IRC logs of #tryton for Monday, 2013-01-07 #tryton log beginning Mon Jan 7 00:00:01 CET 2013
-!- mr_amit(~amit@ has left #tryton12:00
zodmancedk, i recived a admin perms from codereview, are you need something ?15:07
cedkzodman: I give you viewer permission because for some user google doesn't send email except if they are users15:15
sharoonthomascedk: based on the fixture added to doctests when they are added to test suites, it seems like doc tests only work with SQLite16:32
sharoonthomascedk: is that correct ? example:
cedksharoonthomas: no16:34
sharoonthomascedk: are databases created and dropped before each test suite is run ? or is it for each test case ? where is this handled ?16:36
cedksharoonthomas: it depends16:39
cedksharoonthomas: all in one DB except for scenario16:39
sharoonthomascedk: ok16:40
sharoonthomascedk: my problem is I am trying to clean up the test cases of nereid as you suggested to use the existing tryton testing framework which depends on unittest16:40
sharoonthomascedk: the nereid tests exploit the use of class based test fixtures introduced in unittest2 (std unit test in python 2.7+)16:41
sharoonthomascedk: these class based fixtures create and drop the database before and after each test16:41
sharoonthomascedk: what is your suggestion to get around this ?16:41
cedksharoonthomas: it is too slow16:41
cedksharoonthomas: make test transactional and rollback after each test16:42
sharoonthomascedk: well, that is the root of the problem. you cannot rollback because web requests cannot be tested as the tests are written outside the dispatcher which commits code on successful completion of a request16:43
sharoonthomascedk: for example, a POST to the registration page is tested by checking if it created a user record in the database16:43
cedksharoonthomas: skip this layout16:44
cedksharoonthomas: skip this layer16:44
sharoonthomascedk: checking ...16:45
PilouWith postgresql, if the database creation takes a great time, the database can be created once and next used as a template database.16:45
-!- calbasi( has left #tryton16:46
sharoonthomasPilou: thats an interesting approach, but would again require class based testing fixtures16:49
sharoonthomascedk: looks like a possible solution and better solution. Its a level deeper and I guess integration tests should be left elsewhere.16:55
sharoonthomascedk: nereid tests so far seem to be a blend of unit and integration tests16:55
shomonI'm in the translations area in tryton and I only see english translations! What am I missing?17:45
shomonanother question: how do you change the word "party" - the title of the party menu... I want to translate it18:08
cedkshomon: it is the name of the action linked to the menu entry18:23
shomonah found it, thanks cedk!18:39
plantianhi guys, my upgrade of tryton and been postponed repeatedly but now I have time to try and upgrade, i have custom modules and existing database, is the best approach to use --update=all in each successive stable trytond version?23:27
plantianI also see this --  -- I hand apply all these after update in 2.0 version but before 2.2 --update=all right?23:27
cedkplantian: normally you should be able to upgrade directly to the last one23:28
plantiancedk: Okay, so I update all code and xml in modules, then load 1.8 database with psql, then I can just upgrade in 2.6 code base?23:30
plantianWhat about database changes, like the one listed here with regard to changes to column defaults:
plantianIt seems that if my modules created integers, floats or numeric columns I would have to fix this by hand.23:33
cedkplantian: it should not be a real issue23:33
cedkplantian: but you can remove the default if you want23:33
cedkplantian: it only matters if you had such field with required=True because now 0 will be allowed23:34
plantiancedk: Okay, I will have to check all my modules.23:37
cedkplantian: yes there is a lot of change in the API over 4 versions23:39
plantiancedk: Has a process been developed for building mac client for development versions of trytond?23:41
cedkplantian: especially the Active Record, new workflow, new wizard23:41
plantiancedk: yes, new wizard will be a big problem to update, but I'm hoping it will be a lot better than the old wizard23:42
cedkplantian: no, I just build Mac client per release23:42
Pilouabout mac building, you could follow (using updated versions of mentionned software): it works23:42
plantianPilou: Yes problem is I only have 2 users that use mac client and no development machine.23:43
cedkplantian: then you don't need of trunk build23:43
plantiancedk: Typically I have had to deploy trunk version to get fixes.23:44
cedkPilou: by the way, we should not wait too long to include GIS patch23:44
cedkPilou: otherwise it will be too close of the next release23:44
plantianHopefully not this time, but it is problematic if I need trunk version fixes but there is no way to build client.23:44
cedkplantian: fix happens in released series also23:45
Piloucedk: when starts the freeze ?23:45
cedkPilou: 1 month before the release but big patch should be there long before23:46
cedkACTION forgot to choose the next release date23:48
plantiancedk: I was just looking to see where that was.23:48
plantiancedk: The date for 2.8 right?23:48
cedkplantian: done23:49
cedkplantian: I have to go, bye23:50
plantiancedk: Okay thanks.23:50
plantianapril 22nd, I guess23:51

Generated by 2.11.0 by Marius Gedminas - find it at!