cpplib.c (initialize_char_syntax): Move to cppinit.c.

1999-02-04 14:33 -0500  Zack Weinberg  <zack@rabi.phys.columbia.edu>
	* cpplib.c (initialize_char_syntax): Move to cppinit.c.
	(cpp_define): Remove redundant syntax checks.
	(make_assertion): Rename cpp_assert, remove redundant syntax
	checks, export.
	(cpp_options_init): Don't init things to zero twice.
	(cpp_expand_to_buffer): Use memcpy, not a char-by-char loop.
	(do_include): Kill excessively verbose import warning that
	snuck back in in the gcc2 merge.
	(convert_string): Removed.
	(do_line): Rewrite with simple last-name-used cache instead of
	private hashtable.
	(cpp_start_read): Call initialize_char_syntax here, not...
	(cpp_reader_init): ...here.
	(cpp_handle_options): Support the -std switch.
	* cpplib.h (cpp_buffer): Add last_nominal_fname member.
	(cpp_options): Add c9x flag.
	Declare all the is_* tables and trigraph table here, as const.
	Prototype cpp_assert and initialize_char_syntax.
	* cppinit.c: New file.
	* cppfiles.c (read_and_prescan): Optimize.
	* Makefile.in (LIBCPP_OBJS): Add cppinit.o.

From-SVN: r25024

gcc/Makefile.in

@@ -873,10 +873,10 @@ stamp-objlist: $(OBJS)
 # We call this executable `xgcc' rather than `gcc'
 # to avoid confusion if the current directory is in the path
 # and CC is `gcc'.  It is renamed to `gcc' when it is installed.
-xgcc$(exeext): gcc.o version.o intl.o prefix.o \
-   version.o $(LIBDEPS) $(EXTRA_GCC_OBJS)
+xgcc$(exeext): gcc.o version.o choose-temp.o intl.o pexecute.o prefix.o \
+   version.o mkstemp.o $(LIBDEPS) $(EXTRA_GCC_OBJS)
 	$(CC) $(ALL_CFLAGS) $(LDFLAGS) -o $@ gcc.o prefix.o version.o \
-	  $(EXTRA_GCC_OBJS) $(LIBS)
+	  choose-temp.o pexecute.o mkstemp.o $(EXTRA_GCC_OBJS) $(LIBS)
 
 # Dump a specs file to make -B./ read these specs over installed ones.
 specs: xgcc$(exeext)
@@ -1337,11 +1337,11 @@ graph.o: graph.c $(CONFIG_H) system.h toplev.h flags.h output.h $(RTL_H) \
     hard-reg-set.h $(BASIC_BLOCK_H)
 sbitmap.o: sbitmap.c $(CONFIG_H) system.h $(RTL_H) flags.h $(BASIC_BLOCK_H)
 
-collect2$(exeext): collect2.o tlink.o hash.o underscore.o \
-	version.o $(LIBDEPS)
+collect2$(exeext): collect2.o tlink.o hash.o cplus-dem.o underscore.o \
+	version.o choose-temp.o mkstemp.o $(LIBDEPS)
 
-COLLECT2_OBJS = collect2.o tlink.o hash.o \
- intl.o underscore.o version.o 
+COLLECT2_OBJS = collect2.o tlink.o hash.o choose-temp.o cplus-dem.o \
+ intl.o underscore.o version.o mkstemp.o 
 collect2 : $(COLLECT2_OBJS) $(LIBDEPS)
 # Don't try modifying collect2 (aka ld) in place--it might be linking this.
 	-rm -f collect2$(exeext)
@@ -1356,6 +1356,16 @@ collect2.o : collect2.c $(CONFIG_H) system.h gstab.h intl.h \
 tlink.o: tlink.c $(DEMANGLE_H) hash.h $(CONFIG_H) system.h toplev.h collect2.h
 hash.o: hash.c hash.h system.h toplev.h
 
+cplus-dem.o: $(srcdir)/../libiberty/cplus-dem.c $(DEMANGLE_H)
+	rm -f cplus-dem.c
+	$(LN_S) $(srcdir)/../libiberty/cplus-dem.c cplus-dem.c
+	$(CC) -c $(ALL_CFLAGS) $(ALL_CPPFLAGS) $(INCLUDES) cplus-dem.c
+
+pexecute.o: $(srcdir)/../libiberty/pexecute.c $(CONFIG_H) system.h
+	rm -f pexecute.c
+	$(LN_S) $(srcdir)/../libiberty/pexecute.c pexecute.c
+	$(CC) -c $(ALL_CFLAGS) $(ALL_CPPFLAGS) $(INCLUDES) pexecute.c
+
 vfprintf.o: $(srcdir)/../libiberty/vfprintf.c $(CONFIG_H) system.h
 	rm -f vfprintf.c
 	$(LN_S) $(srcdir)/../libiberty/vfprintf.c vfprintf.c
@@ -1423,6 +1433,16 @@ obstack.o: $(srcdir)/../libiberty/obstack.c $(CONFIG_H)
 	$(LN_S) $(srcdir)/../libiberty/obstack.c obstack.c
 	$(CC) -c $(ALL_CFLAGS) $(ALL_CPPFLAGS) $(INCLUDES) obstack.c
 
+choose-temp.o: $(srcdir)/../libiberty/choose-temp.c $(CONFIG_H) system.h
+	rm -f choose-temp.c
+	$(LN_S) $(srcdir)/../libiberty/choose-temp.c choose-temp.c
+	$(CC) -c $(ALL_CFLAGS) $(ALL_CPPFLAGS) $(INCLUDES) choose-temp.c
+
+mkstemp.o: $(srcdir)/../libiberty/mkstemp.c $(CONFIG_H) system.h
+	rm -f mkstemp.c
+	$(LN_S) $(srcdir)/../libiberty/mkstemp.c mkstemp.c
+	$(CC) -c $(ALL_CFLAGS) $(ALL_CPPFLAGS) $(INCLUDES) mkstemp.c
+
 prefix.o: prefix.c $(CONFIG_H) system.h Makefile prefix.h
 	$(CC) $(ALL_CFLAGS) $(ALL_CPPFLAGS) $(INCLUDES) \
 	-DPREFIX=\"$(prefix)\" \
@@ -1959,7 +1979,7 @@ cccp.o: cccp.c $(CONFIG_H) intl.h pcp.h version.c config.status system.h \
 	  -c `echo $(srcdir)/cccp.c | sed 's,^\./,,'`
 
