Make minor improvements

third_party/chibicc/asm.c

@@ -20,6 +20,8 @@
 
 #define PRECIOUS 0b1111000000101000  // bx,bp,r12-r15
 
+StaticAsm *staticasms;
+
 static const char kGreg[4][16][5] = {
     {"al", "cl", "dl", "bl", "spl", "bpl", "sil", "dil", "r8b", "r9b", "r10b",
      "r11b", "r12b", "r13b", "r14b", "r15b"},
@@ -31,8 +33,6 @@ static const char kGreg[4][16][5] = {
      "r11", "r12", "r13", "r14", "r15"},
 };
 
-StaticAsm *staticasms;
-
 static void DecodeAsmConstraints(AsmOperand *op) {
   int i;
   char c;

third_party/chibicc/chibicc.c

@@ -590,7 +590,7 @@ static void run_linker(StringArray *inputs, char *output) {
   run_subprocess(arr.data);
 }
 
-int main(int argc, char **argv) {
+int chibicc(int argc, char **argv) {
   showcrashreports();
   atexit(cleanup);
   init_macros();

third_party/chibicc/chibicc.h

@@ -120,6 +120,10 @@ Token *tokenize_string_literal(Token *, Type *);
 bool consume(Token **, Token *, char *, size_t);
 bool equal(Token *, char *, size_t);
 void convert_pp_tokens(Token *);
+void remove_backslash_newline(char *);
+void canonicalize_newline(char *);
+char *read_file(char *);
+int read_escaped_char(char **, char *);
 
 #define UNREACHABLE()    error("internal error at %s:%d", __FILE__, __LINE__)
 #define EQUAL(T, S)      equal(T, S, strlen(S))
@@ -563,6 +567,8 @@ extern bool opt_sse4;
 extern bool opt_verbose;
 extern char *base_file;
 
+int chibicc(int, char **);
+
 //
 // alloc.c
 //

third_party/chibicc/chibicc.mk

@@ -20,10 +20,16 @@ PKGS += THIRD_PARTY_CHIBICC
 THIRD_PARTY_CHIBICC_ARTIFACTS += THIRD_PARTY_CHIBICC_A
 THIRD_PARTY_CHIBICC = $(THIRD_PARTY_CHIBICC_A_DEPS) $(THIRD_PARTY_CHIBICC_A)
 THIRD_PARTY_CHIBICC_A = o/$(MODE)/third_party/chibicc/chibicc.a
+THIRD_PARTY_CHIBICC2_A = o/$(MODE)/third_party/chibicc/chibicc2.a
 THIRD_PARTY_CHIBICC_A_FILES := $(wildcard third_party/chibicc/*)
 THIRD_PARTY_CHIBICC_A_HDRS = $(filter %.h,$(THIRD_PARTY_CHIBICC_A_FILES))
 THIRD_PARTY_CHIBICC_A_SRCS = $(filter %.c,$(THIRD_PARTY_CHIBICC_A_FILES))
 
+THIRD_PARTY_CHIBICC_DEFINES =						\
+	-DCRT=\"$(CRT)\"						\
+	-DAPE=\"o/$(MODE)/ape/ape.o\"					\
+	-DLDS=\"o/$(MODE)/ape/ape.lds\"
+
 THIRD_PARTY_CHIBICC_BINS =						\
 	o/$(MODE)/third_party/chibicc/chibicc.com.dbg			\
 	o/$(MODE)/third_party/chibicc/chibicc.com			\
@@ -33,9 +39,13 @@ THIRD_PARTY_CHIBICC_BINS =						\
 THIRD_PARTY_CHIBICC_A_OBJS =						\
 	$(THIRD_PARTY_CHIBICC_A_SRCS:%=o/$(MODE)/%.zip.o)		\
 	$(THIRD_PARTY_CHIBICC_A_SRCS:%.c=o/$(MODE)/%.o)
+THIRD_PARTY_CHIBICC2_A_OBJS =						\
+	$(THIRD_PARTY_CHIBICC_A_SRCS:%=o/$(MODE)/%.zip.o)		\
+	$(THIRD_PARTY_CHIBICC_A_SRCS:%.c=o/$(MODE)/%.chibicc.o)
 
