IRC logs of #tryton for Wednesday, 2011-04-20 #tryton log beginning Wed Apr 20 00:00:06 CEST 2011
-!- elbenfreund(~elbenfreu@ has joined #tryton00:06
-!- albertca(~albertca@ has joined #tryton00:12
-!- albertca(~albertca@ has left #tryton00:23
-!- rhubner(~rhubner@ has joined #tryton00:25
-!- elbenfreund(~elbenfreu@ has joined #tryton00:25
-!- elbenfreund1(~elbenfreu@ has joined #tryton00:34
-!- chrue1( has joined #tryton02:09
-!- Cazou(~Cazou@ has joined #tryton02:44
-!- alimon(~alimon@ has joined #tryton03:17
-!- Hyper-Core(~lol@ has joined #tryton04:49
-!- Hyper-Core(~lol@ has left #tryton05:00
-!- yangoon( has joined #tryton05:19
-!- alimon(~alimon@ has joined #tryton05:28
-!- alimon(~alimon@ has joined #tryton05:38
-!- silverfox1971(~sysadmin0@ has joined #tryton06:11
-!- gremly(~gremly@ has joined #tryton06:33
-!- predatell(~predatell@ has joined #tryton07:04
-!- Vladimirek( has joined #tryton07:50
-!- okko(~okko@ has joined #tryton08:42
-!- pjstevns( has joined #tryton08:56
-!- okko(~okko@ has joined #tryton09:23
-!- predatell(~predatell@ has joined #tryton10:06
-!- enlightx( has joined #tryton10:06
-!- cedk(~ced@gentoo/developer/cedk) has joined #tryton10:11
-!- bechamel( has joined #tryton10:14
-!- predatell(~predatell@ has joined #tryton11:01
-!- reichlich( has joined #tryton11:40
-!- blast_hardcheese( has joined #tryton11:50
-!- predatell(~predatell@ has joined #tryton11:51
-!- predatell(~predatell@ has joined #tryton11:58
-!- predatell(~predatell@ has joined #tryton12:05
-!- Vladimirek( has joined #tryton12:05
-!- nicoe( has joined #tryton12:11
-!- Cazou(~Cazou@ has joined #tryton12:56
-!- paepke( has joined #tryton13:17
-!- silverfox1971(~sysadmin0@ has left #tryton13:33
-!- predatell(~predatell@ has joined #tryton14:32
-!- alimon(~alimon@ has joined #tryton14:49
-!- okko(~okko@ has joined #tryton14:53
-!- paepke( has joined #tryton15:41
-!- alimon(~alimon@ has joined #tryton15:52
-!- yangoon( has joined #tryton16:43
-!- trifon( has joined #tryton17:22
-!- trifon( has joined #tryton17:40
sharooncedk: can you explain what you meant by your last tweet ?18:19
cedksharoon: there is index in mongodb ?18:22
sharooncedk: yes18:22
cedksharoon: so it will be great to be able to define it in the field18:22
cedksharoon: like we have select=1 in SQL18:22
sharooncedk: oh yes, resuse select = 1 ?18:22
sharooncedk: yes18:22
sharooncedk: adding a ticket for that18:23
cedksharoon: otherwise your search will not be efficient18:23
sharooncedk: yes18:24
cedksharoon: and you do that because you have a lot of products :-)18:24
sharooncedk: oh yes i have over a million of them :D18:24
cedksharoon: next step is to find a way to edit Mongo field in the GTK client :-)18:25
sharooncedk: its already there18:25
cedksharoon: where?18:26
cedksharoon: wait...18:26
sharooncedk: like MongoAttribute is dislayed using default widget18:26
cedksharoon: ok I understand18:26
sharooncedk: but yes MongoDocument is stillr epresented as a JSON18:26
sharoonACTION out for lunch, brb18:28
cedksharoon: do you think this kind of fields could be added dynamicly?18:29
-!- gremly(~gremly@ has joined #tryton18:32
-!- chrue( has joined #tryton19:07
-!- Vladimirek( has joined #tryton19:18
sharooncedk: it could be a great feature but not sure how we could do that19:23
-!- alimon(~alimon@ has joined #tryton19:24
cedknicoe: could you check this codereview
-!- curlynostrill(~curlynost@ has joined #tryton20:44
-!- okko( has joined #tryton20:45
-!- plantian( has joined #tryton20:46
-!- paepke( has joined #tryton20:57
-!- elbenfreund( has joined #tryton20:59
-!- trifon( has joined #tryton21:10
-!- rhubner(~rhubner@ has joined #tryton21:33