 LIBCPP_OBJS =	cpplib.o cpphash.o cppalloc.o cpperror.o cppexp.o cppfiles.o \
-		cppulp.o prefix.o version.o mbchar.o @extra_cpp_objs@ intl.o
+		cppinit.o cppulp.o prefix.o version.o mbchar.o @extra_cpp_objs@
 
 # All the other archives built/used by this makefile are for targets.  This
 # one is strictly for the host.
@@ -1996,12 +2016,15 @@ cpphash.o: cpphash.c cpplib.h machmode.h cpphash.h $(CONFIG_H) system.h
 
 cppalloc.o: cppalloc.c $(CONFIG_H) cpplib.h machmode.h system.h
 
+cppinit.o: cppalloc.c $(CONFIG_H) cpplib.h machmode.h system.h
+
 # Note for the stamp targets, we run the program `true' instead of
 # having an empty command (nothing following the semicolon).
 
 proto: config.status protoize$(exeext) unprotoize$(exeext) SYSCALLS.c.X
 
-PROTO_OBJS = getpwd.o intl.o version.o 
+PROTO_OBJS = choose-temp.o getopt.o getopt1.o getpwd.o \
+  intl.o pexecute.o version.o mkstemp.o
 
 protoize$(exeext): protoize.o $(PROTO_OBJS) $(LIBDEPS)
 	$(CC) $(ALL_CFLAGS) $(LDFLAGS) -o $@ protoize.o $(PROTO_OBJS) $(LIBS)
@@ -2031,6 +2054,16 @@ unprotoize.o: unprotoize.c protoize.c $(srcdir)/../include/getopt.h \
 	  -DSTD_PROTO_DIR=\"$(libsubdir)\" \
 	  $(srcdir)/unprotoize.c
 
+getopt.o: $(srcdir)/../libiberty/getopt.c $(srcdir)/../include/getopt.h
+	rm -f getopt.c
+	$(LN_S) $(srcdir)/../libiberty/getopt.c getopt.c
+	$(CC) -c $(ALL_CFLAGS) $(ALL_CPPFLAGS) $(INCLUDES) getopt.c
+
+getopt1.o: $(srcdir)/../libiberty/getopt1.c $(srcdir)/../include/getopt.h
+	rm -f getopt1.c
+	$(LN_S) $(srcdir)/../libiberty/getopt1.c getopt1.c
+	$(CC) -c $(ALL_CFLAGS) $(ALL_CPPFLAGS) $(INCLUDES) getopt1.c
+
 # This info describes the target machine, so compile with GCC just built.
 SYSCALLS.c.X: $(srcdir)/sys-types.h $(srcdir)/sys-protos.h $(GCC_PASSES) \
    stmp-int-hdrs

gcc/cppfiles.c

@@ -814,12 +814,11 @@ read_and_prescan (pfile, fp, desc, len)
      int desc;
      size_t len;
 {
-
   U_CHAR *buf = (U_CHAR *) xmalloc (len);
   U_CHAR *ip, *op, *line_base;
   U_CHAR *ibase;
   unsigned int line;
-  int count, seen_eof;
+  int count;
   size_t offset;
   /* 4096 bytes of buffer proper, 2 to detect running off the end without
      address arithmetic all the time, and 2 for pushback in the case there's
@@ -832,35 +831,35 @@ read_and_prescan (pfile, fp, desc, len)
   line_base = buf;
   line = 1;
   ibase = intermed + 2;
-  seen_eof = 0;
 
   for (;;)
     {
     read_next:
+
       count = read (desc, intermed + 2, INTERMED_BUFFER_SIZE);
       if (count < 0)
 	goto error;
-      if (count == 0)
-	seen_eof = 1;
-      count += 2 - (ibase - intermed);
-      if (count == 0)
+      else if (count == 0)
 	break;
 
+      offset += count;
       ip = ibase;
-      ip[count] = ip[count+1] = '\0';
       ibase = intermed + 2;
-      offset += count;
+      ibase[count] = ibase[count+1] = '\0';
       
       if (offset > len)
 	{
-	  size_t delta_op = op - buf;
-	  size_t delta_line_base = line_base - buf;
+	  size_t delta_op;
+	  size_t delta_line_base;
 	  len *= 2;
 	  if (offset > len)
 	      /* len overflowed.
 		 This could happen if the file is larger than half the
 		 maximum address space of the machine. */
 	    goto too_big;
+
+	  delta_op = op - buf;
+	  delta_line_base = line_base - buf;
 	  buf = xrealloc (buf, len);
 	  op = buf + delta_op;
 	  line_base = buf + delta_line_base;
@@ -868,7 +867,7 @@ read_and_prescan (pfile, fp, desc, len)
 
       for (;;)
 	{
-	  U_CHAR c;
+	  unsigned int c;
 	  c = *ip++;
 	  switch (c)
 	    {
@@ -880,16 +879,14 @@ read_and_prescan (pfile, fp, desc, len)
 	      break;
 	      
 	    case '\0':
-	      if (seen_eof)
-		goto eof;
-	      else
 	      goto read_next;
 	    case '\r':
 	      if (*ip == '\n') ip++;
-	      else if (*ip == '\0' && !seen_eof)
+	      else if (*ip == '\0')
 		{
-		  *--ibase = '\r';
-		  break;
+		  --ibase;
+		  intermed[1] = '\r';
+		  goto read_next;
 		}
 	      *op++ = '\n';
 	      line++;
@@ -898,10 +895,11 @@ read_and_prescan (pfile, fp, desc, len)
 
