mirrors/cosmopolitan
1
0
Fork 0
mirror of https://github.com/jart/cosmopolitan.git synced 2025-01-31 11:37:35 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions
cosmopolitan/build/rules.mk

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

206 lines
7.2 KiB
Makefile
Raw Normal View History

Initial import
2020-06-15 14:18:57 +00:00
#-*-mode:makefile-gmake;indent-tabs-mode:t;tab-width:8;coding:utf-8-*-┐
#── vi: set et ft=make ts=8 sw=8 fenc=utf-8 :vi ──────────────────────┘
#
# SYNOPSIS
#
# Cosmopolitan Core Build Rules
#
# DESCRIPTION
#
# This file shows GNU Make how to convert between file types, based on
# their extensions. We use everyday, ordinary, boring, and ubiquitous
# system commands for everything, e.g. as, cc, ld, etc. with plain and
# simple shell-script wrappers, e.g. build/assemble and build/compile.
# Those veneers abstract away most of the boring trivialities, to keep
# our makefiles pristine and readable.
MAKEFLAGS += --no-builtin-rules
Rewrite .zip.o file linker This change takes an entirely new approach to the incremental linking of pkzip executables. The assets created by zipobj.com are now treated like debug data. After a .com.dbg is compiled, fixupobj.com should be run, so it can apply fixups to the offsets and move the zip directory to the end of the file. Since debug data doesn't get objcopy'd, a new tool has been introduced called zipcopy.com which should be run after objcopy whenever a .com file is created. This is all automated by the `cosmocc` toolchain which is rapidly becoming the new recommended approach. This change also introduces the new C23 checked arithmetic macros.
2023-06-10 16:15:19 +00:00
MAKE_ZIPCOPY = $(COMPILE) -AZIPCOPY -wT$@ $(ZIPCOPY) $< $@
Introduce native support for MacOS ARM64 There's a new program named ape/ape-m1.c which will be used to build an embeddable binary that can load ape and elf executables. The support is mostly working so far, but still chasing down ABI issues.
2023-05-19 02:05:08 +00:00
ifneq ($(ARCH), aarch64)
Fix the build
2023-08-11 16:49:39 +00:00
MAKE_OBJCOPY = $(COMPILE) -AOBJCOPY -T$@ $(OBJCOPY) -S -O binary $< $@ && $(MAKE_ZIPCOPY)
Introduce native support for MacOS ARM64 There's a new program named ape/ape-m1.c which will be used to build an embeddable binary that can load ape and elf executables. The support is mostly working so far, but still chasing down ABI issues.
2023-05-19 02:05:08 +00:00
else
Rewrite .zip.o file linker This change takes an entirely new approach to the incremental linking of pkzip executables. The assets created by zipobj.com are now treated like debug data. After a .com.dbg is compiled, fixupobj.com should be run, so it can apply fixups to the offsets and move the zip directory to the end of the file. Since debug data doesn't get objcopy'd, a new tool has been introduced called zipcopy.com which should be run after objcopy whenever a .com file is created. This is all automated by the `cosmocc` toolchain which is rapidly becoming the new recommended approach. This change also introduces the new C23 checked arithmetic macros.
2023-06-10 16:15:19 +00:00
MAKE_OBJCOPY = $(COMPILE) -AOBJCOPY -T$@ $(OBJCOPY) -S $< $@ && $(MAKE_ZIPCOPY)
Introduce native support for MacOS ARM64 There's a new program named ape/ape-m1.c which will be used to build an embeddable binary that can load ape and elf executables. The support is mostly working so far, but still chasing down ABI issues.
2023-05-19 02:05:08 +00:00
endif
Improve posix_spawn() some more
2023-09-12 15:58:57 +00:00
o/$(MODE)/%: o/$(MODE)/%.dbg
@$(MAKE_OBJCOPY)
Prove that Makefile is fully defined The whole repository is now buildable with GNU Make Landlock sandboxing. This proves that no Makefile targets exist which touch files other than their declared prerequisites. In order to do this, we had to: 1. Stop code morphing GCC output in package.com and instead run a newly introduced FIXUPOBJ.COM command after GCC invocations. 2. Disable all the crumby Python unit tests that do things like create files in the current directory, or rename() files between folders. This ended up being a lot of tests, but most of them are still ok. 3. Introduce an .UNSANDBOXED variable to GNU Make to disable Landlock. We currently only do this for things like `make tags`. 4. This change deletes some GNU Make code that was preventing the execve() optimization from working. This means it should no longer be necessary in most cases for command invocations to be indirected through the cocmd interpreter. 5. Missing dependencies had to be declared in certain places, in cases where they couldn't be automatically determined by MKDEPS.COM 6. The libcxx header situation has finally been tamed. One of the things that makes this difficult is MKDEPS.COM only wants to consider the first 64kb of a file, in order to go fast. But libcxx likes to have #include lines buried after huge documentation. 7. An .UNVEIL variable has been introduced to GNU Make just in case we ever wish to explicitly specify additional things that need to be whitelisted which aren't strictly prerequisites. This works in a manner similar to the recently introduced .EXTRA_PREREQS feature. There's now a new build/bootstrap/make.com prebuilt binary available. It should no longer be possible to write invalid Makefile code.
