Kurt Roeckx wrote:
> This bug is indeed blocking the transition.
>
> Version 1.0.20-1 was built succesfully, but the testsuite error
> is now being ignored and the package built succesfully.
Personally I think ignoring errors thrown up by the test suite
is a bad idea. For instance the libsndfile testsuite found a bug
in gcc's code generator when generating ARMEL code for libvorbis.
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=515949
> So the question is:
> - Do you think it should migrate to testing as is or not?
No I don't. There is still a question mark over the functionality
on MIPS. It should be investigated properly.
> Reason why you could think it should migrate:
> - This bug affects the version in stable (1.0.17-4+lenny2) and/or
This bug affects functionality that was not present in 1.0.17.
> testing (1.0.18-2), and so is not a regression. If you
> mark it as found in those version, it will migrate.
This bug did not affect 1.0.18 because Ogg/Vorbis support was not
enabled at build time (missing Build-Depends). Hence, it is not
a regression.
> - It's not a release critical bug, it's only a minor functionaly
> that's broken and it's mostly ussable even with this. You
> could change the severity so it's not release critical.
I'm concerned about people using libsndfile to convert audio
data to Ogg/Vorbis and deleting the source before listening to
output files. Doing this could result in lost input files and
garbage output files.
> Anyway, try to fix this bug, I see this in the various built logs:
> arm:
> Line 231: Actual SNR ( 0.0) > target SNR (-31.0).
Arm we know about. Thats a code generator bug in the ARM gcc.
> mips:
> Line 231: Actual SNR ( 57.4) > target SNR (-31.0).
> hppa:
> Line 231: Actual SNR ( 57.4) > target SNR (-31.0).
> ia64:
> Line 231: Actual SNR ( 58.7) > target SNR (-31.0).
> mipsel:
> Line 231: Actual SNR ( 57.4) > target SNR (-31.0).
> powerpc:
> Line 279: Actual SNR ( 59.3) > target SNR (-29.0).
These are probably instances of the same bug.
> alpha:
> test_wrapper.sh: line 33: 20915 Floating point exception./floating_point_test
And that is another issue.
> If you want access to a certain arch, it's probably best
> to mail the porter list for that arch. If you can make
> a standalone test program for any of those issues I can
> test them on the various arches for you, but it will
> probably go faster if you can test things yourself.
I'll try the porters lists, probably starting with MIPS.
Erik
--
----------------------------------------------------------------------
Erik de Castro Lopo
http://www.mega-nerd.com/
--
To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
