gcc -M bug.

Zack Weinberg zack@wolery.cumb.org
Mon Jul 17 13:11:00 GMT 2000


On Mon, Jul 17, 2000 at 12:52:23PM -0700, H . J . Lu wrote:
> Now with gcc -M,
> 
> # /work/build/gnu/bin/gcc-3.0-libstdc++v3/gcc/xgcc -B/work/build/gnu/bin/gcc-3.0-libstdc++v3/gcc/  -M -I.  -ansi -pedantic-errors t.cc
> # 1 "t.cc"
> #pragma weak foo
> # 1 "t.cc"
> 
> 
> # 3 "t.h" 1 3
> # 33 "/work/build/gnu/bin/gcc-3.0-libstdc++v3/gcc/include/exception" 1 3
> #pragma interface "exception"
> # 33 "/work/build/gnu/bin/gcc-3.0-libstdc++v3/gcc/include/exception" 3
> t.o: t.cc t.h \
>   /work/build/gnu/bin/gcc-3.0-libstdc++v3/gcc/include/exception
> 
> Has the meaning of -M been changed? Why those extra stuff?

That's a bug - a harmless one, since # is Make's comment character,
but still.

Please try this patch.

zw

===================================================================
Index: cpplex.c
--- cpplex.c	2000/07/16 13:35:23	1.76
+++ cpplex.c	2000/07/17 20:09:41
@@ -326,7 +326,7 @@ output_line_command (pfile, print, line)
     putc ('\n', print->outf);
   pfile->need_newline = 0;
 
-  if (CPP_OPTION (pfile, no_line_commands))
+  if (CPP_OPTION (pfile, no_line_commands) || CPP_OPTION (pfile, no_output))
     return;
 
   /* If ip is null, we've been called from cpp_finish, and they just


More information about the Gcc-bugs mailing list