 	    case '\n':
 	      if (*ip == '\r') ip++;
-	      else if (*ip == '\0' && !seen_eof)
+	      else if (*ip == '\0')
 		{
-		  *--ibase = '\n';
-		  break;
+		  --ibase;
+		  intermed[1] = '\n';
+		  goto read_next;
 		}
 	      *op++ = '\n';
 	      line++;
@@ -912,28 +910,30 @@ read_and_prescan (pfile, fp, desc, len)
 	      if (CPP_OPTIONS (pfile)->trigraphs
 		  || CPP_OPTIONS (pfile)->warn_trigraphs)
 		{
+		  unsigned int d;
 		  /* If we're at the end of the intermediate buffer,
 		     we have to shift the ?'s down to the start and
 		     come back next pass. */
-		  c = ip[0];
-		  if (c == '\0' && !seen_eof)
+		  d = ip[0];
+		  if (d == '\0')
 		    {
-		      *--ibase = '?';
-		      break;
+		      --ibase;
+		      intermed[1] = '?';
+		      goto read_next;
 		    }
-		  if (c != '?')
+		  if (d != '?')
 		    {
 		      *op++ = '?';
 		      break;
 		    }
-		  c = ip[1];
-		  if (c == '\0' && !seen_eof)
+		  d = ip[1];
+		  if (d == '\0')
 		    {
-		      *--ibase = '?';
-		      *--ibase = '?';
-		      break;
+		      ibase -= 2;
+		      intermed[0] = intermed[1] = '?';
+		      goto read_next;
 		    }
-		  if (!trigraph_table[c])
+		  if (!trigraph_table[d])
 		    {
 		      *op++ = '?';
 		      break;
@@ -941,33 +941,47 @@ read_and_prescan (pfile, fp, desc, len)
 
 		  if (CPP_OPTIONS (pfile)->warn_trigraphs)
 		    cpp_warning_with_line (pfile, line, op-line_base,
-					   "trigraph ??%c encountered", c);
+					   "trigraph ??%c encountered", d);
 		  if (CPP_OPTIONS (pfile)->trigraphs)
-		    {
-		      *op++ = trigraph_table[c];
-		      ip += 2;
-		      break;
-		    }
+		    *op++ = trigraph_table[d];
 		  else
 		    {
 		      *op++ = '?';
 		      *op++ = '?';
-		      *op++ = c;
-		      ip += 2;
+		      *op++ = d;
 		    }
+		  ip += 2;
 		}
 	      else
 		*op++ = c;
 	    }
 	}
     }
- eof:
 
-  if (op == buf)
+  if (offset == 0)
     return 0;
 
+  /* Deal with pushed-back chars at true EOF.
+     If two chars were pushed back, they must both be ?'s.
+     If one was, it might be ?, \r, or \n, and \r needs to
+     become \n.
+     We know we have space already. */
+  if (ibase == intermed)
+    {
+      *op++ = '?';
+      *op++ = '?';
+    }
+  else if (ibase == intermed + 1)
+    {
+      if (*ibase == '?')
+	*op++ = '?';
+      else
+	*op++ = '\n';
+    }
+
   if (op[-1] != '\n' || op[-2] == '\\')
     {
+      if (CPP_PEDANTIC (pfile))
 	cpp_pedwarn_with_line (pfile, line, op - line_base,
 			       "no newline at end of file");
       if (offset + 2 > len)
@@ -983,8 +997,7 @@ read_and_prescan (pfile, fp, desc, len)
       *op++ = '\n';
     }
 
-  buf = xrealloc (buf, op - buf);
-  fp->buf = buf;
+  fp->buf = (len - offset < 20) ? buf : xrealloc (buf, op - buf);
   return op - buf;
 
  too_big:

gcc/cppinit.c

+/* CPP Library.
+   Copyright (C) 1986, 87, 89, 92-98, 1999 Free Software Foundation, Inc.
+   Contributed by Per Bothner, 1994-95.
+   Based on CCCP program by Paul Rubin, June 1986
+   Adapted to ANSI C, Richard Stallman, Jan 1987
+
+This program is free software; you can redistribute it and/or modify it
+under the terms of the GNU General Public License as published by the
+Free Software Foundation; either version 2, or (at your option) any
+later version.
+
+This program is distributed in the hope that it will be useful,
+but WITHOUT ANY WARRANTY; without even the implied warranty of
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+GNU General Public License for more details.
+
+You should have received a copy of the GNU General Public License
+along with this program; if not, write to the Free Software
+Foundation, 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.  */
+
+/* This file will have more stuff in it eventually, but right now
+   we just have one hack: we move all the is_* table initialization
+   in here, and we can declare them const in cpplib.h, which improves
+   code a bit. */
+
+#include "config.h"
+#include "system.h"
+
+typedef unsigned char U_CHAR;
+
+/* table to tell if char can be part of a C identifier.  */
+U_CHAR is_idchar[256] = { 0 };
+/* table to tell if char can be first char of a c identifier.  */
+U_CHAR is_idstart[256] = { 0 };
+/* table to tell if c is horizontal space.  */
+U_CHAR is_hor_space[256] = { 0 };
+/* table to tell if c is horizontal or vertical space.  */
+U_CHAR is_space[256] = { 0 };
+/* Table to handle trigraph conversion, which occurs before all other
+   processing, everywhere in the file.  (This is necessary since one
+   of the trigraphs encodes backslash.)  Note it's off by default.
+
+	from	to	from	to	from	to
+	?? =	#	?? )	]	?? !	|
+	?? (	[	?? '	^	?? >	}
+	?? /	\	?? <	{	?? -	~
+
+   There is not a space between the ?? and the third char.  I put spaces
+   there to avoid warnings when compiling this file. */
+U_CHAR trigraph_table[256] = { 0 };
+
+/* Initialize syntactic classifications of characters. */
+void
+initialize_char_syntax (int dollar_in_ident)
+{
+  is_idstart['a'] = 1; is_idstart['b'] = 1; is_idstart['c'] = 1;
+  is_idstart['d'] = 1; is_idstart['e'] = 1; is_idstart['f'] = 1;
+  is_idstart['g'] = 1; is_idstart['h'] = 1; is_idstart['i'] = 1;
+  is_idstart['j'] = 1; is_idstart['k'] = 1; is_idstart['l'] = 1;
+  is_idstart['m'] = 1; is_idstart['n'] = 1; is_idstart['o'] = 1;
+  is_idstart['p'] = 1; is_idstart['q'] = 1; is_idstart['r'] = 1;
+  is_idstart['s'] = 1; is_idstart['t'] = 1; is_idstart['u'] = 1;
+  is_idstart['v'] = 1; is_idstart['w'] = 1; is_idstart['x'] = 1;
+  is_idstart['y'] = 1; is_idstart['z'] = 1;
+
+  is_idstart['A'] = 1; is_idstart['B'] = 1; is_idstart['C'] = 1;
+  is_idstart['D'] = 1; is_idstart['E'] = 1; is_idstart['F'] = 1;
+  is_idstart['G'] = 1; is_idstart['H'] = 1; is_idstart['I'] = 1;
+  is_idstart['J'] = 1; is_idstart['K'] = 1; is_idstart['L'] = 1;
+  is_idstart['M'] = 1; is_idstart['N'] = 1; is_idstart['O'] = 1;
+  is_idstart['P'] = 1; is_idstart['Q'] = 1; is_idstart['R'] = 1;
+  is_idstart['S'] = 1; is_idstart['T'] = 1; is_idstart['U'] = 1;
+  is_idstart['V'] = 1; is_idstart['W'] = 1; is_idstart['X'] = 1;
+  is_idstart['Y'] = 1; is_idstart['Z'] = 1;
+
+  is_idstart['_'] = 1;
+
+  is_idchar['a'] = 1; is_idchar['b'] = 1; is_idchar['c'] = 1;
+  is_idchar['d'] = 1; is_idchar['e'] = 1; is_idchar['f'] = 1;
+  is_idchar['g'] = 1; is_idchar['h'] = 1; is_idchar['i'] = 1;
+  is_idchar['j'] = 1; is_idchar['k'] = 1; is_idchar['l'] = 1;
+  is_idchar['m'] = 1; is_idchar['n'] = 1; is_idchar['o'] = 1;
+  is_idchar['p'] = 1;  is_idchar['q'] = 1; is_idchar['r'] = 1;
+  is_idchar['s'] = 1; is_idchar['t'] = 1;  is_idchar['u'] = 1;
+  is_idchar['v'] = 1; is_idchar['w'] = 1; is_idchar['x'] = 1;
+  is_idchar['y'] = 1; is_idchar['z'] = 1;
+
+  is_idchar['A'] = 1; is_idchar['B'] = 1; is_idchar['C'] = 1;
+  is_idchar['D'] = 1; is_idchar['E'] = 1; is_idchar['F'] = 1;
+  is_idchar['G'] = 1; is_idchar['H'] = 1; is_idchar['I'] = 1;
+  is_idchar['J'] = 1; is_idchar['K'] = 1; is_idchar['L'] = 1;
+  is_idchar['M'] = 1; is_idchar['N'] = 1; is_idchar['O'] = 1;
+  is_idchar['P'] = 1; is_idchar['Q'] = 1; is_idchar['R'] = 1;
+  is_idchar['S'] = 1; is_idchar['T'] = 1;  is_idchar['U'] = 1;
+  is_idchar['V'] = 1; is_idchar['W'] = 1; is_idchar['X'] = 1;
+  is_idchar['Y'] = 1; is_idchar['Z'] = 1;
+
+  is_idchar['1'] = 1; is_idchar['2'] = 1; is_idchar['3'] = 1;
+  is_idchar['4'] = 1; is_idchar['5'] = 1; is_idchar['6'] = 1;
+  is_idchar['7'] = 1; is_idchar['8'] = 1; is_idchar['9'] = 1;
+  is_idchar['0'] = 1;
+
+  is_idchar['_']  = 1;
+
+  /* These will be reset later if -$ is in effect. */
+  is_idchar['$']  = dollar_in_ident;
+  is_idstart['$'] = dollar_in_ident;
+
+  /* horizontal space table */
+  is_hor_space[' '] = 1;
+  is_hor_space['\t'] = 1;
+  is_hor_space['\v'] = 1;
+  is_hor_space['\f'] = 1;
+  is_hor_space['\r'] = 1;
+
+  is_space[' '] = 1;
+  is_space['\t'] = 1;
+  is_space['\v'] = 1;
+  is_space['\f'] = 1;
+  is_space['\n'] = 1;
+  is_space['\r'] = 1;
+
+  /* trigraph conversion */
+  trigraph_table['='] = '#';  trigraph_table[')'] = ']';
+  trigraph_table['!'] = '|';  trigraph_table['('] = '[';
+  trigraph_table['\''] = '^'; trigraph_table['>'] = '}';
+  trigraph_table['/'] = '\\'; trigraph_table['<'] = '{';
+  trigraph_table['-'] = '~';
+}

gcc/cpplib.c

@@ -127,10 +127,8 @@ struct cpp_pending {
 extern void cpp_hash_cleanup PARAMS ((cpp_reader *));
 
 static char *my_strerror		PROTO ((int));
-static void make_assertion		PROTO ((cpp_reader *, char *, U_CHAR *));
 static void path_include		PROTO ((cpp_reader *, char *));
 static void initialize_builtins		PROTO ((cpp_reader *));
-static void initialize_char_syntax	PROTO ((void));
 static void validate_else		PROTO ((cpp_reader *, char *));
 static int comp_def_part		PROTO ((int, U_CHAR *, int, U_CHAR *,
 						int, int));
@@ -280,75 +278,6 @@ static struct directive directive_table[] = {
   {  -1, 0, "", T_UNUSED }
 };
 
-/* table to tell if char can be part of a C identifier.  */
-U_CHAR is_idchar[256] = { 0 };
-/* table to tell if char can be first char of a c identifier.  */
-U_CHAR is_idstart[256] = { 0 };
-/* table to tell if c is horizontal space.  */
-U_CHAR is_hor_space[256] = { 0 };
-/* table to tell if c is horizontal or vertical space.  */
-U_CHAR is_space[256] = { 0 };
-/* Table to handle trigraph conversion, which occurs before all other
-   processing, everywhere in the file.  (This is necessary since one
-   of the trigraphs encodes backslash.)  Note it's off by default.
-
-	from	to	from	to	from	to
-	?? =	#	?? )	]	?? !	|
-	?? (	[	?? '	^	?? >	}
-	?? /	\	?? <	{	?? -	~
-
-   There is not a space between the ?? and the third char.  I put spaces
-   there to avoid warnings when compiling this file. */
-U_CHAR trigraph_table[256] = { 0 };
-
-/* Initialize syntactic classifications of characters. */
-static void
-initialize_char_syntax ()
-{
-  register int i;
-
-  /*
-   * Set up is_idchar and is_idstart tables.  These should be
-   * faster than saying (is_alpha (c) || c == '_'), etc.
-   * Set up these things before calling any routines tthat
-   * refer to them.
-   * XXX We should setlocale(LC_CTYPE, "C") here for safety.
-   */
-  for (i = 0; i < 256; i++)
-    {
-      is_idchar[i]  = ISALNUM (i);
-      is_idstart[i] = ISALPHA (i);
-    }
-
-  is_idchar['_']  = 1;
-  is_idstart['_'] = 1;
-
-  /* These will be reset later if -$ is in effect. */
-  is_idchar['$']  = 1;
-  is_idstart['$'] = 1;
-
-  /* horizontal space table */
-  is_hor_space[' '] = 1;
-  is_hor_space['\t'] = 1;
-  is_hor_space['\v'] = 1;
-  is_hor_space['\f'] = 1;
-  is_hor_space['\r'] = 1;
-
-  is_space[' '] = 1;
-  is_space['\t'] = 1;
-  is_space['\v'] = 1;
-  is_space['\f'] = 1;
-  is_space['\n'] = 1;
-  is_space['\r'] = 1;
-
-  /* trigraph conversion */
-  trigraph_table['='] = '#';  trigraph_table[')'] = ']';
-  trigraph_table['!'] = '|';  trigraph_table['('] = '[';
-  trigraph_table['\''] = '^'; trigraph_table['>'] = '}';
-  trigraph_table['/'] = '\\'; trigraph_table['<'] = '{';
-  trigraph_table['-'] = '~';
-}
-
 /* Place into PFILE a quoted string representing the string SRC.
    Caller must reserve enough space in pfile->token_buffer.  */
 