-!- rhubner(~rhubner@ has left #tryton21:33
-!- dfamorato( has joined #tryton21:33
-!- okko( has joined #tryton22:32
-!- nicoe_( has joined #tryton22:41
sharooncedk: I  have some performance improvements to make, its too naive like __slots__ optimisation for some classes like the etc22:42
sharooncedk: is there still time for me to propose a patch ?22:42
cedksharoon: I'm not sure it will improve something23:06
cedksharoon: have you benchmark?23:08
cedksharoon: slots are there to limit memory usage23:11
cedksharoon: but for Transaction, there is only one instance per request23:11
cedkso the gain is very limited23:11
-!- bechamel( has joined #tryton23:13
cedksharoon: for info:
sharooncedk: any issues ?23:19
cedksharoon: ???23:20
sharooncedk: got it now23:21
cedksharoon: it is bad practice to optimise stuff that doesn't need23:21
sharooncedk: i thought it was getting intialised too many times let me recheck a profile count i got23:22
cedksharoon: if you are concern, you can have a look at
cedksharoon: and we should really get your read with tuple for 2.223:24
sharooncedk: thats something we got to discuss23:25
sharooncedk: i was doing some experiments and these are my findings23:25
sharooncedk: 1. creation of dict is slow23:25
sharooncedk: 2. but pickling of tuple, list and dict are fastest and custom objects including named tuples dont match it23:26
cedksharoon: what do you want to discuss?23:27
sharooncedk: 3. when compression is used the dict's repeating keys are removed and hence the compression ensures there is not much overhead just because we use dict23:27
sharooncedk: read with tuple benefits23:27
cedksharoon: I'm convinced that we must go to tuple23:29
sharooncedk: ok23:29
sharooncedk: and returna  list of tuple with the first tuple as column headers ?23:30
cedksharoon: no header23:30
sharooncedk: ok23:30
cedksharoon: the header is the list you give23:30
sharooncedk: we automatically insert id23:31
cedksharoon: we will remove this feature :-)23:31
sharooncedk: :D you get what you ask for and in that case fields_names must be a required argument23:32
cedksharoon: yes23:32
cedksharoon: and we must ensure that the order is the same as the list of ids23:32
sharooncedk: On a separate note what do you think of __slots__ optimisation of browse node ?23:32
sharooncedk: agree23:33
cedksharoon: I think it is a waste of time because we will merge Browse with Model23:33
cedksharoon: and slots will be a no way23:34
sharooncedk: Active Record pattern23:34
cedksharoon: yes it is on the roadmap23:34
cedkso much to do, so little time23:35
sharooncedk: but you do realise that an insatnce of a model is gonna be huge right23:35
sharooncedk: i guess it may be better to have browse record is a descriptor23:36
cedksharoon: why?23:36
sharooncedk: will each record not represent an instance of the Model ?23:36
cedksharoon: yes23:37
sharooncedk: "so much to do, so little time" => "call for a sprint of x days"23:37
sharooncedk: for example an instance of ModelSQL (eq to pool.get('x')) alone has over 100 attributes in methods etc, its huge23:38
cedksharoon: the attributes are not on the instance23:39
sharooncedk: ok ?23:40
cedksharoon: it will not be very different then what we have now23:40
cedksharoon: it will be just more Pythonic23:40
sharooncedk: ok23:40
cedksharoon: I don't think for now that memory consumption is an issue for Tryton23:41
sharooncedk:  a profile log23:42
cedksharoon: I could say that __search_domain_calc need some love23:42
cedksharoon: but it will be improved with pysql23:43
-!- blast_hardcheese(~blast_har@ has joined #tryton23:45

Generated by 2.11.0 by Marius Gedminas - find it at!