2022-08-06 10:51:50 +00:00
o/$(MODE)/%.o: %.c
@$(COMPILE) -AOBJECTIFY.c $(OBJECTIFY.c) $(OUTPUT_OPTION) $<
Avoid creating temporary output files This change also removes the futimens() call on the Landlock Make output file workaround, since it caused problems with commands like fixupobj which modify-in-place. It turns out if a file is opened for writing and then no writes actually occur, then the modified time doesn't change.
2022-08-11 05:08:43 +00:00
@$(COMPILE) -AFIXUPOBJ -wT$@ $(FIXUPOBJ) $@
Prove that Makefile is fully defined The whole repository is now buildable with GNU Make Landlock sandboxing. This proves that no Makefile targets exist which touch files other than their declared prerequisites. In order to do this, we had to: 1. Stop code morphing GCC output in package.com and instead run a newly introduced FIXUPOBJ.COM command after GCC invocations. 2. Disable all the crumby Python unit tests that do things like create files in the current directory, or rename() files between folders. This ended up being a lot of tests, but most of them are still ok. 3. Introduce an .UNSANDBOXED variable to GNU Make to disable Landlock. We currently only do this for things like `make tags`. 4. This change deletes some GNU Make code that was preventing the execve() optimization from working. This means it should no longer be necessary in most cases for command invocations to be indirected through the cocmd interpreter. 5. Missing dependencies had to be declared in certain places, in cases where they couldn't be automatically determined by MKDEPS.COM 6. The libcxx header situation has finally been tamed. One of the things that makes this difficult is MKDEPS.COM only wants to consider the first 64kb of a file, in order to go fast. But libcxx likes to have #include lines buried after huge documentation. 7. An .UNVEIL variable has been introduced to GNU Make just in case we ever wish to explicitly specify additional things that need to be whitelisted which aren't strictly prerequisites. This works in a manner similar to the recently introduced .EXTRA_PREREQS feature. There's now a new build/bootstrap/make.com prebuilt binary available. It should no longer be possible to write invalid Makefile code.
2022-08-06 10:51:50 +00:00
o/$(MODE)/%.o: o/$(MODE)/%.c
@$(COMPILE) -AOBJECTIFY.c $(OBJECTIFY.c) $(OUTPUT_OPTION) $<
Avoid creating temporary output files This change also removes the futimens() call on the Landlock Make output file workaround, since it caused problems with commands like fixupobj which modify-in-place. It turns out if a file is opened for writing and then no writes actually occur, then the modified time doesn't change.
2022-08-11 05:08:43 +00:00
@$(COMPILE) -AFIXUPOBJ -wT$@ $(FIXUPOBJ) $@
Prove that Makefile is fully defined The whole repository is now buildable with GNU Make Landlock sandboxing. This proves that no Makefile targets exist which touch files other than their declared prerequisites. In order to do this, we had to: 1. Stop code morphing GCC output in package.com and instead run a newly introduced FIXUPOBJ.COM command after GCC invocations. 2. Disable all the crumby Python unit tests that do things like create files in the current directory, or rename() files between folders. This ended up being a lot of tests, but most of them are still ok. 3. Introduce an .UNSANDBOXED variable to GNU Make to disable Landlock. We currently only do this for things like `make tags`. 4. This change deletes some GNU Make code that was preventing the execve() optimization from working. This means it should no longer be necessary in most cases for command invocations to be indirected through the cocmd interpreter. 5. Missing dependencies had to be declared in certain places, in cases where they couldn't be automatically determined by MKDEPS.COM 6. The libcxx header situation has finally been tamed. One of the things that makes this difficult is MKDEPS.COM only wants to consider the first 64kb of a file, in order to go fast. But libcxx likes to have #include lines buried after huge documentation. 7. An .UNVEIL variable has been introduced to GNU Make just in case we ever wish to explicitly specify additional things that need to be whitelisted which aren't strictly prerequisites. This works in a manner similar to the recently introduced .EXTRA_PREREQS feature. There's now a new build/bootstrap/make.com prebuilt binary available. It should no longer be possible to write invalid Makefile code.
