I just spent half an hour to debug mysterious build failures (make complaining about missing dependencies) in some third party code. (The reason was a missing -Ipath.) Given what the documentation of -MG in the preprocessor says I'm not sure this is a real bug, but as an enhancement request I think gcc shouldn't pass -MG to the preprocessor if it is used in combination with -c (or -S etc.), or produce an error (it doesn't really make sense to use -MG with -c). This behaviour is the same in gcc --version: gcc-3.3 (GCC) 3.3.6 (Debian 1:3.3.6-7) gcc-3.4 (GCC) 3.4.5 20050613 (prerelease) (Debian 3.4.4-1) gcc-4.0 (GCC) 4.0.1 20050617 (prerelease) (Debian 4.0.0-10) $ cat >t.c #include "foo.h" $ gcc -MD -c t.c t.c:1:17: foo.h: No such file or directory $ ls -l t.* -rw-r--r-- 1 js js 17 Jun 29 15:35 t.c -rw-r--r-- 1 js js 0 Jun 29 15:36 t.d $ gcc -MD -MG -c t.c $ ls -l t.* -rw-r--r-- 1 js js 17 Jun 29 15:35 t.c -rw-r--r-- 1 js js 15 Jun 29 15:35 t.d -rw-r--r-- 1 js js 637 Jun 29 15:35 t.o $ cat t.d t.o: t.c foo.h $
Confirmed. I'm inclined to agree that we should disallow -MG in this situation. Perhaps this is best done in c-opts.c:sanitize_cpp_opts()
I am testing a patch for this.
I finally submitted this patch. http://gcc.gnu.org/ml/gcc-patches/2008-05/msg00520.html
Subject: Bug 22231 Author: tromey Date: Fri May 9 22:17:00 2008 New Revision: 135128 URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=135128 Log: gcc PR preprocessor/22231: * c-opts.c (sanitize_cpp_opts): Disallow -MG if compilation is proceeding. gcc/testsuite PR preprocessor/22231: * gcc.dg/pr22231.c: New file. Added: trunk/gcc/testsuite/gcc.dg/pr22231.c Modified: trunk/gcc/ChangeLog trunk/gcc/c-opts.c trunk/gcc/testsuite/ChangeLog
Fixed on trunk. It is very unlikely that this will be back-ported, so I'm closing the bug.