| Summary: | dev-python/pyro-4.20: version bump | ||
|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | Amadeusz Żołnowski (RETIRED) <aidecoe> |
| Component: | New packages | Assignee: | Amadeusz Żołnowski (RETIRED) <aidecoe> |
| Status: | RESOLVED FIXED | ||
| Severity: | enhancement | CC: | python |
| Priority: | Normal | ||
| Version: | unspecified | ||
| Hardware: | All | ||
| OS: | Linux | ||
| URL: | http://pythonhosted.org/Pyro4/changelog.html | ||
| Whiteboard: | |||
| Package list: | Runtime testing required: | --- | |
| Bug Depends on: | 474622, 476346 | ||
| Bug Blocks: | |||
|
Description
Amadeusz Żołnowski (RETIRED)
2013-06-24 17:45:16 UTC
There is the issue of KEYWORDS="~amd64 ~ppc ~x86 ~amd64-linux ~x86-linux". Now I have added serpent, it needs a keyword request to support ppc. Alternately you can indicate to elect to drop ~ppc, make the version bump viable and keyword request both packages. What be yout pleasure? Also there is one more test that seems to require seding out, but I we can fix that. (In reply to Ian Delaney from comment #1) > There is the issue of > KEYWORDS="~amd64 ~ppc ~x86 ~amd64-linux ~x86-linux". > > Now I have added serpent, it needs a keyword request to support ppc. Thank you! :-) > Alternately you can indicate to elect to drop ~ppc, make the version bump > viable and keyword request both packages. What be yout pleasure? We can wait until serpent gets the keyword. > Also there is one more test that seems to require seding out, but I we can > fix that. OK, next weekend I might have a bit of spare time, to so I can take a look at the broken test. (In reply to Amadeusz Żołnowski from comment #2) > (In reply to Ian Delaney from comment #1) > > Also there is one more test that seems to require seding out, but I we can > > fix that. > > OK, next weekend I might have a bit of spare time, to so I can take a look > at the broken test. In place of working the error, which I already did, there's an alternate challenge. On the basis that it is just another net seeking tests that needs seding, it's seded currently here by -e "s/testGetIpVersion/_&/" \ However, I attempted to make a patch of the actioned sed statements, made it in the 'usual' way, and it errored on being run with Hunk #x FAILED at xx (different line endings)? ex temp/pyro-4.20-skip-tests.patch.out I am totally miffed. Please try making a patch that does what the sed statements do. I'll now make a keyword request but don't hold yr breath ok I managed to fast track serpent. Here's some homework;
1. dev-python/pyro $ PYTHON_TARGETS="python2_6" ebuild pyro-4.20.ebuild clean test
* python2_6: running distutils-r1_run_phase python_test
ERROR: testCallCustomObjAsArg (PyroTests.test_serialize.SerializeTests_marshal)
----------------------------------------------------------------------
Traceback (most recent call last):
File "/mnt/gen2/TmpDir/portage/dev-python/pyro-4.20/work/Pyro4-4.20/tests/PyroTests/test_serialize.py", line 221, in testCallCustomObjAsArg
self.skipTest("marshal can't serialize custom objects")
AttributeError: 'SerializeTests_marshal' object has no attribute 'skipTest'
=================================================================
FAILED (errors=4)
Clumsy testsuite is failing tests that are intended to be skipped. These are not real test failures but testsuite process stuffups.
2. Make the patch described above and substitute the sed statements at your leisure.
*pyro-4.20 (10 Jul 2013)
10 Jul 2013; Ian Delaney <idella4@gentoo.org> +pyro-4.20.ebuild:
bump, drop pypy1_9 support, add new dep serpent', addresses Bug #474624 by
aidecoe
Then please close the bug FIXED
(In reply to Ian Delaney from comment #4) > * python2_6: running distutils-r1_run_phase python_test > > ERROR: testCallCustomObjAsArg > (PyroTests.test_serialize.SerializeTests_marshal) > ---------------------------------------------------------------------- > Traceback (most recent call last): > File > "/mnt/gen2/TmpDir/portage/dev-python/pyro-4.20/work/Pyro4-4.20/tests/ > PyroTests/test_serialize.py", line 221, in testCallCustomObjAsArg > self.skipTest("marshal can't serialize custom objects") > AttributeError: 'SerializeTests_marshal' object has no attribute 'skipTest' > > ================================================================= > FAILED (errors=4) > > Clumsy testsuite is failing tests that are intended to be skipped. These are > not real test failures but testsuite process stuffups. https://github.com/irmen/Pyro4/issues/10 If irmen is not going to fix it quickly his way, I'll prepare some patch in the CEST evening/tomorrow. I've made a pull request with a fix: https://github.com/irmen/Pyro4/pull/11 + 18 Jul 2013; Amadeusz Żołnowski <aidecoe@gentoo.org> pyro-4.20.ebuild, + +files/4.20-0001-Use-unittest2-for-older-Python-version.patch: + Fixed tests for Python 2.6 and 3.1. + Wrt sed I'd provide a patch which would be applied by upstream as well to avoid manually finding out which tests use networking and which not. But that later. The pull request, all good. (In reply to Amadeusz Żołnowski from comment #7) > > Wrt sed I'd provide a patch which would be applied by upstream as well to > avoid manually finding out which tests use networking and which not. But > that later. On this matter we are fundamentally on totally different pages. Upstream don't give a toss about their tests connecting to the net, that is purely an added gentoo/portage layer. The sed works at any rate, so really it doesn't matter. It all looks fine to me. Close at your lesiure (In reply to Amadeusz Żołnowski from comment #6) > I've made a pull request with a fix: https://github.com/irmen/Pyro4/pull/11 https://github.com/irmen/Pyro4/pull/12 - merged to upstream master. |