This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Building with -ansi -D_POSIX_C_SOURCE=2; results encouraging
- To: gcc at gcc dot gnu dot org, gcc-patches at gcc dot gnu dot org
- Subject: Building with -ansi -D_POSIX_C_SOURCE=2; results encouraging
- From: "Zack Weinberg" <zackw at stanford dot edu>
- Date: Sat, 7 Jul 2001 11:24:25 -0700
In a spirit of scientific inquiry I ran a bootstrap of the gcc
subdirectory only, with BOOT_CFLAGS set to include -ansi
-D_POSIX_C_SOURCE=2, which has the effect of limiting system headers
to routines and types specified in C89, POSIX.1, and POSIX.2 (the
originals, not the 1993 and 1995 revisions).
The only problem encountered was a single use of // commentary in
java/verify.c. We didn't even get any additional missing-prototype
warnings. I've changed that to /* comments; patch appended.
The target libraries (esp. libstdc++) may have more problems with this
stunt, but then it's not clear that it makes sense for them. I do
think it makes sense for the compiler itself.
--
zw The master came to a yatai which was selling hot dogs. 'What do you
want on your hot dog?' he was asked. 'Nothing,' he replied. Then the
hot dog was enlightened.
-- a koan by Eric Hallstrom
* java/verify.c: Don't use // comments.
===================================================================
Index: java/verify.c
--- java/verify.c 2001/04/29 11:24:37 1.43
+++ java/verify.c 2001/07/07 18:23:36
@@ -1052,8 +1052,8 @@ verify_jvm_instructions (jcf, byte_ops,
if (!nargs || notZero)
VERIFICATION_ERROR
("invalid argument number in invokeinterface");
- // If we verify/resolve the constant pool, as we should,
- // this test (and the one just following) are redundant.
+ /* If we verify/resolve the constant pool, as we should,
+ this test (and the one just following) are redundant. */
if (! self_is_interface)
VERIFICATION_ERROR ("invokeinterface calls method not in interface");
break;