PySide Bugzilla Closed for New Bugs

PySide is now a Qt Add-on and uses the Qt Project's JIRA Bug Tracker instead of this Bugzilla instance. This Bugzilla is left for reference purposes.

First Last Prev Next    This bug is not in your last search results.
Bug 733 - Meta-bug for tests failures on "special" architectures.
: Meta-bug for tests failures on "special" architectures.
Status: NEW
Product: PySide
Classification: Unclassified
Component: PySide
: 1.0.1
: Other Linux
: P3 normal
Assigned To: renato filho
: http://alioth.debian.org/~odyx-guest/...
:
:
  Show dependency tree / graph
 
Reported: 2011-03-15 17:21 EET by Didier Raboud
Modified: 2012-03-08 16:57 EET (History)
9 users (show)

See Also:

Attachments
Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Didier Raboud 2011-03-15 17:21:49 EET 
Hi dear PySide developers, 

as mentionned on bug 730:  now that a Qt >= 4.7. reached Debian unstable, I
pushed the whole PySide stack to Debian unstable, where it slowly gets built on
many different architecture. Unfortunately, there is a set of tests that fail
to build on several "special" architectures, with tough errors (Segmentation
faults, Illegal Access or Bus errors).

I set up a 4-times-a-day cron job that publishes the list of failing tests per
architectures (and reversely) there (so it's not yet an accurate picture of all
failing tests):

http://alioth.debian.org/~odyx-guest/packages/pyside/pyside_1.0.0-2_analysis.txt

There are already several tests failed on alpha, ia64 and sparc, but not all
builds are done yet; see there (two first lines are the current state for Qt
and QtWebkit (both build-depends of PySide)):

https://buildd.debian.org/status/package.php?p=qt4-x11%2Cqtwebkit%2Chorizontal-void-line%2Capiextractor%2Cgeneratorrunner%2Cshiboken%2Cpyside&suite=sid&compact=compact

So I'm hereby filing this bug as a reference, with a "normal" priority (as the
bug 730). I can produce gdb backtraces for all those tests on all
architectures, but it takes a considerable amount of time, so I don't intend to
produce all of them pro-actively: please ask me directly (IRC, mail or this bug
will make it) and I'll produce a given test backtrace in ~1 day.

Cheers,

OdyX
Comment 1 Luciano Wolf 2011-03-16 19:29:24 EET 
Due to the importance of having this on Debian we are rising priority to P3.
Comment 2 Didier Raboud 2011-03-28 13:00:37 EEST 
I have just updated the analysing script to not spot the architectures that do
not matter for Debian testing (alpha, hppa). The 

http://alioth.debian.org/~odyx-guest/packages/pyside/pyside_1.0.0-2_analysis.txt

link now lists only the tests that must be fixed for PySide to enter testing.
Comment 3 Hugo Parente Lima 2011-03-28 21:06:05 EEST 
Nice, but better regenerate this list after 1.0.1 (to be released by the end of
this week), because Renato fixed some hidden bugs that occurred only on the new
Ubuntu.
Comment 4 Didier Raboud 2011-04-04 11:21:20 EEST 
1.0.1 got released to Debian unstable and is currently being built by the
various build daemons.

http://alioth.debian.org/~odyx-guest/packages/pyside/pyside_1.0.1-1_analysis.txt

shows somehow similar test failures (but still missing mips{,el} builds).
Comment 5 Didier Raboud 2011-04-07 11:56:39 EEST 
Oookay. Now, according to
https://buildd.debian.org/status/logs.php?pkg=pyside&ver=1.0.1-1 , PySide 1.0.1
has been built on all release architectures (aka all but alpha and hppa which
will get removed from Debian soon anyway).

As you can see from
http://alioth.debian.org/~odyx-guest/packages/pyside/pyside_1.0.1-1_analysis.txt
, the good new is that many architectures have successful tests: armel, amd64,
i386,  kfreebsd-*, amd64, mips, mipsel, powerpc and s390.

There are actually two architectures that have failing tests : ia64 and sparc. 

So now it's your turn ! As told I can prepare backtraces on those
architectures, but I need you to ask.

Finally, note that I reported a "Release-Critical" bug against the pyside
Debian package to forbid its migration to Wheezy (our next stable release) :
http://bugs.debian.org/618347 . I will close it once all architectures have all
their tests run successfully.

Cheers, OdyX
Comment 6 Hugo Parente Lima 2011-04-07 16:12:08 EEST 
The backtraces or at least the tests verbose output would be nice to have since
we don't have access to any sparc or ia64 machines.
Comment 8 Didier Raboud 2011-05-12 11:04:12 EEST 
I just enhanced my buildlogs parsing script to also detect timeouts, so you get
a slight number of new errors. :-)

http://alioth.debian.org/~odyx-guest/packages/pyside/pyside_1.0.2-1_analysis.txt
Comment 9 anatoly techtonik 2011-10-14 11:21:47 EEST 
How hard is to patch your script to extract tracebacks into separate dir?

Updated link:
http://alioth.debian.org/~odyx-guest/packages/pyside/pyside_1.0.7-2_analysis.txt
Comment 10 Didier Raboud 2011-10-14 11:37:06 EEST 
Hmm. Tracebacks are done manually, on developer boxes. The point is that we
don't want to automatically run debugging sessions when some tests fail, as
that would imply a build-dependency on gdb and is not necessarily automatable.
Comment 11 anatoly techtonik 2011-10-14 11:48:37 EEST 
I see. How about standard Python traceback? It should be available from stderr.
Comment 12 Matti Airas 2012-03-08 16:57:29 EET 
PySide is now a Qt-addon and uses Qt Project's JIRA tool for tracking bugs.
Please verify that the bug is still valid and re-submit it in the address
below:

https://bugreports.qt-project.org/

Sorry for the inconvenience!
First Last Prev Next    This bug is not in your last search results.