2022-08-06 10:51:50 +00:00
Improve posix_spawn() some more
2023-09-12 15:58:57 +00:00
o/$(MODE)/%.h: %.c
@$(COMPILE) -AAMALGAMATE $(PREPROCESS) $(OUTPUT_OPTION) -fdirectives-only -P $<
o/$(MODE)/%.o: %.cc
Prove that Makefile is fully defined The whole repository is now buildable with GNU Make Landlock sandboxing. This proves that no Makefile targets exist which touch files other than their declared prerequisites. In order to do this, we had to: 1. Stop code morphing GCC output in package.com and instead run a newly introduced FIXUPOBJ.COM command after GCC invocations. 2. Disable all the crumby Python unit tests that do things like create files in the current directory, or rename() files between folders. This ended up being a lot of tests, but most of them are still ok. 3. Introduce an .UNSANDBOXED variable to GNU Make to disable Landlock. We currently only do this for things like `make tags`. 4. This change deletes some GNU Make code that was preventing the execve() optimization from working. This means it should no longer be necessary in most cases for command invocations to be indirected through the cocmd interpreter. 5. Missing dependencies had to be declared in certain places, in cases where they couldn't be automatically determined by MKDEPS.COM 6. The libcxx header situation has finally been tamed. One of the things that makes this difficult is MKDEPS.COM only wants to consider the first 64kb of a file, in order to go fast. But libcxx likes to have #include lines buried after huge documentation. 7. An .UNVEIL variable has been introduced to GNU Make just in case we ever wish to explicitly specify additional things that need to be whitelisted which aren't strictly prerequisites. This works in a manner similar to the recently introduced .EXTRA_PREREQS feature. There's now a new build/bootstrap/make.com prebuilt binary available. It should no longer be possible to write invalid Makefile code.
2022-08-06 10:51:50 +00:00
@$(COMPILE) -AOBJECTIFY.cxx $(OBJECTIFY.cxx) $(OUTPUT_OPTION) $<
Make improvements - Introduce portable sched_getcpu() api - Support GCC's __target_clones__ feature - Make fma() go faster on x86 in default mode - Remove some asan checks from core libraries - WinMain() now ensures $HOME and $USER are defined
2024-02-01 11:39:46 +00:00
@$(COMPILE) -AFIXUPOBJ -wT$@ $(FIXUPOBJ) $@
Prove that Makefile is fully defined The whole repository is now buildable with GNU Make Landlock sandboxing. This proves that no Makefile targets exist which touch files other than their declared prerequisites. In order to do this, we had to: 1. Stop code morphing GCC output in package.com and instead run a newly introduced FIXUPOBJ.COM command after GCC invocations. 2. Disable all the crumby Python unit tests that do things like create files in the current directory, or rename() files between folders. This ended up being a lot of tests, but most of them are still ok. 3. Introduce an .UNSANDBOXED variable to GNU Make to disable Landlock. We currently only do this for things like `make tags`. 4. This change deletes some GNU Make code that was preventing the execve() optimization from working. This means it should no longer be necessary in most cases for command invocations to be indirected through the cocmd interpreter. 5. Missing dependencies had to be declared in certain places, in cases where they couldn't be automatically determined by MKDEPS.COM 6. The libcxx header situation has finally been tamed. One of the things that makes this difficult is MKDEPS.COM only wants to consider the first 64kb of a file, in order to go fast. But libcxx likes to have #include lines buried after huge documentation. 7. An .UNVEIL variable has been introduced to GNU Make just in case we ever wish to explicitly specify additional things that need to be whitelisted which aren't strictly prerequisites. This works in a manner similar to the recently introduced .EXTRA_PREREQS feature. There's now a new build/bootstrap/make.com prebuilt binary available. It should no longer be possible to write invalid Makefile code.
2022-08-06 10:51:50 +00:00
Add OpenMP support
2024-01-29 05:57:40 +00:00
o/$(MODE)/%.o: %.cpp
@$(COMPILE) -AOBJECTIFY.cxx $(OBJECTIFY.cxx) $(OUTPUT_OPTION) $<
Improve posix_spawn() some more
2023-09-12 15:58:57 +00:00
o/$(MODE)/%.lds: %.lds
@$(COMPILE) -APREPROCESS $(PREPROCESS.lds) $(OUTPUT_OPTION) $<
ifneq ($(ARCH), aarch64)
o/$(MODE)/%.o: %.S
@$(COMPILE) -AOBJECTIFY.S $(OBJECTIFY.S) $(OUTPUT_OPTION) $<
o/$(MODE)/%.o: o/$(MODE)/%.S
@$(COMPILE) -AOBJECTIFY.S $(OBJECTIFY.S) $(OUTPUT_OPTION) $<
else
o/$(MODE)/%.o: %.S libc/empty.s
@$(COMPILE) -AOBJECTIFY.S $(OBJECTIFY.s) $(OUTPUT_OPTION) libc/empty.s
o/$(MODE)/%.o: o/$(MODE)/%.S libc/empty.s
@$(COMPILE) -AOBJECTIFY.S $(OBJECTIFY.s) $(OUTPUT_OPTION) libc/empty.s
endif
Introduce native support for MacOS ARM64 There's a new program named ape/ape-m1.c which will be used to build an embeddable binary that can load ape and elf executables. The support is mostly working so far, but still chasing down ABI issues.
