[Pkg-qof-maintainers] Bug#383746: qof: FTBFS (hppa, sparc, m68k): timezone failures - help required

[Neil Williams]

Clint Adams wrote:
>> The 'main' function at the end of qof-0.7.0/qof/test/test-date.c
>> contains a DEBUG section that can be enabled by either using #define
>> TEST_DEBUG 1 or by commenting out the #ifdef and #endif in each case.
>> Compiling and running the test (either directly or using make check)
>> will then create a debug file, "test-date.trace". Would you mind
>> attaching "test-date.trace" please?
> 
> The attachment is from a build on etch rather than sid.

Thanks, but I do need the output from the sid machine too, please.

This output indicates a timezone failure. All results are exactly 5
hours out.

Either the timezone is not properly set on this box or it is not being
correctly determined.

Please can you run this simple series (on all affected boxes):
date -d"1995-03-26 06:00:00.000000000 +0000" +"%H %Z"
date -u -d"1995-03-26 06:00:00.000000000 +0000" +"%H %Z"
date -d"1995-03-26 06:00:00.000000000 +0500" +"%H %Z"
date -u -d"1995-03-26 06:00:00.000000000 +0500" +"%H %Z"
date -d"1995-03-26 06:00:00.000000000 -0500" +"%H %Z"
date -u -d"1995-03-26 06:00:00.000000000 -0500" +"%H %Z"

As these errors are machine-specific, I need to ensure that the debug
output is from the correct machine. Is it possible to get the debug
output on the original machine as well?

>     Error: stamp_and_scan():  df=5 str=1995-03-26T06:00:00Z scan=1995-03-26T11:00:00Z
>     Error: stamp_and_scan():  df=5 str=1995-03-26T05:58:20Z scan=1995-03-26T10:58:20Z
>     Error: stamp_and_scan():  df=6 str=1995-03-26 06:00:00.000000000 +0000 scan=1995-03-26 11:00:00.000000000 +0000
>     Error: stamp_and_scan():  df=6 str=1995-03-26 05:58:20.000072000 +0000 scan=1995-03-26 10:58:20.000072000 +0000

Margarita: Could you check the timezone setting for the hppa box that
completed the QOF build successfully, please? Would you also be able to
run the above test 'date ... ' commands, just for comparison? Thanks.

I'm trying to arrange ssh access to a hppa and possibly sparc machine
(as I don't have direct access to the Debian machines yet) via my local
LUG but until then, it's difficult to locate the real problem with the
timezone determination. Suffice to say it works on my current test
machines, the other buildd machines and at least one hppa machine.

The rest of the debug output and buildd logs was predictable as arising
from the timezone issue. This should mean a simple fix is available,
once someone can test directly on the affected machines.

-- 

Neil Williams
=============
http://www.data-freedom.org/
http://www.nosoftwarepatents.com/
http://www.linux.codehelp.co.uk/


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 252 bytes
Desc: OpenPGP digital signature
Url : http://lists.alioth.debian.org/pipermail/pkg-qof-maintainers/attachments/20060821/0f51922d/signature-0002.pgp