@@ -400,13 +329,10 @@ cpp_grow_buffer (pfile, n)
   CPP_SET_WRITTEN (pfile, old_written);
 }
 
-
-/*
- * process a given definition string, for initialization
- * If STR is just an identifier, define it with value 1.
- * If STR has anything after the identifier, then it should
- * be identifier=definition.
- */
+/* Process the string STR as if it appeared as the body of a #define
+   If STR is just an identifier, define it with value 1.
+   If STR has anything after the identifier, then it should
+   be identifier=definition. */
 
 void
 cpp_define (pfile, str)
@@ -414,101 +340,37 @@ cpp_define (pfile, str)
      U_CHAR *str;
 {
   U_CHAR *buf, *p;
+  size_t count;
 
-  buf = str;
-  p = str;
-  if (!is_idstart[*p])
-    {
-      cpp_error (pfile, "malformed option `-D %s'", str);
-      return;
-    }
-  while (is_idchar[*++p])
-    ;
-  if (*p == '(') {
-    while (is_idchar[*++p] || *p == ',' || is_hor_space[*p])
-      ;
-    if (*p++ != ')')
-      p = (U_CHAR *) str;			/* Error */
-  }
-  if (*p == 0)
-    {
-      buf = (U_CHAR *) alloca (p - buf + 4);
-      strcpy ((char *)buf, str);
-      strcat ((char *)buf, " 1");
-    }
-  else if (*p != '=')
-    {
-      cpp_error (pfile, "malformed option `-D %s'", str);
-      return;
-    }
-  else
-    {
-      U_CHAR *q;
   /* Copy the entire option so we can modify it.  */
-      buf = (U_CHAR *) alloca (2 * strlen (str) + 1);
-      strncpy (buf, str, p - str);
-      /* Change the = to a space.  */
-      buf[p - str] = ' ';
-      /* Scan for any backslash-newline and remove it.  */
-      p++;
-      q = &buf[p - str];
-      while (*p)
+  count = strlen (str) + 3;
+  buf = (U_CHAR *) alloca (count);
+  memcpy (buf, str, count - 2);
+  /* Change the first "=" in the string to a space.  If there is none,
+     tack " 1" on the end. */
+  p = strchr (buf, '=');
+  if (p)
     {
-      if (*p == '\\' && p[1] == '\n')
-	p += 2;
-      else
-	*q++ = *p++;
-    }
-    *q = 0;
+      *p = ' ';
+      count -= 2;
     }
+  else
+      strcpy (&buf[count-3], " 1");
   
-  if (cpp_push_buffer (pfile, buf, strlen (buf)) != NULL)
+  if (cpp_push_buffer (pfile, buf, count - 1) != NULL)
     {
       do_define (pfile, NULL);
       cpp_pop_buffer (pfile);
     }
 }
 
-/* Process the string STR as if it appeared as the body of a #assert.
-   OPTION is the option name for which STR was the argument.  */
-
-static void
-make_assertion (pfile, option, str)
+/* Process the string STR as if it appeared as the body of a #assert. */
+void
+cpp_assert (pfile, str)
      cpp_reader *pfile;
-     char *option;
      U_CHAR *str;
 {
-  U_CHAR *buf, *p, *q;
-
-  /* Copy the entire option so we can modify it.  */
-  buf = (U_CHAR *) alloca (strlen (str) + 1);
-  strcpy ((char *) buf, str);
-  /* Scan for any backslash-newline and remove it.  */
-  p = q = buf;
-  while (*p) {
-#if 0
-    if (*p == '\\' && p[1] == '\n')
-      p += 2;
-    else
-#endif
-      *q++ = *p++;
-  }
-  *q = 0;
-
-  p = buf;
-  if (!is_idstart[*p]) {
-    cpp_error (pfile, "malformed option `%s %s'", option, str);
-    return;
-  }
-  while (is_idchar[*++p])
-    ;
-  while (*p == ' ' || *p == '\t') p++;
-  if (! (*p == 0 || *p == '(')) {
-    cpp_error (pfile, "malformed option `%s %s'", option, str);
-    return;
-  }
-
-  if (cpp_push_buffer (pfile, buf, strlen (buf)) != NULL)
+  if (cpp_push_buffer (pfile, str, strlen (str)) != NULL)
     {
       do_assert (pfile, NULL);
       cpp_pop_buffer (pfile);
@@ -563,32 +425,10 @@ cpp_options_init (opts)
      cpp_options *opts;
 {
   bzero ((char *) opts, sizeof *opts);
-  opts->in_fname = NULL;
-  opts->out_fname = NULL;
 
   opts->dollars_in_ident = 1;
-  initialize_char_syntax ();
-
-  opts->no_line_commands = 0;
-  opts->trigraphs = 0;
-  opts->put_out_comments = 0;
-  opts->print_include_names = 0;
-  opts->dump_macros = dump_none;
-  opts->no_output = 0;
-  opts->remap = 0;
-  opts->cplusplus = 0;
   opts->cplusplus_comments = 1;
-
-  opts->verbose = 0;
-  opts->objc = 0;
-  opts->lang_asm = 0;
-  opts->for_lint = 0;
-  opts->chill = 0;
-  opts->pedantic_errors = 0;
-  opts->inhibit_warnings = 0;
-  opts->warn_comments = 0;
   opts->warn_import = 1;
-  opts->warnings_are_errors = 0;
 }
 