 THIRD_PARTY_CHIBICC_A_CHECKS =						\
 	$(THIRD_PARTY_CHIBICC_A).pkg					\
+	$(THIRD_PARTY_CHIBICC2_A).pkg					\
 	$(THIRD_PARTY_CHIBICC_A_HDRS:%=o/$(MODE)/%.ok)
 
 THIRD_PARTY_CHIBICC_A_DIRECTDEPS =					\
@@ -53,7 +63,9 @@ THIRD_PARTY_CHIBICC_A_DIRECTDEPS =					\
 	LIBC_STUBS							\
 	LIBC_TIME							\
 	LIBC_UNICODE							\
+	LIBC_SYSV							\
 	LIBC_X								\
+	TOOL_BUILD_LIB							\
 	THIRD_PARTY_COMPILER_RT						\
 	THIRD_PARTY_DLMALLOC						\
 	THIRD_PARTY_GDTOA
@@ -65,39 +77,49 @@ $(THIRD_PARTY_CHIBICC_A):						\
 		third_party/chibicc/					\
 		$(THIRD_PARTY_CHIBICC_A).pkg				\
 		$(THIRD_PARTY_CHIBICC_A_OBJS)
-
 $(THIRD_PARTY_CHIBICC_A).pkg:						\
 		$(THIRD_PARTY_CHIBICC_A_OBJS)				\
 		$(foreach x,$(THIRD_PARTY_CHIBICC_A_DIRECTDEPS),$($(x)_A).pkg)
 
+$(THIRD_PARTY_CHIBICC2_A):						\
+		third_party/chibicc/					\
+		$(THIRD_PARTY_CHIBICC2_A).pkg				\
+		$(THIRD_PARTY_CHIBICC2_A_OBJS)
+$(THIRD_PARTY_CHIBICC2_A).pkg:						\
+		$(THIRD_PARTY_CHIBICC2_A_OBJS)				\
+		$(foreach x,$(THIRD_PARTY_CHIBICC_A_DIRECTDEPS),$($(x)_A).pkg)
+
 o/$(MODE)/third_party/chibicc/chibicc.com.dbg:				\
 		$(THIRD_PARTY_CHIBICC_A_DEPS)				\
 		$(THIRD_PARTY_CHIBICC_A)				\
 		$(APE)							\
 		$(CRT)							\
-		o/$(MODE)/third_party/chibicc/chibicc.o			\
+		o/$(MODE)/third_party/chibicc/main.o			\
+		$(THIRD_PARTY_CHIBICC_A).pkg
+	@$(APELINK)
+o/$(MODE)/third_party/chibicc/chibicc2.com.dbg:				\
+		$(THIRD_PARTY_CHIBICC_A_DEPS)				\
+		$(THIRD_PARTY_CHIBICC2_A)				\
+		$(APE)							\
+		$(CRT)							\
+		o/$(MODE)/third_party/chibicc/main.chibicc.o		\
+		$(THIRD_PARTY_CHIBICC2_A).pkg
+	@$(APELINK)
+
+o/$(MODE)/third_party/chibicc/as.com.dbg:				\
+		$(THIRD_PARTY_CHIBICC_A_DEPS)				\
+		$(THIRD_PARTY_CHIBICC_A)				\
+		$(APE)							\
+		$(CRT)							\
+		o/$(MODE)/third_party/chibicc/as.o			\
 		$(THIRD_PARTY_CHIBICC_A).pkg
 	@$(APELINK)
 
-o/$(MODE)/third_party/chibicc/chibicc2.com.dbg:				\
-		$(THIRD_PARTY_CHIBICC_A_DEPS)				\
-		$(THIRD_PARTY_CHIBICC_A_SRCS:%.c=o/$(MODE)/%.chibicc.o)	\
-		$(THIRD_PARTY_CHIBICC_A).pkg				\
-		$(CRT)							\
-		$(APE)
-	@$(APELINK)
-
 o/$(MODE)/third_party/chibicc/chibicc.o:				\
-		CPPFLAGS +=						\
-			-DCRT=\"$(CRT)\"				\
-			-DAPE=\"o/$(MODE)/ape/ape.o\"			\
-			-DLDS=\"o/$(MODE)/ape/ape.lds\"
+		CPPFLAGS += $(THIRD_PARTY_CHIBICC_DEFINES)
 