2023-05-19 02:05:08 +00:00
o/$(MODE)/%.o: %.greg.c
@$(COMPILE) -AOBJECTIFY.greg $(OBJECTIFY.greg.c) $(OUTPUT_OPTION) $<
@$(COMPILE) -AFIXUPOBJ -wT$@ $(FIXUPOBJ) $@
o/$(MODE)/%.greg.o: %.greg.c
@$(COMPILE) -AOBJECTIFY.greg $(OBJECTIFY.greg.c) $(OUTPUT_OPTION) $<
@$(COMPILE) -AFIXUPOBJ -wT$@ $(FIXUPOBJ) $@
Improve memory safety This commit makes numerous refinements to cosmopolitan memory handling. The default stack size has been reduced from 2mb to 128kb. A new macro is now provided so you can easily reconfigure the stack size to be any value you want. Work around the breaking change by adding to your main: STATIC_STACK_SIZE(0x00200000); // 2mb stack If you're not sure how much stack you need, then you can use: STATIC_YOINK("stack_usage_logging"); After which you can `sort -nr o/$MODE/stack.log`. Based on the unit test suite, nothing in the Cosmopolitan repository (except for Python) needs a stack size greater than 30kb. There are also new macros for detecting the size and address of the stack at runtime, e.g. GetStackAddr(). We also now support sigaltstack() so if you want to see nice looking crash reports whenever a stack overflow happens, you can put this in main(): ShowCrashReports(); Under `make MODE=dbg` and `make MODE=asan` the unit testing framework will now automatically print backtraces of memory allocations when things like memory leaks happen. Bugs are now fixed in ASAN global variable overrun detection. The memtrack and asan runtimes also handle edge cases now. The new tools helped to identify a few memory leaks, which are fixed by this change. This change should fix an issue reported in #288 with ARG_MAX limits. Fixing this doubled the performance of MKDEPS.COM and AR.COM yet again.
2021-10-14 00:27:13 +00:00
o/%.a:
Work around Landlock output inode in compile.com This change fixes Landlock Make so that only the output target file is unveiled, rather than unveiling the directory that contains it. This gives us a much stronger sandbox. It also helped identify problematic build code in our repo that should have been using o/tmp instead. Landlock isn't able to let us unveil files that don't exist. Even if they do, then once a file is deleted, the sandboxing for it goes away. This caused problems for Landlock Make because tools like GNU LD will repeatedly delete and recreate the output file. This change uses the compile.com wrapper to ensure on changes happen to the output inode. New binary available on https://justine.lol/make/ Fixes #528
2022-08-09 13:28:41 +00:00
$(file >$(TMPDIR)/$(subst /,_,$@),$^)
Avoid creating temporary output files This change also removes the futimens() call on the Landlock Make output file workaround, since it caused problems with commands like fixupobj which modify-in-place. It turns out if a file is opened for writing and then no writes actually occur, then the modified time doesn't change.
2022-08-11 05:08:43 +00:00
@$(COMPILE) -AARCHIVE -wT$@ $(AR) $(ARFLAGS) $@ @$(TMPDIR)/$(subst /,_,$@)
Improve memory safety This commit makes numerous refinements to cosmopolitan memory handling. The default stack size has been reduced from 2mb to 128kb. A new macro is now provided so you can easily reconfigure the stack size to be any value you want. Work around the breaking change by adding to your main: STATIC_STACK_SIZE(0x00200000); // 2mb stack If you're not sure how much stack you need, then you can use: STATIC_YOINK("stack_usage_logging"); After which you can `sort -nr o/$MODE/stack.log`. Based on the unit test suite, nothing in the Cosmopolitan repository (except for Python) needs a stack size greater than 30kb. There are also new macros for detecting the size and address of the stack at runtime, e.g. GetStackAddr(). We also now support sigaltstack() so if you want to see nice looking crash reports whenever a stack overflow happens, you can put this in main(): ShowCrashReports(); Under `make MODE=dbg` and `make MODE=asan` the unit testing framework will now automatically print backtraces of memory allocations when things like memory leaks happen. Bugs are now fixed in ASAN global variable overrun detection. The memtrack and asan runtimes also handle edge cases now. The new tools helped to identify a few memory leaks, which are fixed by this change. This change should fix an issue reported in #288 with ARG_MAX limits. Fixing this doubled the performance of MKDEPS.COM and AR.COM yet again.
2021-10-14 00:27:13 +00:00
Make build hermetic without shell scripts - Fix some minor issues in ar.com - Have execve() look for `ape` command - Rewrite NT paths using /c/ rather /??/c:/ - Replace broken GCC symlinks with .sym files - Rewrite $PATH environment variables on startup - Make $(APE_NO_MODIFY_SELF) the default bootloader - Add all build command dependencies to build/bootstrap - Get the repository mostly building from source on non-Linux
2022-05-25 18:31:08 +00:00
o/%.pkg:
Restore zip.com and .symtab files This change restores the .symtab symbol table files in our flagship programs (e.g. redbean.com, python.com) needed to show backtraces. This also rolls back earlier changes to zip.com w.r.t. temp directories since the right way to do it turned out to be the -b DIR flag. This change also improves the performance of zip.com. It turned out mmap() wasn't being used, because zip.com was assuming a 4096-byte granularity, but cosmo requires 65536. There was also a chance to speed up stdio scanning using the unlocked functions.
2022-08-21 07:17:20 +00:00
$(file >$(TMPSAFE).args,$(filter %.o,$^))
@$(COMPILE) -APACKAGE -wT$@ $(PKG) $(OUTPUT_OPTION) $(addprefix -d,$(filter %.pkg,$^)) @$(TMPSAFE).args
Make build hermetic without shell scripts - Fix some minor issues in ar.com - Have execve() look for `ape` command - Rewrite NT paths using /c/ rather /??/c:/ - Replace broken GCC symlinks with .sym files - Rewrite $PATH environment variables on startup - Make $(APE_NO_MODIFY_SELF) the default bootloader - Add all build command dependencies to build/bootstrap - Get the repository mostly building from source on non-Linux
2022-05-25 18:31:08 +00:00
o/$(MODE)/%.pkg:
Restore zip.com and .symtab files This change restores the .symtab symbol table files in our flagship programs (e.g. redbean.com, python.com) needed to show backtraces. This also rolls back earlier changes to zip.com w.r.t. temp directories since the right way to do it turned out to be the -b DIR flag. This change also improves the performance of zip.com. It turned out mmap() wasn't being used, because zip.com was assuming a 4096-byte granularity, but cosmo requires 65536. There was also a chance to speed up stdio scanning using the unlocked functions.
2022-08-21 07:17:20 +00:00
$(file >$(TMPSAFE).args,$(filter %.o,$^))
@$(COMPILE) -APACKAGE -wT$@ $(PKG) $(OUTPUT_OPTION) $(addprefix -d,$(filter %.pkg,$^)) @$(TMPSAFE).args
Make build hermetic without shell scripts - Fix some minor issues in ar.com - Have execve() look for `ape` command - Rewrite NT paths using /c/ rather /??/c:/ - Replace broken GCC symlinks with .sym files - Rewrite $PATH environment variables on startup - Make $(APE_NO_MODIFY_SELF) the default bootloader - Add all build command dependencies to build/bootstrap - Get the repository mostly building from source on non-Linux
2022-05-25 18:31:08 +00:00
Stop using .com extension in monorepo The WIN32 CreateProcess() function does not require an .exe or .com suffix in order to spawn an executable. Now that we have Cosmo bash we're no longer so dependent on the cmd.exe prompt.
2024-03-03 00:57:56 +00:00
o/$(MODE)/%.o: %.py o/$(MODE)/third_party/python/pyobj
@$(COMPILE) -wAPYOBJ o/$(MODE)/third_party/python/pyobj $(PYFLAGS) -o $@ $<
Introduce Python objectifier (#259)
2021-09-04 22:44:00 +00:00
Stop using .com extension in monorepo The WIN32 CreateProcess() function does not require an .exe or .com suffix in order to spawn an executable. Now that we have Cosmo bash we're no longer so dependent on the cmd.exe prompt.
2024-03-03 00:57:56 +00:00
o/$(MODE)/%.pyc: %.py o/$(MODE)/third_party/python/pycomp
@$(COMPILE) -wAPYCOMP o/$(MODE)/third_party/python/pycomp $(PYCFLAGS) -o $@ $<
Add Lua compiler Redbean Lua Server Pages may now be stored in the zip as byte code. This can improve performance, since redbean currently doesn't cache byte code but it might be more useful for anyone wanting to create a closed source redbean. The .lua extension should be used for byte code files. Lua will tell them apart based on a magic number at the start of the file. This change also improves some Lua error reporting conditions. See #97
2021-08-09 20:09:14 +00:00
Stop using .com extension in monorepo The WIN32 CreateProcess() function does not require an .exe or .com suffix in order to spawn an executable. Now that we have Cosmo bash we're no longer so dependent on the cmd.exe prompt.