 enum cpp_token
@@ -1707,7 +1547,6 @@ cpp_expand_to_buffer (pfile, buf, length)
 #if 0
   cpp_buffer obuf;
 #endif
-  U_CHAR *limit = buf + length;
   U_CHAR *buf1;
 #if 0
   int odepth = indepth;
@@ -1719,13 +1558,7 @@ cpp_expand_to_buffer (pfile, buf, length)
   /* Set up the input on the input stack.  */
 
   buf1 = (U_CHAR *) alloca (length + 1);
-  {
-    register U_CHAR *p1 = buf;
-    register U_CHAR *p2 = buf1;
-
-    while (p1 != limit)
-      *p2++ = *p1++;
-  }
+  memcpy (buf1, buf, length);
   buf1[length] = 0;
 
   ip = cpp_push_buffer (pfile, buf1, length);
@@ -1739,11 +1572,6 @@ cpp_expand_to_buffer (pfile, buf, length)
   /* Scan the input, create the output.  */
   cpp_scan_buffer (pfile);
 
-#if 0
-  if (indepth != odepth)
-    abort ();
-#endif
-
   CPP_NUL_TERMINATE (pfile);
 }
 
@@ -2932,19 +2760,7 @@ do_include (pfile, keyword)
       && !CPP_BUFFER (pfile)->system_header_p && !pfile->import_warning)
     {
       pfile->import_warning = 1;
-      cpp_warning (pfile, "using `#import' is not recommended");
-      cpp_notice ("The fact that a certain header file need not be processed more than once\n\
-should be indicated in the header file, not where it is used.\n\
-The best way to do this is with a conditional of this form:\n\
-\n\
-  #ifndef _FOO_H_INCLUDED\n\
-  #define _FOO_H_INCLUDED\n\
-  ... <real contents of file> ...\n\
-  #endif /* Not _FOO_H_INCLUDED */\n\
-\n\
-Then users can use `#include' any number of times.\n\
-GNU C automatically avoids processing the file more than once\n\
-when it is equipped with such a conditional.\n");
+      cpp_warning (pfile, "`#import' is obsolete, use an #ifndef wrapper in the header file");
     }
 
   pfile->parsing_include_directive++;
@@ -3139,57 +2955,9 @@ when it is equipped with such a conditional.\n");
   return 0;
 }
 
-
-/* Convert a character string literal into a nul-terminated string.
-   The input string is [IN ... LIMIT).
-   The result is placed in RESULT.  RESULT can be the same as IN.
-   The value returned in the end of the string written to RESULT,
-   or NULL on error.  */
-
-static U_CHAR *
-convert_string (pfile, result, in, limit, handle_escapes)
-     cpp_reader *pfile;
-     register U_CHAR *result, *in, *limit;
-     int handle_escapes;
-{
-  U_CHAR c;
-  c = *in++;
-  if (c != '\"')
-    return NULL;
-  while (in < limit)
-    {
-      U_CHAR c = *in++;
-      switch (c)
-	{
-	case '\0':
-	  return NULL;
-	case '\"':
-	  limit = in;
-	  break;
-	case '\\':
-	  if (handle_escapes)
-	    {
-	      char *bpc = (char *) in;
-	      int i = (U_CHAR) cpp_parse_escape (pfile, &bpc, 0x00ff);
-	      in = (U_CHAR *) bpc;
-	      if (i >= 0)
-		*result++ = (U_CHAR)c;
-	      break;
-	    }
-	  /* else fall through */
-	default:
-	  *result++ = c;
-	}
-    }
-  *result = 0;
-  return result;
-}
-
-/*
- * interpret #line command.  Remembers previously seen fnames
- * in its very own hash table.
- */
-#define FNAME_HASHSIZE 37
+/* Interpret #line command.
+   Note that the filename string (if any) is treated as if it were an
+   include filename.  That means no escape handling.  */
 
 static int
 do_line (pfile, keyword)
@@ -3201,59 +2969,39 @@ do_line (pfile, keyword)
   long old_written = CPP_WRITTEN (pfile);
   enum file_change_code file_change = same_file;
   enum cpp_token token;
+  char *x;
 
   token = get_directive_token (pfile);
 
-  if (token != CPP_NUMBER
-      || !ISDIGIT(pfile->token_buffer[old_written]))
+  if (token != CPP_NUMBER)
     {
-      cpp_error (pfile, "invalid format `#line' command");
+      cpp_error (pfile, "token after `#line' is not an integer");
       goto bad_line_directive;
     }
 
-  /* The Newline at the end of this line remains to be processed.
-     To put the next line at the specified line number,
-     we must store a line number now that is one less.  */
-  new_lineno = atoi ((char *)(pfile->token_buffer + old_written)) - 1;
+  new_lineno = strtol (pfile->token_buffer + old_written, &x, 10);
+  if (x[0] != '\0')
+    {
+      cpp_error (pfile, "token after `#line' is not an integer");
+      goto bad_line_directive;
+    }      
   CPP_SET_WRITTEN (pfile, old_written);
 
-  /* NEW_LINENO is one less than the actual line number here.  */
-  if (CPP_PEDANTIC (pfile) && new_lineno < 0)
+  if (CPP_PEDANTIC (pfile) && new_lineno <= 0)
     cpp_pedwarn (pfile, "line number out of range in `#line' command");
 
-#if 0 /* #line 10"foo.c" is supposed to be allowed.  */
-  if (PEEKC() && !is_space[PEEKC()]) {
-    cpp_error (pfile, "invalid format `#line' command");
-    goto bad_line_directive;
-  }
-#endif
-
   token = get_directive_token (pfile);
 
-  if (token == CPP_STRING) {
-    U_CHAR *fname = pfile->token_buffer + old_written;
-    U_CHAR *end_name;
-    static HASHNODE *fname_table[FNAME_HASHSIZE];
-    HASHNODE *hp, **hash_bucket;
-    U_CHAR *p;
-    long num_start;
-    int fname_length;
-
-    /* Turn the file name, which is a character string literal,
-       into a null-terminated string.  Do this in place.  */
-    end_name = convert_string (pfile, fname, fname, CPP_PWRITTEN (pfile), 1);
-    if (end_name == NULL)
+  if (token == CPP_STRING)
     {
-	cpp_error (pfile, "invalid format `#line' command");
-	goto bad_line_directive;
-    }
-
-    fname_length = end_name - fname;
+      U_CHAR *fname = pfile->token_buffer + old_written + 1;
+      U_CHAR *end_name = CPP_PWRITTEN (pfile) - 1;
+      long num_start = CPP_WRITTEN (pfile);
 
