Bug#873783: wcc: FTBFS on non-amd64 architectures
Le 31 ao?t 2017 05:10:10 GMT+02:00, "Aaron M. Ucko" <ucko at debian.org> a ?crit :
>Source: wcc
>Version: 0.0.2+dfsg-1
>Severity: important
>Tags: upstream
>Justification: fails to build from source
>
>Builds of wcc for architectures other than (Linux) amd64 have been
>failing.
>
>On non-x86 architectures, there are two considerations: the
>embedded copy of openlibm under src/wsh generally has no
>$(ARCH)/Make.files, and GCC doesn't support -masm=intel regardless.
>You could sidestep the former by building against separately packaged
>libopenlibm-dev (as called for by Policy 4.13), but the latter may be
>more of a problem.
>
>On non-Linux architectures (kFreeBSD and presumably also the Hurd if
>and when clang becomes installable there), there's no <linux/elf-em.h>
>for arch.h to include.
>
>On i386, wsh somehow winds up compiled for the wrong architecture,
>leading to link errors:
>
>/usr/bin/ld: skipping incompatible
>/usr/lib/gcc/i686-linux-gnu/7/../../../i386-linux-gnu/libiberty.a when
>searching for -liberty
>/usr/bin/ld: skipping incompatible /usr/lib/i386-linux-gnu/libiberty.a
>when searching for -liberty
> /usr/bin/ld: cannot find -liberty
>
>On x32 (admittedly not a release architecture), wcc is still in the
>Needs-Build queue; I'm not sure what will happen there.
>
>At any rate, I'm reporting this as a single bug because I suspect you
>may just want to declare Architecture: amd64 (with the possible
>addition of x32) and be done with it. That said, you're certainly
>welcome to address any or all of these portability issues if feasible.
>
>Thanks!
Hello Aaron,
Thanks for the report !
I'll update the package to use the packaged libopenlibm-dev.
For the elf-em.h, I'll check for the corresponding file in Freebsd and Hurd and add some preprocessing to support all of them.
I'll test on arm, but I dunno if there is a debomatic for others arches than amd64. If you know if there is such testbed, it would be great.
--
O Philippe Thierry.
/Y\/ GPG: 7010 9a3c e210 763e 6341 4581 c257 b91b cdaf c1ea
o#o (sent from my smartphone)
Reply to: