On Fri, 2011-07-01 at 14:15 +0200, Dan Horák wrote: > g++ -DHAVE_CONFIG_H -I/usr/lib64/wx/include/gtk2-unicode-release-2.8 > -I/usr/include/wx-2.8 -D_FILE_OFFSET_BITS=64 -D_LARGE_FILES -D__WXGTK__ > -pthread -I../../src/include -I../../src/sdk/wxscintilla/include > -I../../src/include/scripting/include > -I../../src/include/scripting/sqplus -I../../src/include/mozilla_chardet > -Ulinux -Uunix -O2 -ffast-math -DCB_AUTOCONF -O2 -g -pipe -Wall > -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -fstack-protector > --param=ssp-buffer-size=4 -Wl,-z,relro -m64 -mtune=generic -DCB_PRECOMP > -Winvalid-pch -fPIC -DPIC -fexceptions -o ../../src/include/sdk.h.gch > -xc++-header ./sdk.h > In file included from ./sdk_common.h:40:0, > from ./sdk_precomp.h:13, > from ./sdk.h:17: > ./prep.h:32:1: warning: identifier ‘nullptr’ will become a keyword in C > ++0x [-Wc++0x-compat] > In file included from ./sdk_common.h:40:0, > from ./sdk_precomp.h:13, > from ./sdk.h:17: > ./prep.h: In member function ‘ID::operator void*() const’: > ./prep.h:333:45: warning: cast to pointer from integer of different size > [-Wint-to-pointer-cast] > /usr/lib/gcc/x86_64-redhat-linux/4.6.1/../../../../lib64/crt1.o: In > function `_start': > (.text+0x20): undefined reference to `main' > collect2: ld returned 1 exit status Precompiled headers are, of course, a crutch for a broken language. But given that... It looks like gcc is treating the mere presence of -Wl,* as a request to invoke the linker. This is a bit silly for precompiling headers, since they're not objects, the linker couldn't link them if it wanted to. But for example: % gcc -c -o stdio.gch /usr/include/stdio.h % gcc -c -o stdio.gch /usr/include/stdio.h -Wl,-z,relro % gcc -o stdio.gch /usr/include/stdio.h -Wl,-z,relro /usr/lib/gcc/x86_64-redhat-linux/4.6.0/../../../../lib64/crt1.o: In function `_start': (.text+0x20): undefined reference to `main' collect2: ld returned 1 exit status So, ick. Probably we should do this the other way around: default ld to -z relro (instead of -z norelro), and remove the -Wl bits from the global cflags. Whether this is a gcc bug or not is debateable, I guess. - ajax
Attachment:
signature.asc
Description: This is a digitally signed message part
-- devel mailing list devel@xxxxxxxxxxxxxxxxxxxxxxx https://admin.fedoraproject.org/mailman/listinfo/devel