-    num_start = CPP_WRITTEN (pfile);
       token = get_directive_token (pfile);
-    if (token != CPP_VSPACE && token != CPP_EOF && token != CPP_POP) {
-      p = pfile->token_buffer + num_start;
+      if (token != CPP_VSPACE && token != CPP_EOF && token != CPP_POP)
+	{
+	  U_CHAR *p = pfile->token_buffer + num_start;
 	  if (CPP_PEDANTIC (pfile))
 	    cpp_pedwarn (pfile, "garbage at end of `#line' command");
 
@@ -3274,53 +3022,68 @@ do_line (pfile, keyword)
 	  CPP_SET_WRITTEN (pfile, num_start);
 	  token = get_directive_token (pfile);
 	  p = pfile->token_buffer + num_start;
-      if (token == CPP_NUMBER && p[1] == '\0' && (*p == '3' || *p== '4')) {
+	  if (token == CPP_NUMBER && p[1] == '\0' && (*p == '3' || *p== '4'))
+	    {
 	      ip->system_header_p = *p == '3' ? 1 : 2;
 	      token = get_directive_token (pfile);
 	    }
-      if (token != CPP_VSPACE) {
+	  if (token != CPP_VSPACE)
+	    {
 	      cpp_error (pfile, "invalid format `#line' command");
 	      goto bad_line_directive;
 	    }
 	}
       
-    hash_bucket = &fname_table[hashf (fname, fname_length, FNAME_HASHSIZE)];
-    for (hp = *hash_bucket; hp != NULL; hp = hp->next)
-      if (hp->length == fname_length
-	  && strncmp (hp->value.cpval, fname, fname_length) == 0) {
-	ip->nominal_fname = hp->value.cpval;
-	break;
-      }
-    if (hp == 0) {
-      /* Didn't find it; cons up a new one.  */
-      hp = (HASHNODE *) xcalloc (1, sizeof (HASHNODE) + fname_length + 1);
-      hp->next = *hash_bucket;
-      *hash_bucket = hp;
+      *end_name = '\0';
+      
+      if (strcmp (fname, ip->nominal_fname))
+	{
+	  char *newname, *oldname;
+	  if (!strcmp (fname, ip->fname))
+	    newname = ip->fname;
+	  else if (ip->last_nominal_fname
+		   && !strcmp (fname, ip->last_nominal_fname))
+	    newname = ip->last_nominal_fname;
+	  else
+	    newname = xstrdup (fname);
 
-      hp->length = fname_length;
-      ip->nominal_fname = hp->value.cpval = ((char *) hp) + sizeof (HASHNODE);
-      bcopy (fname, hp->value.cpval, fname_length);
+	  oldname = ip->nominal_fname;
+	  ip->nominal_fname = newname;
+
+	  if (ip->last_nominal_fname
+	      && ip->last_nominal_fname != oldname
+	      && ip->last_nominal_fname != newname)
+	    free (ip->last_nominal_fname);
+
+	  if (newname == ip->fname)
+	    ip->last_nominal_fname = NULL;
+	  else
+	    ip->last_nominal_fname = oldname;
 	} 
     }
-  else if (token != CPP_VSPACE && token != CPP_EOF) {
-    cpp_error (pfile, "invalid format `#line' command");
+  else if (token != CPP_VSPACE && token != CPP_EOF)
+    {
+      cpp_error (pfile, "token after `#line %d' is not a string", new_lineno);
       goto bad_line_directive;
     }
 
-  ip->lineno = new_lineno;
+  /* The Newline at the end of this line remains to be processed.
+     To put the next line at the specified line number,
+     we must store a line number now that is one less.  */
+  ip->lineno = new_lineno - 1;
+  CPP_SET_WRITTEN (pfile, old_written);
+  output_line_command (pfile, 0, file_change);
+  return 0;
+
  bad_line_directive:
   skip_rest_of_line (pfile);
   CPP_SET_WRITTEN (pfile, old_written);
-  output_line_command (pfile, 0, file_change);
   return 0;
 }
 
-/*
- * remove the definition of a symbol from the symbol table.
- * according to un*x /lib/cpp, it is not an error to undef
- * something that has no definitions, so it isn't one here either.
- */
-
+/* Remove the definition of a symbol from the symbol table.
+   According to the C standard, it is not an error to undef
+   something that has no definitions. */
 static int
 do_undef (pfile, keyword)
      cpp_reader *pfile;
@@ -4711,22 +4474,21 @@ cpp_start_read (pfile, fname)
   cpp_buffer *fp;
   struct include_hash *ih_fake;
 
-  /* The code looks at the defaults through this pointer, rather than through
-     the constant structure above.  This pointer gets changed if an environment
-     variable specifies other defaults.  */
+  /* The code looks at the defaults through this pointer, rather than
+     through the constant structure above.  This pointer gets changed
+     if an environment variable specifies other defaults.  */
   struct default_include *include_defaults = include_defaults_array;
 
-  /* Now that we know dollars_in_ident for real,
-     reset is_idchar/is_idstart. */
-  is_idchar['$'] = opts->dollars_in_ident;
-  is_idstart['$'] = opts->dollars_in_ident;
+  /* Now that we know dollars_in_ident, we can initialize the syntax
+     tables. */
+  initialize_char_syntax (opts->dollars_in_ident);
   