2024-03-03 00:57:56 +00:00
o/$(MODE)/%.lua: %.lua o/$(MODE)/third_party/lua/luac
@$(COMPILE) -wALUAC o/$(MODE)/third_party/lua/luac -s -o $@ $<
Add Deflate() / Inflate() to redbean and fix bugs The Compress() and Uncompress() APIs were a mistake. The functions themselves work fine, but it's a design blemish and does superfluous work. Since they were only introduced in the last few weeks, they're now deprecated and references to them have been scrubbed from the website and other documentation. Please use the new APIs since the old APIs will be removed at some point in the future. This change introduces automated Lua unit tests for the Redbean APIs. There's a few functions that were broken which have now been fixed, e.g. Underlong() and Decimate().
2022-07-08 16:47:04 +00:00
Stop using .com extension in monorepo The WIN32 CreateProcess() function does not require an .exe or .com suffix in order to spawn an executable. Now that we have Cosmo bash we're no longer so dependent on the cmd.exe prompt.
2024-03-03 00:57:56 +00:00
o/$(MODE)/%.lua.runs: %.lua o/$(MODE)/tool/net/redbean
@$(COMPILE) -wALUA -tT$@ o/$(MODE)/tool/net/redbean $(LUAFLAGS) -i $<
Remove plenty of makefile misconfigurations
2022-07-21 16:16:38 +00:00
Avoid sandboxing directory prerequisites Landlock Make will no longer sandbox prerequisites that end with a trailing slash. This means you can use use directory prerequisites for detecting deleted files when using using globbing, without the effect of unveiling the entire directory. When you do want make to unveil directories, you can omit the trailing slash.
2022-08-19 17:00:41 +00:00
################################################################################
Create unit test framework for subprocesses
2022-09-04 01:01:38 +00:00
# LOCAL UNIT TESTS
#
# We always run unit tests as part of the normal `make` invocation. you
# may override the $(TESTARGS) variable to do things such as enable the
# benchmarking feature. For example:
#
# TESTARGS = -b
#
# May be specified in your ~/.cosmo.mk file. You can also use this to
# enable things like function tracing. For example:
#
Improve quality of raise(), abort(), and tkill() This change fixes a nasty bug where SIG_IGN and SIG_DFL weren't working as advertised on BSDs. This change also fixes the tkill() definition on MacOS so it maps to __pthread_kill().
2022-09-04 01:12:01 +00:00
# TESTARGS = ----ftrace
Create unit test framework for subprocesses
2022-09-04 01:01:38 +00:00
# .PLEDGE += prot_exec
#
# You could then run a command like:
#
Stop using .com extension in monorepo The WIN32 CreateProcess() function does not require an .exe or .com suffix in order to spawn an executable. Now that we have Cosmo bash we're no longer so dependent on the cmd.exe prompt.
2024-03-03 00:57:56 +00:00
# make -j8 o//test/libc/calls/openbsd_test.runs
Create unit test framework for subprocesses
2022-09-04 01:01:38 +00:00
#
# You need PROT_EXEC permission since ftrace morphs the binary. It's
Stop using .com extension in monorepo The WIN32 CreateProcess() function does not require an .exe or .com suffix in order to spawn an executable. Now that we have Cosmo bash we're no longer so dependent on the cmd.exe prompt.
2024-03-03 00:57:56 +00:00
# also worth mentioning that the pledge command can simulate what
Create unit test framework for subprocesses
2022-09-04 01:01:38 +00:00
# Landlock Make does:
#
Stop using .com extension in monorepo The WIN32 CreateProcess() function does not require an .exe or .com suffix in order to spawn an executable. Now that we have Cosmo bash we're no longer so dependent on the cmd.exe prompt.
2024-03-03 00:57:56 +00:00
# o//tool/build/pledge \
Create unit test framework for subprocesses
2022-09-04 01:01:38 +00:00
# -v. -p 'stdio rpath wpath cpath tty prot_exec' \
Stop using .com extension in monorepo The WIN32 CreateProcess() function does not require an .exe or .com suffix in order to spawn an executable. Now that we have Cosmo bash we're no longer so dependent on the cmd.exe prompt.
2024-03-03 00:57:56 +00:00
# o//test/libc/calls/openbsd_test \
Create unit test framework for subprocesses
2022-09-04 01:01:38 +00:00
# ----ftrace
#
# This is useful in the event a test binary should run by itself, but
# fails to run beneath Landlock Make. It's also useful sometimes to
# override the verbosity when running tests:
#
Stop using .com extension in monorepo The WIN32 CreateProcess() function does not require an .exe or .com suffix in order to spawn an executable. Now that we have Cosmo bash we're no longer so dependent on the cmd.exe prompt.
2024-03-03 00:57:56 +00:00
# make V=5 TESTARGS=-b o//test/libc/calls/openbsd_test.runs
Create unit test framework for subprocesses
2022-09-04 01:01:38 +00:00
#
# This way, if for some reason a test should fail but calls exit(0),
# then the stdout/stderr output, which would normally be suppressed,
# will actually be displayed.
Remove VM variable You need to use qemu-user in binfmt_misc. For non-Linux we'll update execve() to spawn under the appropriate blink or qemu when needed.