 o/$(MODE)/third_party/chibicc/chibicc.chibicc.o:			\
-		CHIBICC_FLAGS +=					\
-			-DCRT=\"$(CRT)\"				\
-			-DAPE=\"o/$(MODE)/ape/ape.o\"			\
-			-DLDS=\"o/$(MODE)/ape/ape.lds\"
+		CHIBICC_FLAGS += $(THIRD_PARTY_CHIBICC_DEFINES)
 
 o/$(MODE)/%.chibicc.o: %.c o/$(MODE)/third_party/chibicc/chibicc.com.dbg
 	@ACTION=CHIBICC TARGET=$@ build/do $(CHIBICC) $(CHIBICC_FLAGS) -c -o $@ $<

third_party/chibicc/codegen.c

@@ -1770,9 +1770,9 @@ void gen_expr(Node *node) {
           println("\tdiv\t%s", di);
         } else {
           if (node->lhs->ty->size == 8) {
-            emitlin("\tcqo");
+            emitlin("\tcqto");
           } else {
-            emitlin("\tcdq");
+            emitlin("\tcltd");
           }
           println("\tidiv\t%s", di);
         }

third_party/chibicc/main.c

@@ -0,0 +1,5 @@
+#include "third_party/chibicc/chibicc.h"
+
+int main(int argc, char **argv) {
+  return chibicc(argc, argv);
+}

third_party/chibicc/test/macro_test.c

@@ -5,6 +5,8 @@ char *main_filename1 = __FILE__;
 int main_line1 = __LINE__;
 #define LINE() __LINE__
 int main_line2 = LINE();
+#define add2 add2
+#define add6(a, b, c, d, e, f) add6(a, b, c, d, e, f)
 
 #
 

third_party/chibicc/test/test.mk

@@ -79,7 +79,6 @@ o/$(MODE)/third_party/chibicc/test/%.com.dbg:					\
 		$(THIRD_PARTY_CHIBICC_TEST_A)					\
 		o/$(MODE)/third_party/chibicc/test/%.chibicc.o			\
 		$(THIRD_PARTY_CHIBICC_TEST_A).pkg				\
-		$(LIBC_TESTMAIN)						\
 		$(CRT)								\
 		$(APE)
 	@$(APELINK)
@@ -89,7 +88,6 @@ o/$(MODE)/third_party/chibicc/test/%2.com.dbg:					\
 		$(THIRD_PARTY_CHIBICC_TEST2_A)					\
 		o/$(MODE)/third_party/chibicc/test/%.chibicc2.o			\
 		$(THIRD_PARTY_CHIBICC_TEST2_A).pkg				\
-		$(LIBC_TESTMAIN)						\
 		$(CRT)								\
 		$(APE)
 	@$(APELINK)

third_party/chibicc/tokenize.c

@@ -128,7 +128,7 @@ static int read_ident(char *start) {
   for (;;) {
     char *q;
     c = decode_utf8(&q, p);
-    if (!('a' <= c && c <= 'f') && !is_ident2(c)) {
+    if (!is_ident2(c)) {
       return p - start;
     }
     p = q;
@@ -142,13 +142,14 @@ static int from_hex(char c) {
 }
 
 // Read a punctuator token from p and returns its length.
-static int read_punct(char *p) {
-  static char *kw[] = {"<<=", ">>=", "...", "==", "!=", "<=", ">=", "->",
-                       "+=",  "-=",  "*=",  "/=", "++", "--", "%=", "&=",
-                       "|=",  "^=",  "&&",  "||", "<<", ">>", "##"};
+int read_punct(char *p) {
+  static char kw[][4] = {"<<=", ">>=", "...", "==", "!=", "<=", ">=", "->",
+                         "+=",  "-=",  "*=",  "/=", "++", "--", "%=", "&=",
+                         "|=",  "^=",  "&&",  "||", "<<", ">>", "##"};
   for (int i = 0; i < sizeof(kw) / sizeof(*kw); i++) {
-    if (startswith(p, kw[i])) {
-      return strlen(kw[i]);
+    for (int j = 0;;) {
+      if (p[j] != kw[i][j]) break;
+      if (!kw[i][++j]) return j;
     }
   }
   return ispunct(*p) ? 1 : 0;
@@ -181,7 +182,7 @@ static bool is_keyword(Token *tok) {
   return hashmap_get2(&map, tok->loc, tok->len);
 }
 
