| chat.freenode.net #tryton log beginning Thu Jan 7 00:00:02 CET 2010 | ||
| 2010-01-07 00:51 -!- vengfulsquirrel(n=ian@c-69-181-194-95.hsd1.ca.comcast.net) has left #tryton | ||
| 2010-01-07 01:56 -!- ikks(n=ikks@190.158.102.38) has joined #tryton | ||
| 2010-01-07 05:20 -!- yangoon(n=mathiasb@p549F6D19.dip.t-dialin.net) has joined #tryton | ||
| 2010-01-07 06:28 -!- udono(n=udono@dynamic-unidsl-85-197-24-177.westend.de) has joined #tryton | ||
| 2010-01-07 06:30 -!- udono(n=udono@dynamic-unidsl-85-197-24-177.westend.de) has joined #tryton | ||
| 2010-01-07 08:25 -!- bechamel(n=user@host-85-201-159-186.brutele.be) has joined #tryton | ||
| 2010-01-07 08:31 -!- Timitos(n=timitos@88.217.184.172) has joined #tryton | ||
| 2010-01-07 09:05 -!- FWiesing(n=FWiesing@194-208-185-012.tele.net) has joined #tryton | ||
| 2010-01-07 09:11 -!- FWiesing(n=FWiesing@194-208-185-012.tele.net) has left #tryton | ||
| 2010-01-07 10:48 -!- sharoon(n=sharoont@117.197.58.161) has joined #tryton | ||
| 2010-01-07 10:59 -!- paepke(n=paepke@79.235.39.152) has joined #tryton | ||
| 2010-01-07 11:55 -!- paepke(n=paepke@p4FEB2798.dip0.t-ipconnect.de) has joined #tryton | ||
| 2010-01-07 12:12 -!- sharoon(n=sharoont@117.197.57.172) has joined #tryton | ||
| 2010-01-07 14:53 -!- yangoon(n=mathiasb@p549F6D19.dip.t-dialin.net) has joined #tryton | ||
| 2010-01-07 15:33 -!- essich(n=essich@p4FCF9611.dip0.t-ipconnect.de) has joined #tryton | ||
| 2010-01-07 16:05 -!- johbo(n=joh@statdsl-085-016-072-173.ewe-ip-backbone.de) has joined #tryton | ||
| 2010-01-07 16:48 -!- cedk(n=ced@gentoo/developer/cedk) has joined #tryton | ||
| 2010-01-07 17:28 -!- woakas(n=woakas@190.144.69.234) has joined #tryton | ||
| 2010-01-07 17:36 -!- essich(n=essich@p4FCF9611.dip0.t-ipconnect.de) has joined #tryton | ||
| 2010-01-07 18:47 -!- sharoon(n=sharoont@117.197.57.172) has left #tryton | ||
| 2010-01-07 20:06 -!- LucaSub1(n=LucaSub@host158-190-dynamic.9-87-r.retail.telecomitalia.it) has joined #tryton | ||
| 2010-01-07 20:13 -!- vengfulsquirrel(n=ian@c-69-181-194-95.hsd1.ca.comcast.net) has joined #tryton | ||
| 2010-01-07 20:33 <vengfulsquirrel> Hi what does select=1 do on a model property? | ||
| 2010-01-07 20:39 <cedk> vengfulsquirrel: it adds index on column of table | ||
| 2010-01-07 20:39 <vengfulsquirrel> So for each model property it adds a single index of just that column ? Or does it add composite indices as well somehow? | ||
| 2010-01-07 20:40 <vengfulsquirrel> I was looking at this patch: '''Add select=1 on res and field of ir.property | ||
| 2010-01-07 20:40 <vengfulsquirrel> It is the two fields on which get/set functions make search ''' and was curious if it would help my slowness situation. | ||
| 2010-01-07 20:43 <cedk> yes, search must be faster | ||
| 2010-01-07 20:49 <vengfulsquirrel> Yeah, well there are 23,000 rows in ir_property. I don't really understand how that comes into play. | ||
| 2010-01-07 20:50 <cedk> vengfulsquirrel: I don't understand what you don't understand | ||
| 2010-01-07 21:02 <cedk> vengfulsquirrel: any way, tell us if it improves the situation | ||
| 2010-01-07 21:19 <vengfulsquirrel> cedk: Hmm yeah doesn't seem to have any effect at least in the specific case I use to recreate my problem. | ||
| 2010-01-07 21:20 <vengfulsquirrel> Ir is responsible for just tracking models and their properties right? Those probably are not accessed repeatedly in one client request besides a few times per model right? | ||
