { /* integer loop? */
|
|
|
|
lua_Unsigned count = l_castS2U(ivalue(s2v(ra + 1)));
|
|
|
|
if (count > 0) { /* still more iterations? */
|
|
|
|
lua_Integer step = ivalue(s2v(ra + 2));
|
|
|
|
lua_Integer idx = ivalue(s2v(ra)); /* internal index */
|
|
|
|
chgivalue(s2v(ra + 1), count - 1); /* update counter */
|
|
|
|
idx = intop(+, idx, step); /* add step to index */
|
|
|
|
chgivalue(s2v(ra), idx); /* update internal index */
|
|
|
|
setivalue(s2v(ra + 3), idx); /* and control variable */
|
|
|
|
pc -= GETARG_Bx(i); /* jump back */
|
|
|
|
}
|
|
|
|
}
|
|
|
|
else if (floatforloop(ra)) /* float loop */
|
|
|
|
pc -= GETARG_Bx(i); /* jump back */
|
|
|
|
updatetrap(ci); /* allows a signal to break the loop */
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_FORPREP) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
savestate(L, ci); /* in case of errors */
|
|
|
|
if (forprep(L, ra))
|
|
|
|
pc += GETARG_Bx(i) + 1; /* skip the loop */
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_TFORPREP) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
/* create to-be-closed upvalue (if needed) */
|
|
|
|
halfProtect(luaF_newtbcupval(L, ra + 3));
|
|
|
|
pc += GETARG_Bx(i);
|
|
|
|
i = *(pc++); /* go to next instruction */
|
|
|
|
lua_assert(GET_OPCODE(i) == OP_TFORCALL && ra == RA(i));
|
|
|
|
goto l_tforcall;
|
|
|
|
}
|
|
|
|
vmcase(OP_TFORCALL) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
l_tforcall: {
|
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
/* 'ra' has the iterator function, 'ra + 1' has the state,
|
|
|
|
'ra + 2' has the control variable, and 'ra + 3' has the
|
|
|
|
to-be-closed variable. The call will use the stack after
|
|
|
|
these values (starting at 'ra + 4')
|
|
|
|
*/
|
|
|
|
/* push function, state, and control variable */
|
|
|
|
memcpy(ra + 4, ra, 3 * sizeof(*ra));
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
L->top.p = ra + 4 + 3;
|
2021-03-02 13:51:10 +00:00
|
|
|
ProtectNT(luaD_call(L, ra + 4, GETARG_C(i))); /* do the call */
|
|
|
|
updatestack(ci); /* stack may have changed */
|
|
|
|
i = *(pc++); /* go to next instruction */
|
|
|
|
lua_assert(GET_OPCODE(i) == OP_TFORLOOP && ra == RA(i));
|
|
|
|
goto l_tforloop;
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
}}
|
2021-03-02 13:51:10 +00:00
|
|
|
vmcase(OP_TFORLOOP) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
l_tforloop: {
|
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
if (!ttisnil(s2v(ra + 4))) { /* continue loop? */
|
|
|
|
setobjs2s(L, ra + 2, ra + 4); /* save control variable */
|
|
|
|
pc -= GETARG_Bx(i); /* jump back */
|
|
|
|
}
|
|
|
|
vmbreak;
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
}}
|
2021-03-02 13:51:10 +00:00
|
|
|
vmcase(OP_SETLIST) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
int n = GETARG_B(i);
|
|
|
|
unsigned int last = GETARG_C(i);
|
|
|
|
Table *h = hvalue(s2v(ra));
|
|
|
|
if (n == 0)
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
n = cast_int(L->top.p - ra) - 1; /* get up to the top */
|
2021-03-02 13:51:10 +00:00
|
|
|
else
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
L->top.p = ci->top.p; /* correct top in case of emergency GC */
|
2021-03-02 13:51:10 +00:00
|
|
|
last += n;
|
|
|
|
if (TESTARG_k(i)) {
|
|
|
|
last += GETARG_Ax(*pc) * (MAXARG_C + 1);
|
|
|
|
pc++;
|
|
|
|
}
|
|
|
|
if (last > luaH_realasize(h)) /* needs more space? */
|
|
|
|
luaH_resizearray(L, h, last); /* preallocate it at once */
|
|
|
|
for (; n > 0; n--) {
|
|
|
|
TValue *val = s2v(ra + n);
|
|
|
|
setobj2t(L, &h->array[last - 1], val);
|
|
|
|
last--;
|
|
|
|
luaC_barrierback(L, obj2gco(h), val);
|
|
|
|
}
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_CLOSURE) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
Proto *p = cl->p->p[GETARG_Bx(i)];
|
|
|
|
halfProtect(pushclosure(L, p, cl->upvals, base, ra));
|
|
|
|
checkGC(L, ra + 1);
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_VARARG) {
|
Bring Lua to 5.4.6. (#1214)
This essentially re-does the work of #875 on top of master.