   /* Add dirs from CPATH after dirs from -I.  */
   /* There seems to be confusion about what CPATH should do,
      so for the moment it is not documented.  */
-  /* Some people say that CPATH should replace the standard include dirs,
-     but that seems pointless: it comes before them, so it overrides them
-     anyway.  */
+  /* Some people say that CPATH should replace the standard include
+     dirs, but that seems pointless: it comes before them, so it
+     overrides them anyway.  */
   GET_ENV_PATH_LIST (p, "CPATH");
   if (p != 0 && ! opts->no_standard_includes)
     path_include (pfile, p);
@@ -4810,7 +4572,7 @@ cpp_start_read (pfile, fname)
 	save_char = *termination;
 	*termination = '\0';
 	/* Install the assertion.  */
-	make_assertion (pfile, "-A", assertion);
+	cpp_assert (pfile, assertion);
 	*termination = (char) save_char;
 	p = termination;
 	while (*p == ' ' || *p == '\t')
@@ -4844,7 +4606,7 @@ cpp_start_read (pfile, fname)
 	      cpp_define (pfile, pend->arg);
 	      break;
 	    case 'A':
-	      make_assertion (pfile, "-A", pend->arg);
+	      cpp_assert (pfile, pend->arg);
 	      break;
 	    }
 	}
@@ -5273,6 +5035,10 @@ print_help ()
   printf ("  -lang-objc++              Assume that the input sources are in ObjectiveC++\n");
   printf ("  -lang-asm                 Assume that the input sources are in assembler\n");
   printf ("  -lang-chill               Assume that the input sources are in Chill\n");
+  printf ("  -std=<std name>           Specify the conformance standard; one of:\n");
+  printf ("                            gnu89, gnu9x, c89, c9x, iso9899:1990,\n");
+  printf ("                            iso9899:199409, iso9899:199x\n");
+
   printf ("  -+                        Allow parsing of C++ style features\n");
   printf ("  -w                        Inhibit warning messages\n");
   printf ("  -Wtrigraphs               Warn if trigraphs are encountered\n");
@@ -5520,19 +5286,19 @@ cpp_handle_option (pfile, argc, argv)
     case 'l':
       if (! strcmp (argv[i], "-lang-c"))
 	opts->cplusplus = 0, opts->cplusplus_comments = 1, opts->c89 = 0,
-	  opts->objc = 0;
+	  opts->c9x = 1, opts->objc = 0;
       if (! strcmp (argv[i], "-lang-c89"))
 	opts->cplusplus = 0, opts->cplusplus_comments = 0, opts->c89 = 1,
-	  opts->objc = 0;
+	  opts->c9x = 0, opts->objc = 0;
       if (! strcmp (argv[i], "-lang-c++"))
 	opts->cplusplus = 1, opts->cplusplus_comments = 1, opts->c89 = 0,
-	  opts->objc = 0;
+	  opts->c9x = 0, opts->objc = 0;
       if (! strcmp (argv[i], "-lang-objc"))
 	opts->cplusplus = 0, opts->cplusplus_comments = 1, opts->c89 = 0,
-	  opts->objc = 1;
+	  opts->c9x = 0, opts->objc = 1;
       if (! strcmp (argv[i], "-lang-objc++"))
 	opts->cplusplus = 1, opts->cplusplus_comments = 1, opts->c89 = 0,
-	  opts->objc = 1;
+	  opts->c9x = 0, opts->objc = 1;
       if (! strcmp (argv[i], "-lang-asm"))
 	opts->lang_asm = 1;
       if (! strcmp (argv[i], "-lint"))
@@ -5546,6 +5312,20 @@ cpp_handle_option (pfile, argc, argv)
       opts->cplusplus = 1, opts->cplusplus_comments = 1;
       break;
 
+    case 's':
+      if (!strcmp (argv[i], "-std=iso9899:1990")
+	  || !strcmp (argv[i], "-std=iso9899:199409")
+	  || !strcmp (argv[i], "-std=c89")
+	  || !strcmp (argv[i], "-std=gnu89"))
+	  opts->cplusplus = 0, opts->cplusplus_comments = 0,
+	    opts->c89 = 1, opts->c9x = 0, opts->objc = 0;
+      else if (!strcmp (argv[i], "-std=iso9899:199x")
+	       || !strcmp (argv[i], "-std=c9x")
+	       || !strcmp (argv[i], "-std=gnu9x"))
+	opts->cplusplus = 0, opts->cplusplus_comments = 1, opts->c89 = 0,
+	  opts->c9x = 1, opts->objc = 0;
+      break;
+
     case 'w':
       opts->inhibit_warnings = 1;
       break;

gcc/cpplib.h

@@ -108,6 +108,8 @@ struct cpp_buffer {
   char *fname;
   /* Filename specified with #line command.  */
   char *nominal_fname;
+  /* Last filename specified with #line command.  */
+  char *last_nominal_fname;
   /* Actual directory of this file, used only for "" includes */
   struct file_name_list *actual_dir;
 
@@ -422,6 +424,9 @@ struct cpp_options {
   /* Nonzero for the 1989 C Standard, including corrigenda and amendments.  */
   char c89;
 
+  /* Nonzero for the 199x C Standard, including corrigenda and amendments.  */
+  char c9x;
+  
   /* Nonzero means give all the error messages the ANSI standard requires.  */
   char pedantic;
 
@@ -639,10 +644,11 @@ struct definition {
   } args;
 };
 
-extern unsigned char is_idchar[256];
-extern unsigned char is_hor_space[256];
-extern unsigned char is_space[256];
-extern unsigned char trigraph_table[256];
+extern const unsigned char is_idstart[256];
+extern const unsigned char is_idchar[256];
+extern const unsigned char is_hor_space[256];
+extern const unsigned char is_space[256];
+extern const unsigned char trigraph_table[256];
 
 /* Stack of conditionals currently in progress
    (including both successful and failing conditionals).  */
@@ -669,6 +675,7 @@ typedef struct if_stack IF_STACK_FRAME;
 extern void cpp_buf_line_and_col PARAMS((cpp_buffer *, long *, long *));
 extern cpp_buffer* cpp_file_buffer PARAMS((cpp_reader *));
 extern void cpp_define PARAMS ((cpp_reader*, unsigned char *));
+extern void cpp_assert PARAMS ((cpp_reader *, unsigned char *));
 
 extern void cpp_error PVPROTO ((cpp_reader *, const char *, ...))
   ATTRIBUTE_PRINTF_2;
@@ -729,6 +736,9 @@ extern int finclude			PROTO ((cpp_reader *, int,
 extern void deps_output			PROTO ((cpp_reader *, char *, int));
 extern struct include_hash *include_hash PROTO ((cpp_reader *, char *, int));
 
+/* cppinit.c */
+extern void initialize_char_syntax	PROTO ((int));
+
 #ifndef INCLUDE_LEN_FUDGE
 #define INCLUDE_LEN_FUDGE 0
 #endif