| *** replaceafill has quit IRC | 00:13 | |
| *** th1a has quit IRC | 01:33 | |
| *** menesis has quit IRC | 02:57 | |
| *** yvl has joined #schooltool | 09:09 | |
| *** th1a has joined #schooltool | 13:00 | |
| *** menesis has joined #schooltool | 13:35 | |
| *** khildin has joined #schooltool | 14:23 | |
| *** replaceafill has joined #schooltool | 15:22 | |
| th1a | hi replaceafill, menesis, yvl. | 15:30 |
|---|---|---|
| replaceafill | good afternoon | 15:30 |
| menesis | hello | 15:31 |
| th1a | OK, I just pushed half the journal documentation update to the book. | 15:31 |
| th1a | replaceafill was riding my ass about it last night, justifiably. | 15:31 |
| replaceafill | :D | 15:31 |
| th1a | I also filed some bugs in the process, as usual. | 15:32 |
| th1a | A couple easy ones that should be jammed into 2.6.2 if possible. | 15:33 |
| replaceafill | i don't understand the "split cell view" one | 15:33 |
| replaceafill | is that one of the "new" views in the journal? | 15:33 |
| replaceafill | maybe i haven't seen it yet | 15:33 |
| th1a | It is the homeroom view. | 15:34 |
| replaceafill | ah ok | 15:34 |
| th1a | Or, the attendance view when there is homeroom data. | 15:34 |
| th1a | It is jammed in the top half of the cell. | 15:34 |
| replaceafill | i'll set up some journal data to see it | 15:35 |
| th1a | We should tweak it a bit, but it isn't critical. | 15:35 |
| th1a | OK, the book revisions should be uploaded. | 15:37 |
| th1a | Where are we on the bugfix release menesis? | 15:37 |
| yvl | good morning / afternnoon | 15:37 |
| menesis | th1a: everything that I expected was fixed | 15:38 |
| menesis | I made a few report tests work, and nothing broke | 15:39 |
| th1a | hi yvl. | 15:39 |
| menesis | so today I was merging translations | 15:39 |
| menesis | and updating changes | 15:40 |
| menesis | will start uploading the releases shortly | 15:41 |
| menesis | will be finished tonight | 15:41 |
| th1a | Can we get that quick "Done" > "Save" fix and the journal navigation fix (warning about unsaved data) in first? | 15:42 |
| th1a | replaceafill: I assume the unsaved data thing is pretty much copy paste from the other gradebooks? | 15:42 |
| replaceafill | th1a, yeah, probably | 15:42 |
| * replaceafill goes look | 15:42 | |
| menesis | th1a: I have changed to Save already | 15:43 |
| th1a | OK. Cool | 15:44 |
| th1a | Excellent. menesis, anything else? | 15:45 |
| menesis | I have played a little with selenium testing of reports | 15:46 |
| menesis | But the options I tried, to bypass tasks queue, did not help | 15:47 |
| menesis | testing of new-style reports from the old skin did not work either. | 15:48 |
| menesis | so quite unsuccessful :( | 15:48 |
| * menesis done | 15:49 | |
| th1a | OK. thanks menesis. | 15:49 |
| th1a | yvl? | 15:49 |
| yvl | tried to evolve the database replaceafill gave me | 15:50 |
| yvl | gave up after 24 hours | 15:50 |
| yvl | total memory consumption did no exceed 7 GB | 15:50 |
| yvl | resident memory did not exceed 4 FB | 15:51 |
| yvl | GB | 15:51 |
| yvl | so at least in theory it should be possible to evolve it in a system with little ram | 15:51 |
| yvl | total being RAM + allocated swap | 15:51 |
| yvl | also did some profiling | 15:52 |
| yvl | temporal relationships do slow things down, but I can speed them up | 15:52 |
| yvl | now writing an experimental relationships implementation that should make everything | 15:52 |
| yvl | a) faster | 15:52 |
| yvl | b) use less memory | 15:52 |
| yvl | but I don't know by how much yet | 15:53 |