2023-09-12 08:27:30 +00:00
o/$(MODE)/%.runs: o/$(MODE)/%
@$(COMPILE) -ACHECK -wtT$@ $< $(TESTARGS)
Create unit test framework for subprocesses
2022-09-04 01:01:38 +00:00
################################################################################
# ELF ZIP FILES
Avoid sandboxing directory prerequisites Landlock Make will no longer sandbox prerequisites that end with a trailing slash. This means you can use use directory prerequisites for detecting deleted files when using using globbing, without the effect of unveiling the entire directory. When you do want make to unveil directories, you can omit the trailing slash.
2022-08-19 17:00:41 +00:00
#
Stop using .com extension in monorepo The WIN32 CreateProcess() function does not require an .exe or .com suffix in order to spawn an executable. Now that we have Cosmo bash we're no longer so dependent on the cmd.exe prompt.
2024-03-03 00:57:56 +00:00
# zipobj lets us do fast incremental linking of compressed data.
Avoid sandboxing directory prerequisites Landlock Make will no longer sandbox prerequisites that end with a trailing slash. This means you can use use directory prerequisites for detecting deleted files when using using globbing, without the effect of unveiling the entire directory. When you do want make to unveil directories, you can omit the trailing slash.
2022-08-19 17:00:41 +00:00
# it's nice because if we link a hundred binaries that use the time zone
# database, then that database only needs to be DEFLATE'd once.
o/%.zip.o: o/%
@$(COMPILE) -wAZIPOBJ $(ZIPOBJ) $(ZIPOBJ_FLAGS) $(OUTPUT_OPTION) $<
o/$(MODE)/%.zip.o: %
@$(COMPILE) -wAZIPOBJ $(ZIPOBJ) $(ZIPOBJ_FLAGS) $(OUTPUT_OPTION) $<
o/$(MODE)/%.zip.o: %
@$(COMPILE) -wAZIPOBJ $(ZIPOBJ) $(ZIPOBJ_FLAGS) $(OUTPUT_OPTION) $<
# an issue with sandboxing arises when creating directory entries in the
# zip file. we need the trailing slash (e.g. o//foo/.zip.o) but Landlock
# Make avoids sandboxing directory names that have a trailing slash (so
# they can be used to watch for deleted files, without creating overly
# broad unveiling). such rules need to be written more explicitly.
o/$(MODE)%/.zip.o: %
@$(COMPILE) -wAZIPOBJ $(ZIPOBJ) $(ZIPOBJ_FLAGS) $(OUTPUT_OPTION) $<
Fix Landlock Make so it can read pattern rule vars It turned out that specifying all SRCS and INCS as dependencies on the pattern rules for all headers, caused `make` memory usage to skyrocket from 40mb ot 160mb. This change also reduces the build graph another 4%.
2022-08-14 00:20:50 +00:00
################################################################################
Create unit test framework for subprocesses
2022-09-04 01:01:38 +00:00
# STRICT HEADER CHECKING
Avoid sandboxing directory prerequisites Landlock Make will no longer sandbox prerequisites that end with a trailing slash. This means you can use use directory prerequisites for detecting deleted files when using using globbing, without the effect of unveiling the entire directory. When you do want make to unveil directories, you can omit the trailing slash.
2022-08-19 17:00:41 +00:00
#
# these rules are unsandboxed since they're already a sandboxing test,
# and it would be too costly in terms of make latency to have every
# header file depend on $(HDRS) and $(INCS).
Fix Landlock Make so it can read pattern rule vars It turned out that specifying all SRCS and INCS as dependencies on the pattern rules for all headers, caused `make` memory usage to skyrocket from 40mb ot 160mb. This change also reduces the build graph another 4%.
2022-08-14 00:20:50 +00:00
Add *NSYNC to libc/isystem
2022-09-13 08:46:29 +00:00
o/%.h.ok: private .UNSANDBOXED = 1
Fix Landlock Make so it can read pattern rule vars It turned out that specifying all SRCS and INCS as dependencies on the pattern rules for all headers, caused `make` memory usage to skyrocket from 40mb ot 160mb. This change also reduces the build graph another 4%.
2022-08-14 00:20:50 +00:00
o/%.h.ok: %.h
@$(COMPILE) -ACHECK.h $(COMPILE.c) -xc -g0 -o $@ $<
Add *NSYNC to libc/isystem
2022-09-13 08:46:29 +00:00
o/$(MODE)/%.h.ok: private .UNSANDBOXED = 1
Fix Landlock Make so it can read pattern rule vars It turned out that specifying all SRCS and INCS as dependencies on the pattern rules for all headers, caused `make` memory usage to skyrocket from 40mb ot 160mb. This change also reduces the build graph another 4%.