-static int read_escaped_char(char **new_pos, char *p) {
+int read_escaped_char(char **new_pos, char *p) {
   if ('0' <= *p && *p <= '7') {
     // Read an octal number.
     unsigned c = *p++ - '0';
@@ -592,7 +593,7 @@ Token *tokenize(File *file) {
 }
 
 // Returns the contents of a given file.
-static char *read_file(char *path) {
+char *read_file(char *path) {
   FILE *fp;
   if (strcmp(path, "-") == 0) {
     // By convention, read from stdin if a given filename is "-".
@@ -639,7 +640,7 @@ File *new_file(char *name, int file_no, char *contents) {
 }
 
 // Replaces \r or \r\n with \n.
-static void canonicalize_newline(char *p) {
+void canonicalize_newline(char *p) {
   int i = 0, j = 0;
   while (p[i]) {
     if (p[i] == '\r' && p[i + 1] == '\n') {
@@ -656,7 +657,7 @@ static void canonicalize_newline(char *p) {
 }
 
 // Removes backslashes followed by a newline.
-static void remove_backslash_newline(char *p) {
+void remove_backslash_newline(char *p) {
   int i = 0, j = 0;
   // We want to keep the number of newline characters so that
   // the logical line number matches the physical one.

third_party/chibicc/unicode.c

@@ -61,8 +61,11 @@ uint32_t decode_utf8(char **new_pos, char *p) {
 }
 
 static bool in_range(uint32_t *range, uint32_t c) {
-  for (int i = 0; range[i] != -1; i += 2)
-    if (range[i] <= c && c <= range[i + 1]) return true;
+  for (int i = 0; range[i] != -1; i += 2) {
+    if (range[i] <= c && c <= range[i + 1]) {
+      return true;
+    }
+  }
   return false;
 }
 
@@ -78,7 +81,6 @@ static bool in_range(uint32_t *range, uint32_t c) {
 // (U+3000, full-width space) are allowed because they are out of range.
 bool is_ident1(uint32_t c) {
   static uint32_t range[] = {
-      'a',     'z',     'A',     'Z',     '_',     '_',     '$',     '$',
       0x00A8,  0x00A8,  0x00AA,  0x00AA,  0x00AD,  0x00AD,  0x00AF,  0x00AF,
       0x00B2,  0x00B5,  0x00B7,  0x00BA,  0x00BC,  0x00BE,  0x00C0,  0x00D6,
       0x00D8,  0x00F6,  0x00F8,  0x00FF,  0x0100,  0x02FF,  0x0370,  0x167F,
@@ -93,17 +95,26 @@ bool is_ident1(uint32_t c) {
       0xA0000, 0xAFFFD, 0xB0000, 0xBFFFD, 0xC0000, 0xCFFFD, 0xD0000, 0xDFFFD,
       0xE0000, 0xEFFFD, -1,
   };
-  return in_range(range, c);
+  if (c < 128) {
+    return ('a' <= c && c <= 'z') || ('A' <= c && c <= 'Z') || c == '_' ||
+           c == '$';
+  } else {
+    return in_range(range, c);
+  }
 }
 
 // Returns true if a given character is acceptable as a non-first
 // character of an identifier.
 bool is_ident2(uint32_t c) {
   static uint32_t range[] = {
-      '0',    '9',    '$',    '$',    0x0300, 0x036F, 0x1DC0,
-      0x1DFF, 0x20D0, 0x20FF, 0xFE20, 0xFE2F, -1,
+      0x0300, 0x036F, 0x1DC0, 0x1DFF, 0x20D0, 0x20FF, 0xFE20, 0xFE2F, -1,
   };
-  return is_ident1(c) || in_range(range, c);
+  if (is_ident1(c)) return true;
+  if (c < 128) {
+    return '0' <= c && c <= '9';
+  } else {
+    return in_range(range, c);
+  }
 }
 
 // Returns the number of columns needed to display a given