[02:14] <[ESphynx]> hey guys, I'm hitting this extremely annoying hang which seems to be new...
[02:15] <[ESphynx]> https://gist.github.com/anonymous/cbcac0e6413fc30d6a75  calling XCreateIC seems to jam
[07:19] <dholbach> good morning
[18:35] <picca> Hello, I am the DD of the tango package. the 14.04 version is affected by this debian bug #743508, which was fixed in the 8.1.2c+dfsg-5 version. I would like an update of this package in 14.04. Is is possible ?
[18:43] <jtaylor> picca: sure if you have  patch
[18:45] <picca> hello julian, which can of patch do you need, the version I would like uploaded to 14.04 is the current developpement version
[18:47] <jtaylor> the whole version?
[18:47] <jtaylor> that somewhat depends on the size of the diff
[18:50] <picca> Yes, is adds two patches from the upstream, and also this fix  #657146
[18:51] <jtaylor> the upgrade procedure is described here: https://wiki.ubuntu.com/StableReleaseUpdates
[18:52] <jtaylor> first we need the fixed version in utopic (dev release)
[18:52] <picca> utopic already contain the right version
[18:54] <jtaylor> oh tango, I looked at pytango
[18:55] <picca> In fact the problem is in the tango library, not in the python binding
[18:55] <picca> the real problem is with zeromq3 4.0.4., but tango upstream provided a patch which fix this bad interaction
[18:56] <picca> 14.04 is affected because zeromq is the 4.0.4 version...
[19:00] <jtaylor> file a bug with a minimal debdiff and I'll see to uploading it
[19:00] <jtaylor> bug description should have a testcase as described in thewiki
[19:03] <picca> I do not have an ubuntu machine, I can open a bug from the we interface ?
[19:03] <picca> the wiki is StableReleaseUpdates previour link right ?
[19:06] <jtaylor> yes
[19:06] <jtaylor> there is a webinterface
[19:07] <jtaylor> do you have a launchpad account?
[19:08] <picca> yes
[19:14] <picca> ok, I filled a bug but I do not have the debdiff now Bug #1355394 , I need more time to build the -3 -> -5 debdiff
[19:18] <jtaylor> picca: the backportpackage script from ubuntu-dev-tools should give you a start to work with
[19:18] <jtaylor> should also work in debian
[19:20] <picca> ok I will try, thanks
[19:36] <picca> jtaylor: I attached something to the bug, can you check that it is ok
[19:38] <jtaylor> not really, you would need to debdiff the trusty package against the backport
[19:38] <jtaylor> and then the backport is not yet a proper backport though in this case probably just some minor adjustments are required
[19:41] <jtaylor> oh some stuff is renamed in the package, that should not be part of the sru
[19:43] <picca> so I miss used the backportpackahe script...
[19:45] <jtaylor> no the debdiff
[19:46] <jtaylor> actually I'd probably skip the unowned file backport
[19:46] <jtaylor> doesn't seem important enough
[19:46] <jtaylor> only the zmq fix should be fine
[19:47] <jtaylor> picca: why was --disable-static added to rules?
[19:48] <picca> to reduce the compilation time, I do not install the static library
[19:49] <picca> so in fact the important point is just to add the two upstream pacthes and that's all
[19:49] <jtaylor> yes
[19:49] <jtaylor> I can do so, seems simple enough
[19:52] <jtaylor> picca: your testcase does not run
[19:52] <jtaylor> I need to do some setup first?
[19:52] <picca> you need to install first a mysql-server
[19:53] <picca> then the tango-db package, which should start a tango database usign the tango user
[19:53] <picca> during the configuration it should ask for the TANGO_HOST
[19:54] <picca> normally this should be ok if you use the default value
[19:54] <picca> then you install python-pytango
[19:54] <picca> and at last you should try the python snipset furnish by the bug reporter
[19:57] <jtaylor> mh now how do I get a mysql server to start in a chroot
[19:57] <picca> this is a problem indeed...
[20:02] <picca> Is it possible to start the mysql-server outside the chroot and mont the mysql socket inside the chroot ?
[20:03] <jtaylor> ok I think I got it to work
[20:03] <jtaylor> its now hanging as expected
[20:03] <jtaylor> lets see if my debdiff fixes it (still building)
[20:05] <picca> thanks :)
[20:12] <jtaylor> picca: are there upstream bug logs?
[20:12] <jtaylor> not sure how p812_3 is related
[20:12] <jtaylor> the sleep in the connect is also dodgy
[20:15] <picca> http://sourceforge.net/p/tango-cs/bugs/662/
[20:16] <picca> p 812_3 is not related, but it is better to provide all the upstream pacth available. upstream produce a few hotfix for a release, usually they are  important
[20:20] <jtaylor> ok the patches fixed the hang but I get: 2014-08-11 20:19:37.923333 SYS/DATABASE/2 STATE CHANGE [API_AttributePollingNotStarted] The polling (necessary to send events) for the attribute state is not started
[20:22] <picca> I will ask the main bug reporter what is the expected result of the method. I updatethe bug report with your informations
[20:24] <jtaylor> is there a testsuite that can be run to check its working?
[20:24] <picca> not yet, I asked the upstream for a test suite, but it is not yet part of tango source package. I filled a bug for this.
[20:26] <picca> Now the problem is that they are using a dedicated mysql-server with pre-filled tables for the test
[20:26] <picca> so I need a script which populate this database.
[20:26] <picca> It would be a nice test case for debci :)
[20:26] <picca> b
[20:32] <picca> from what I know about tango it seems to me that the return values are ok, the state attribut is a special one and users need to configure specially the polling of this attribute, to obtain events
[20:33] <picca> the -5 version is working on my machine, but indeed I would be nuch confident If I would have a test suite... maybe one day
[20:37] <jtaylor> do you get the same message?
[20:42] <picca> this is strange on my debian unstable it does not work ???
[20:43] <picca> I do not have the output message
[20:44] <picca> ouf I got it >>> id = dev.subscribe_event("state", PyTango.EventType.CHANGE_EVENT, cb, [], True)
[20:44] <picca> >>> 2014-08-11 22:43:09.745188 SYS/DATABASE/2 STATE CHANGE [API_AttributePollingNotStarted] The polling (necessary to send events) for the attribute state is not started
[20:47] <jtaylor> I also get it in stable without zmq
[20:47] <jtaylor> so its probably fine
[20:50] <picca> Yes in stable I never activated the event system... whcih require a non maintained library omninotify
[20:50] <jtaylor> picca: I uploaded it to -proposed
[20:50] <picca> thanks a lot for your efforts
[20:51] <picca> also for ipython ;)
[20:53] <jtaylor> should have probably looked at it better earlier
[20:53] <jtaylor> I knew of the bug before trusty was released
[20:53] <jtaylor> but for some reason I though trusty is not affected
[20:53] <Logan_> Noskcaj: stealing my syncs, I see :P
[20:53] <jtaylor> must have been before zeromq4 landed
[20:54] <jtaylor> or me just being stupid :)
[20:54] <Noskcaj> Logan_, Saves you work
[20:55] <jtaylor> mh, right I still wanted to fix gccs sanatizers :/
[20:58] <picca> To my opinion zeromq3 should be reverted to the lastest 3.2.4 version and a new zeromq4 should be introduce...
[20:59] <jtaylor> probably
[20:59] <jtaylor> but its too late now
[20:59] <jtaylor> debian would probably prefer to only ship one zmq and thus remove 2 and 3
[21:00] <picca> Indeed, it would be nice to have only once version of zeromq
[21:01] <picca> but the zeromq maintainer does not seems to be very active.
[21:01] <jtaylor> I'm still waiting for sodium to finally get uploaded so I can build pyzmq with encryption
[21:01] <jtaylor> there seems to be some progress on that now
[21:03] <picca> have a good night, and thanks again for your help