| yvl | in any case, views should not do 10.000.000 relationship checks, they will always be slow | 15:53 |
| th1a | What should they do instead? | 15:53 |
| th1a | Smarter querying? | 15:54 |
| th1a | Indexes? | 15:54 |
| yvl | yes | 15:54 |
| yvl | smarter querying | 15:54 |
| th1a | Less brute force, basically? | 15:54 |
| yvl | yep | 15:54 |
| th1a | Meaning? | 15:54 |
| yvl | I don't know how to answer that | 15:55 |
| yvl | basically just look at the offending views | 15:55 |
| yvl | and optimize | 15:55 |
| yvl | see if you can query some things once instead of many times | 15:55 |
| th1a | OK. | 15:56 |
| yvl | also relationships I | 15:56 |
| yvl | I'm playing with | 15:56 |
| yvl | should load less objects into memory | 15:56 |
| th1a | I mean, basically with ZODB you don't have the option of just doing a SQL-ish query? You pretty much always have to sort through all the things at least once? | 15:57 |
| yvl | also, their caches will be smaller | 15:57 |
| th1a | I'm not sure this has ever completely sunk in to me... | 15:57 |
| yvl | also, they'll be a bit more careful not to unghost objects unless necessary | 15:57 |
| yvl | well | 15:57 |
| yvl | yes, you can't do a SQL-ish query | 15:57 |
| yvl | and no | 15:57 |
| yvl | you can, but you have to, well, implement it manually | 15:58 |
| yvl | or you can just walk through all the objects | 15:58 |
| yvl | but that's why we put objects in different "containers" | 15:58 |
| yvl | for example: | 15:59 |
| yvl | direct person relationships are stored on person object | 15:59 |
| yvl | and only links to related things are stored there | 15:59 |
| th1a | OK. | 15:59 |
| yvl | so - less walking there | 15:59 |
| th1a | Well, again, this is the time to do this mucking about. | 16:00 |
| th1a | Just make sure you get done before the end of the year! | 16:00 |
| yvl | I'd rather get that done earlier ;) | 16:00 |
| th1a | Indeed. | 16:00 |
| yvl | I'll keep you updated as things progress | 16:01 |
| th1a | OK. | 16:02 |
| *** replaceafill has quit IRC | 16:02 | |
| th1a | Thanks yvl. | 16:02 |
| th1a | And... replaceafill... | 16:02 |
| th1a | So, in the meantime, I guess things might be happening in Nigeria?!?!? | 16:04 |
| th1a | Always a little hard to tell there... | 16:04 |
| th1a | I did forward that to you, right? | 16:05 |
| th1a | ... | 16:09 |
| *** th1a has quit IRC | 16:10 | |
| *** th1a has joined #schooltool | 16:10 | |
| th1a | Hello? | 16:11 |
| yvl | hello | 16:12 |
| yvl | how are you this fine day? | 16:12 |
| yvl | welcome to this fine parallel universe | 16:12 |
| th1a | Did I disappear there? | 16:12 |
| yvl | replaceafill is still gone :| | 16:13 |
| yvl | and it seems you just relaunched your xchat | 16:13 |
| th1a | OK. I wasn't sure if I was in a parallel univers. | 16:13 |
| th1a | e | 16:13 |
| th1a | OK... I need to go to the bank. | 16:15 |
| * th1a drops the bag of gravel. | 16:15 | |
| th1a | Have a good week/end guys! | 16:15 |
| yvl | thanks guys! | 16:16 |
| *** replaceafill has joined #schooltool | 17:27 | |
| replaceafill | th1a, just sent you an email | 17:34 |
| replaceafill | see you guys tomorrow | 17:35 |
| *** replaceafill has quit IRC | 17:35 | |
| *** mibofra has quit IRC | 20:53 | |
| *** mibofra has joined #schooltool | 20:56 | |
| *** menesis has quit IRC | 21:19 | |
| *** menesis has joined #schooltool | 21:52 | |
| *** mibofra has quit IRC | 21:55 | |
| *** mibofra has joined #schooltool | 21:57 | |
| *** th1a has quit IRC | 21:58 | |
| *** khildin has quit IRC | 23:31 | |
Generated by irclog2html.py 2.15.1 by Marius Gedminas - find it at mg.pov.lt!