This is what I did to check that Cosmo's Lua extensions still worked:
```
$ build/bootstrap/make MODE=aarch64 o/aarch64/third_party/lua/lua
$ ape o/aarch64/third_party/lua/lua
>: 10
10
>: 010
8
>: 0b10
2
>: string.byte("\e")
27
>: "Hello, %s" % {"world"}
Hello, world
>: "*" * 3
***
```
`luaL_traceback2` was used to show the stack trace with parameter
values; it's used in `LuaCallWithTrace`, which is used in Redbean to run
Lua code. You should be able to see the extended stack trace by running
something like this: `redbean -e "function a(b)c()end a(2)"` (with
"params" indicating the extended stack trace):
```
stack traceback:
[string "function a(b)c()end a(2)"]:1: in function 'a', params: b = 2;
[string "function a(b)c()end a(2)"]:1: in main chunk
```
@pkulchenko confirmed that I get the expected result with the updated
code.
This is what I did to check that Lua itself still worked:
```
$ cd third_party/lua/test/
$ ape ../../../o/aarch64/third_party/lua/lua all.lua
```
There's one test failure, in `files.lua`:
```
***** FILE 'files.lua'*****
testing i/o
../../../o/aarch64/third_party/lua/lua: files.lua:84: assertion failed!
stack traceback:
[C]: in function 'assert'
files.lua:84: in main chunk
(...tail calls...)
all.lua:195: in main chunk
[C]: in ?
.>>> closing state <<<
```
That isn't a result of these changes; the same test is failing in
master.
The failure is here:
```lua
if not _port then -- invalid seek
local status, msg, code = io.stdin:seek("set", 1000)
assert(not status and type(msg) == "string" and type(code) == "number")
end
```
The test expects a seek to offset 1,000 on stdin to fail — but it
doesn't. `status` ends up being the new offset rather than `nil`.
If I comment out that one test, the remaining tests succeed.
2024-06-16 00:13:08 +00:00
|
|
|
StkId ra = RA(i);
|
2021-03-02 13:51:10 +00:00
|
|
|
int n = GETARG_C(i) - 1; /* required results */
|
|
|
|
Protect(luaT_getvarargs(L, ci, ra, n));
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_VARARGPREP) {
|
|
|
|
ProtectNT(luaT_adjustvarargs(L, GETARG_A(i), ci, cl->p));
|
2021-07-28 16:26:35 +00:00
|
|
|
if (l_unlikely(trap)) { /* previous "Protect" updated trap */
|
2021-03-02 13:51:10 +00:00
|
|
|
luaD_hookcall(L, ci);
|
|
|
|
L->oldpc = 1; /* next opcode will be seen as a "new" line */
|
|
|
|
}
|
|
|
|
updatebase(ci); /* function has new base after adjustment */
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
vmcase(OP_EXTRAARG) {
|
|
|
|
lua_assert(0);
|
|
|
|
vmbreak;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* }================================================================== */
|