2022-08-14 00:20:50 +00:00
o/$(MODE)/%.h.ok: %.h
@$(COMPILE) -ACHECK.h $(COMPILE.c) -xc -g0 -o $@ $<
Add *NSYNC to libc/isystem
2022-09-13 08:46:29 +00:00
o/$(MODE)/%.hh.ok: private .UNSANDBOXED = 1
Fix Landlock Make so it can read pattern rule vars It turned out that specifying all SRCS and INCS as dependencies on the pattern rules for all headers, caused `make` memory usage to skyrocket from 40mb ot 160mb. This change also reduces the build graph another 4%.
2022-08-14 00:20:50 +00:00
o/$(MODE)/%.hh.ok: %.hh
@$(COMPILE) -ACHECK.h $(COMPILE.cxx) -xc++ -g0 -o $@ $<
o/%.okk: .UNSANDBOXED = 1
o/%.okk: %
@$(COMPILE) -ACHECK.h $(COMPILE.cxx) -xc++ -g0 -o $@ $<
Add *NSYNC to libc/isystem
2022-09-13 08:46:29 +00:00
o/$(MODE)/%.okk: private .UNSANDBOXED = 1
Fix Landlock Make so it can read pattern rule vars It turned out that specifying all SRCS and INCS as dependencies on the pattern rules for all headers, caused `make` memory usage to skyrocket from 40mb ot 160mb. This change also reduces the build graph another 4%.
2022-08-14 00:20:50 +00:00
o/$(MODE)/%.okk: %
@$(COMPILE) -ACHECK.h $(COMPILE.cxx) -xc++ -g0 -o $@ $<
Restore zip.com and .symtab files This change restores the .symtab symbol table files in our flagship programs (e.g. redbean.com, python.com) needed to show backtraces. This also rolls back earlier changes to zip.com w.r.t. temp directories since the right way to do it turned out to be the -b DIR flag. This change also improves the performance of zip.com. It turned out mmap() wasn't being used, because zip.com was assuming a 4096-byte granularity, but cosmo requires 65536. There was also a chance to speed up stdio scanning using the unlocked functions.
2022-08-21 07:17:20 +00:00
Unsandbox assembly generation
2022-09-05 02:30:05 +00:00
################################################################################
# EMACS ASSEMBLY GENERATION
Add *NSYNC to libc/isystem
2022-09-13 08:46:29 +00:00
o/$(MODE)/%-gcc.asm: private .UNSANDBOXED = 1
Unsandbox assembly generation
2022-09-05 02:30:05 +00:00
o/$(MODE)/%-gcc.asm: %.c
@$(COMPILE) -AOBJECTIFY.c $(OBJECTIFY.c) -S -g0 $(OUTPUT_OPTION) $<
Add *NSYNC to libc/isystem
2022-09-13 08:46:29 +00:00
o/$(MODE)/%-gcc.asm: private .UNSANDBOXED = 1
Unsandbox assembly generation
2022-09-05 02:30:05 +00:00
o/$(MODE)/%-gcc.asm: %.cc
@$(COMPILE) -AOBJECTIFY.c $(OBJECTIFY.cxx) -S -g0 $(OUTPUT_OPTION) $<
Add *NSYNC to libc/isystem
2022-09-13 08:46:29 +00:00
o/$(MODE)/%-clang.asm: private .UNSANDBOXED = 1
Unsandbox assembly generation
2022-09-05 02:30:05 +00:00
o/$(MODE)/%-clang.asm: %.c
@$(COMPILE) -AOBJECTIFY.c $(OBJECTIFY.c) -S -g0 $(OUTPUT_OPTION) $<
Port a lot more code to AARCH64 - Introduce epoll_pwait() - Rewrite -ftrapv and ffs() libraries in C code - Use more FreeBSD code in math function library - Get significantly more tests passing on qemu-aarch64 - Fix many Musl long double functions that were broken on AARCH64
2023-05-14 16:32:15 +00:00
# TODO(jart): Make intrinsics support Clang.
Fix close(1) bug on Windows
2023-07-28 10:56:56 +00:00
o/$(MODE)/%-clang.asm: CC = $(CLANG)
Add *NSYNC to libc/isystem
2022-09-13 08:46:29 +00:00
o/$(MODE)/%-clang.asm: private .UNSANDBOXED = 1
Unsandbox assembly generation
2022-09-05 02:30:05 +00:00
o/$(MODE)/%-clang.asm: %.cc
@$(COMPILE) -AOBJECTIFY.c $(OBJECTIFY.cxx) -S -g0 $(OUTPUT_OPTION) $<
Reference in a new issue Copy permalink