Package: libffi
Version: 3.0.9-2
Severity: wishlist
Hi
While preparing an armhf port on debian-ports.org, the autobuilders
built libffi:
http://buildd.debian-ports.org/fetch.php?pkg=libffi&arch=armhf&ver=3.0.9-2&stamp=1284234759&file=log&as=raw
but it you look at the testsuite, many tests FAIL.
This is because libffi needs porting to hard-float. An upstream patch
was developed by Chung-Lin Tang (Linaro) and sent upstream; will open a
separate bug for that one.
I think it would be best if libffi defaulted to failing the build when
the testsuite fails. Looking at Debian build logs, on most arches the
build passes without any failure. I think the other failures should be
marked as expected failures in Debian until they are sorted out.
I looked at the build logs of 3.0.9-2 at
https://buildd.debian.org/build.cgi?pkg=libffi&dist=unstable and found
that these arches had FAILs in the testsuite:
* ia64
FAIL: libffi.call/closure_fn0.c -O0 -W -Wall (test for excess errors)
[...]
# of unexpected failures 598
* kfreebsd-amd64
FAIL: libffi.call/closure_fn0.c -O0 -W -Wall execution test
[...]
# of unexpected failures 419
* mips, mipsel
FAIL: libffi.call/cls_double_va.c -O0 -W -Wall output pattern test, is 0.0
[...]
# of unexpected failures 10
debian-ports logs at
http://buildd.debian-ports.org/status/package.php?p=libffi&suite=unstable
note that only armhf is available
* armhf
FAIL: libffi.call/closure_fn0.c -O0 -W -Wall output pattern test, is 1 2 3 4
127 429 7 8 0 13 19 0 1 69104 69080 69104 3: 207905
[...]
# of unexpected failures 143
Cheers,
--
Loïc Minier
--
To UNSUBSCRIBE, email to debian-gcc-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/20100924093808.ga9...@bee.dooz.org
