IRC log of #schooltool for Thursday, 2013-10-03

*** th1a has quit IRC00:28
*** ignas has quit IRC01:54
*** replaceafill has quit IRC01:57
*** ignas has joined #schooltool02:21
*** ignas has quit IRC02:29
*** khildin has joined #schooltool10:45
*** ignas has joined #schooltool11:03
*** ignas has quit IRC11:49
*** yvl has quit IRC12:01
*** yvl has joined #schooltool12:02
*** ignas has joined #schooltool12:35
*** ignas has quit IRC13:22
*** th1a has joined #schooltool13:35
*** khildin has quit IRC14:13
*** menesis has quit IRC14:16
*** khildin has joined #schooltool14:55
*** menesis has joined #schooltool15:07
*** ignas has joined #schooltool15:10
*** replaceafill has joined #schooltool15:24
th1areplaceafill: Were we meeting today?15:44
th1aAre we?15:45
replaceafillth1a, yes15:45
replaceafilli have some "news"15:45
replaceafillabout the social skills instance15:45
replaceafilli merged the old code with the trunk branch, right?15:45
replaceafillbut when i try to evolve the database the process gets killed :(15:46
replaceafill8 GB of RAM + 2 of swap are not enough15:46
th1aOh, crap.15:46
th1aWhat's in there?15:46
replaceafilli'm now trying to improve the evolve scripts in core and cando15:46
replaceafillto not use .values() calls15:46
replaceafillwhich i think creates huges lists in memory15:47
replaceafilland also, i've set the generations version in cando to 115:47
replaceafillso only the first evolve gets called15:47
replaceafillfor core, evolve40 has the permission remap15:48
replaceafillassigning people to correct groups15:48
replaceafilland evolve41 puts manager in the correct groups15:49
replaceafillfor cando, evolve 1 is what copies skills to sections15:49
th1aThere's got to be some potential optimization in there.15:51
replaceafillimho the issue are those .values() calls15:52
replaceafilli'll finish removing those and try running the evolve again15:52
replaceafillrunning out of RAM! ;)15:54
replaceafill(green indicator)15:54
th1aI'm familiar with that display!15:54
th1areplaceafill:  I'm afraid that trying to generalize the autism views too much is just going to lead to node sniffing craziness.16:00
replaceafillth1a, i agree16:00
th1aOK.  We've got too many other things on our plates as it is.16:01
replaceafillout of ram (swap in purple)16:01
replaceafillth1a, kk, so we put those views in a custom package?16:02
replaceafillwell, there's already a custom schooltool.socialskills package16:02
replaceafillthat they're using16:02
th1aWe shouldn't need to do any rearranging.16:02
replaceafillwe should move their custom logic there16:03
replaceafillok, i'll stop this evolve before my laptop freezes again!16:03
th1aOh, is there other custom logic in their core?16:07
th1aYes, move that.16:07
replaceafillno, i meant, their custom menus, views and reports16:08
replaceafillthey don't have any custom *data* logic afaict16:08
replaceafillah btw16:08
replaceafillthey do have a custom IEP demographis16:08
replaceafillmaybe they could be interested in the whole IEP feature16:09
replaceafillnot sure if it means the same for them though16:09
th1aAh, well, actually I'm having a skype chat with someone from VA DOE about special ed in general.16:09
th1aThe state-wide thing.16:10
th1aIt is hard to tell if IEP is relevant, since this is essentially evaluating them in the context of their special needs, whereas IEP generally is just about defining the needs vis a vis regular ed.16:10
th1aAnyhow, we can see.16:11
replaceafillbtw, just removing app['persons'].values() improves running evolve40 by a lot!16:11
th1aHow often does that run?16:12
replaceafillbut they have like  8736 people in the db16:13
th1aThat doesn't seem like it should be too taxing...16:15
th1aWhat does it do, have to get the whole context of each person?16:16
replaceafillgets each person and queries groups that the person is member of16:18
th1aInstead of just doing one at a time as needed?16:19
replaceafillwhile trying t reproduce this bug:
replaceafilli just found that you can't delete courses with non-ascii __name__s16:56
*** replaceafill has quit IRC17:57
*** ignas has quit IRC18:51
*** ignas has joined #schooltool19:57
*** menesis has quit IRC19:57
*** ignas has quit IRC20:01
*** menesis has joined #schooltool21:09
*** replaceafill has joined #schooltool21:19
*** khildin has quit IRC21:21
*** ignas has joined #schooltool22:19
*** replaceafill has left #schooltool22:35
*** replaceafill has joined #schooltool22:47
*** menesis has quit IRC22:56

Generated by 2.15.1 by Marius Gedminas - find it at!