| 2010-01-07 21:23 <udono> vengfulsquirrel: hi, maybe you need to reindex the ir_sequence table? There is a function in pgadmin. But I don't know if it is related | ||
| 2010-01-07 21:24 <udono> sorry, ir_property... | ||
| 2010-01-07 21:24 <vengfulsquirrel> Well I guess my question is do you think indexing ir_property or not will really have tremendous impact on performance? | ||
| 2010-01-07 21:25 <vengfulsquirrel> I mean it will be "faster" but would it be noticeable ? | ||
| 2010-01-07 21:26 <udono> vengfulsquirrel: good question... | ||
| 2010-01-07 21:56 <cedk> ir_property is used for each Property field | ||
| 2010-01-07 21:56 <cedk> like cost_price on product etc. | ||
| 2010-01-07 22:01 <vengfulsquirrel> cedk,udono: Sorry, okay I have another question, the times returned when I do this(in trytond/backend/postgresql/database.py : execute) don't seem correct, does self.cursor.execute not happen immediately ? http://paste.pocoo.org/show/163070/ | ||
| 2010-01-07 22:03 <cedk> vengfulsquirrel: it is | ||
| 2010-01-07 22:04 <vengfulsquirrel> It does you mean? | ||
| 2010-01-07 22:04 <cedk> vengfulsquirrel: yes, it does | ||
| 2010-01-07 22:05 <cedk> vengfulsquirrel: why do you say it doesn't seem correct? | ||
| 2010-01-07 22:05 <vengfulsquirrel> Well my crude time printing method prints this : 0:00:00.003716 | ||
| 2010-01-07 22:05 <vengfulsquirrel> but EXPLAIN ANALYZE says it takes 7 seconds | ||
| 2010-01-07 22:07 <vengfulsquirrel> Oh I'm an idiot | ||
| 2010-01-07 22:07 <cedk> vengfulsquirrel: I think the time is the time of your query "explain" | ||
| 2010-01-07 22:15 <vengfulsquirrel> Okay well there were a lot of things wrong with what I was doing but the conclusion is those queries most likely aren't causing the 5 - 10 minute lag. | ||
| 2010-01-07 22:16 <vengfulsquirrel> I misread the analyzer and its 7 milliseconds and I had the datetime.now() BEFORE the execute instead of after. | ||
| 2010-01-07 22:31 -!- FWiesing(n=FWiesing@194-208-185-012.tele.net) has joined #tryton | ||
| 2010-01-07 22:31 <vengfulsquirrel> Maybe its possible that the large amount of data is difficult to load in python even though the database can pull it in a fair amount of time? | ||
| 2010-01-07 22:32 <vengfulsquirrel> I asked this the other day too, why are all the fields on my model pulled(lines) when I return from viewing a single model using switch, here is the output of the client using -v, 1. I get the list of inventories(no lines field is requested), 2. I view the inventory (the lines are requested for that single inventory) 3. I click switch to go back to the list of inventories(all the inventories AND their lines are requested--why?) ht | ||
| 2010-01-07 22:42 <vengfulsquirrel> That's about 80,000 records. | ||
| 2010-01-07 23:07 <vengfulsquirrel> I think it might be a bug because sales also exhibit the same behaviour. | ||
| 2010-01-07 23:13 -!- juanfer(n=juanfer@190.27.146.29) has joined #tryton | ||
| 2010-01-07 23:25 -!- cedk(n=ced@gentoo/developer/cedk) has joined #tryton | ||
| 2010-01-07 23:27 <cedk> vengfulsquirrel: I tried on sale with 160 orders and when I switch back to list view it prefetch only the 80 first sales | ||
| 2010-01-07 23:28 <vengfulsquirrel> Yeah you are right, its not getting all the sales. | ||
| 2010-01-07 23:29 <vengfulsquirrel> Is there a way I can make the switch behave the same as the initial view> | ||
| 2010-01-07 23:29 <vengfulsquirrel> *? | ||
| 2010-01-07 23:29 <cedk> vengfulsquirrel: it should do by default | ||
| 2010-01-07 23:29 <vengfulsquirrel> Yeah but do you see how it fetches different columns/fields. | ||
| 2010-01-07 23:29 <vengfulsquirrel> *? | ||
| 2010-01-07 23:30 <cedk> vengfulsquirrel: yes because now the client knows more fields of the model | ||
| 2010-01-07 23:32 <CIA-5> C?dric Krier <ced@b2ck.com> default * 1479:10a10dc4e869 tryton/tryton/gui/window/view_form/view/form_gtk/one2many.py: Use the same behavior for F3 on one2many even if add_remove is set. | ||
| 2010-01-07 23:32 <CIA-5> http://hg.tryton.org/tryton/rev/10a10dc4e869 | ||
| 2010-01-07 23:35 <cedk> vengfulsquirrel: but it must not affect the speed | ||
| 2010-01-07 23:36 <cedk> vengfulsquirrel: except if you have function fields that doesn't scale | ||
| 2010-01-07 23:42 <vengfulsquirrel> How will I know if they don't scale? | ||
| 2010-01-07 23:43 <cedk> vengfulsquirrel: by algorithm analysis | ||
| 2010-01-07 23:44 -!- paepke(n=paepke@p54A2A4F3.dip0.t-ipconnect.de) has joined #tryton | ||
| 2010-01-07 23:45 <cedk> vengfulsquirrel: it can be of the orders: O(1), O(log(n)), O(n), O(n²) etc | ||
| 2010-01-07 23:46 <CIA-5> C?dric Krier <ced@b2ck.com> default * 2127:b3a76ce4635f trytond/bin/trytond: Fix profile mode and use --profile argv | ||
| 2010-01-07 23:46 <CIA-5> http://hg.tryton.org/trytond/rev/b3a76ce4635f | ||
| 2010-01-07 23:46 <CIA-5> C?dric Krier <ced@b2ck.com> default * 1480:32b0c400fb98 tryton/ (bin/tryton tryton/client.py tryton/config.py): Add profile mode and use --profile argv | ||
| 2010-01-07 23:46 <CIA-5> http://hg.tryton.org/tryton/rev/32b0c400fb98 | ||
| 2010-01-07 23:47 <vengfulsquirrel> Yeah I had a check_ method in _constraints, but that only applies when saving the record right ? | ||
| 2010-01-07 23:47 <vengfulsquirrel> I have get_uom and get_digits and those are the only two Function fields in the Inventory or InventoryLine models in my code. | ||
| 2010-01-07 23:49 <CIA-5> C?dric Krier <ced@b2ck.com> default * 1481:33f3b34f0ae0 tryton/tryton/gui/main.py: | ||
| 2010-01-07 23:49 <CIA-5> Don't call sys.exit as it breaks profile | ||
| 2010-01-07 23:49 <CIA-5> See http://bugs.python.org/issue5978 | ||
| 2010-01-07 23:49 <CIA-5> http://hg.tryton.org/tryton/rev/33f3b34f0ae0 | ||
| 2010-01-07 23:49 <cedk> vengfulsquirrel: yes for check_ and _constraints | ||
| 2010-01-07 23:50 <cedk> vengfulsquirrel: you should find where the server spends time | ||
| 2010-01-07 23:50 <vengfulsquirrel> Yeah I have tried the profiler, its really hard to sift through all the stuff that's going on. | ||
| 2010-01-07 23:51 <cedk> vengfulsquirrel: first what the call from the client? | ||
| 2010-01-07 23:51 -!- vengfulsquirrel(n=ian@c-69-181-194-95.hsd1.ca.comcast.net) has left #tryton | ||
| 2010-01-07 23:51 -!- vengfulsquirrel(n=ian@c-69-181-194-95.hsd1.ca.comcast.net) has joined #tryton | ||
| 2010-01-07 23:51 <vengfulsquirrel> Sorry pressed emacs key | ||
| 2010-01-07 23:52 <vengfulsquirrel> What do you mean what is the call ? | ||
| 2010-01-07 23:53 <cedk> vengfulsquirrel: the call that the client does that take long time | ||
| 2010-01-07 23:56 <vengfulsquirrel> I only ran the profiler on the server | ||
| 2010-01-07 23:56 <vengfulsquirrel> side | ||
| 2010-01-07 23:56 <vengfulsquirrel> Should I run it on the client side? | ||
| 2010-01-07 23:56 <cedk> vengfulsquirrel: no |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!