On 14 February 2017 at 17:23, Jakub Jelinek <jakub@xxxxxxxxxx> wrote: > On Mon, Feb 13, 2017 at 08:39:00PM +0100, Francisco J. Tsao Santin wrote: >> Hi all, >> >> We have a problem with the hardlink package. The koji build made by Fedora >> Release Engineering failed in armv7hl and i686 architectures. I saw the logs >> and I tried a mock build in my own machine too, the problem is always the same: >> >> + gcc -O2 -g -pipe -Wall -Werror=format-security -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -fstack-protector-strong --param=ssp-buffer-size=4 >> -grecord-gcc-switches -specs=/usr/lib/rpm/redhat/redhat-hardened-cc1 -m32 -march=i686 -fasynchronous-unwind-tables hardlink.c -o hardlink >> /tmp/cc1xWk9j.o: In function `rf': >> /builddir/build/BUILD/hardlink-1.1/hardlink.c:257: undefined reference to `stcmp' >> collect2: error: ld returned 1 exit status >> >> Of course, stcmp is declared. It's very weird, because the other architectures >> don't have the issue, nor f25 builds. >> >> Any idea about what can be happening? > > GNU89 vs. C99 inlining, that code is really old and apparently nobody > touched it since it has been written. > You don't touch a work of art... you only let the master do touch ups. Thank you for writing it way back then.. it made my life of ftp serving so much easier. > As everything is in a single file, either change all the inlines in the file > to static inline, or turn them into extern inline (in C99/C11 that means > if the function can't be inlined, an out of line copy is then emitted), > or change inline into inline __attribute__((__gnu_inline__)), or compile > with -fgnu89-inline. > > Jakub > _______________________________________________ > devel mailing list -- devel@xxxxxxxxxxxxxxxxxxxxxxx > To unsubscribe send an email to devel-leave@xxxxxxxxxxxxxxxxxxxxxxx -- Stephen J Smoogen. _______________________________________________ devel mailing list -- devel@xxxxxxxxxxxxxxxxxxxxxxx To unsubscribe send an email to devel-leave@xxxxxxxxxxxxxxxxxxxxxxx
