IRC logs of #tryton for Friday, 2009-05-01 #tryton log beginning Fri May 1 00:00:01 CEST 2009
-!- _TiN_(n=usuario@ has joined #tryton00:06
-!- racke( has joined #tryton00:09
CIA-48ced roundup * #1030/Check stdout and stderr with py2exe: [new] User have not necessary the right to write in install directory.00:23
-!- carlos_(n=carlos@ has joined #tryton00:26
-!- _TiN_(n=usuario@ has joined #tryton00:48
-!- opensoft(n=elkafil@ has joined #tryton01:26
-!- racke( has joined #tryton01:54
-!- racke( has joined #tryton02:32
-!- saxa_( has joined #tryton04:02
-!- yangoon( has joined #tryton05:18
-!- vengfulsquirrel( has joined #tryton07:10
-!- vengfulsquirrel( has left #tryton08:50
-!- paola( has joined #tryton09:03
-!- cedk(n=ced@gentoo/developer/cedk) has joined #tryton09:04
-!- sharkcz( has joined #tryton09:07
-!- racke( has joined #tryton09:29
-!- enlightx(n=enlightx@ has joined #tryton10:41 * r603 /wiki/
CIA-48Add GroupDAV11:13
CIA-48Add tablespace11:13
CIA-48Add BETWEEN function11:13
-!- sharkcz( has joined #tryton11:22
-!- cedk(n=ced@gentoo/developer/cedk) has joined #tryton11:59
-!- sharkcz( has joined #tryton12:02
udonocedk: hi12:17
udonocedk: about inheritance again12:17
udonocedk: dbrenk told me that tinyerp has some distingtion on _inherit=''12:18
udonocedk: 1. if _name == _inherit12:18
udonocedk: 2. if _name != _inherit12:19
udonoI don't know for now how the second works, and if it work well.12:21
udonoBut it seems that we removed this inheritance with the collapse of _name and _inherit to soly _name12:23
cedkudono: We can do the same with a simple python inherit12:51
cedkudono: look at ir.sequence and ir.sequence.strict12:52
cedkudono: and the OpenERP implementation, there is some indetermination when you mix the both inheritance12:53
cedkwhat happend if an other module inherit simply the object other.material12:53
cedkoops I mean network.material12:53
cedkwhat do it happen to other.material12:54
cedkI guess it depends of the order of the module which is not determinist12:54
cedkwith our implementation, it is always the basic model that is used12:55
cedkudono: and I don't find any use of the second in OpenERP base modules12:58
cedkudono: by the way, I put some patches for pywebdav on
cedkudono: it is with davfs2 that you have issue?13:02
udonocedk: yes13:08
udonocedk: should I go deeper into introspection now?13:08
cedkudono: I have an issue with a missing group davfs213:10
udonocedk: do you recognize the slowness of davfs2?13:11
cedkudono: not yet working13:12
udonocedk: as a test collection I use ir.translation13:14
udonocedk: with firefox it needed ~3s to load all entries13:15
cedkudono: this patch must fix a part
cedkudono: and translations is very too big for pywebdav and I think also for davfs213:20
udonocedk: can not download the patch, 403 - Forbidden13:22
udonocedk: with firefox it takes 3s to load13:22
cedkudono: download fixed13:23
udonocedk: works13:23
cedkudono: davfs2 make some cache at first time, so it downloads all collection at depth 113:24
cedkudono: and pywebdav uses minidom to create the xml and Tryton use string concatenation to create web page13:24
udonocedk: yes, but last time I debugged, davfs2 retries 3times to load something, after this it works. DAVFS said something like wrong header.13:32
cedkudono: I think it is because pywebdav on HEAD for non-existing url return a 40413:53
cedkudono: it doesn't seem to be that14:00
cedkudono: it looks like if davfs2 is bugged14:00
udonocedk: yes, possible14:10
cedkudono: by the way, where did you find log from davfs214:15
udonocedk: you need to tweak /etc/davfs2/davfs2.conf14:17
udonocedk: man davfs2.conf14:17
udonoIt is stores in syslog or messages14:17
cedkudono: strange davfs2 makes PUT with empty body14:17
udonohere is a complete log
udonowith all debug modes14:21
cedkudono: ok, I think I find one issue with "100-continue" which is not handle by pywebdav or BaseHTTPRequest14:41
cedkIt works better when setting use_expect100 to 0 in davfs2.conf14:53
udonocedk: what do you understand with 'better'?15:20
udonocedk: what strange is, my trytond process is getting bigger and bigger15:23
udonocedk: I do nothing, but it grows 295mb + 2mb/second15:24
udonomaybe it is the caching of davfs2...15:25
cedkudono: it doesn't work if you enable expect10015:26
udonocedk: ok15:27
udonothe growing trytond seems to depend on davfs215:35
-!- Odd_Bloke( has left #tryton15:36
udonowith firefox(cache disabled) there is not this huge growing. After displaying the contents, the size of trytond shrinks small15:36
cedkudono: I guess it is the minidom15:41
udonocedk: did you have entries in /var/cache/davfs2/<CONNECTION>/index ?15:45
cedkudono: not for index15:46
udonocedk: you have other entreis as index?15:46
cedkudono: yes, it is the cache of davfs15:47
udonomine is empty15:47
udonoBut I only have the collection ir.translation15:47
udonocedk: is davfs2 in general much slower then firefox on your system?15:50
cedkudono: yes but firefox idoesn't use WebDAV15:50
cedkudono: I think we should remove minidom from pywebdav and use lxml15:55
cedkudono: I will see with pywebdav devs if they will accept the change15:58
udonocedk: I find out that when I decrease the read_timeout  to  15, then it is more quick...strange16:06
CIA-48C?dric Krier <> default * 1783:e96a04849390 trytond/trytond/ir/ Skip set_datas on attachment only if value is False or None16:07
CIA-48C?dric Krier <> default * 1784:c99c3c4ff4ae trytond/trytond/webdav/
CIA-48Fix _uri2object when there is more than one attachment on the same record16:07
CIA-48Fix uninitialized variables16:07
cedkudono: if you still use translations, it is normal16:10
udonono, this time I use just a refresh of the dir list in my dav root16:11
cedkudono: but if you have anyweher a collection with all translations, this is still normal16:11
udonocedk: yes, you are right, I destroyed the ir.translation collection and now it works better16:21
udonocedk: for me it seems to work with smaller (<2900) collections16:42
-!- enlightx( has joined #tryton17:21
cedkudono: I have put a pywebdav_expect100.patch to hanle expect 100-continue17:38
-!- cedk(n=ced@gentoo/developer/cedk) has joined #tryton18:14
-!- vengfulsquirrel( has joined #tryton20:37
cedkudono: I made a test by replacing minidom with lxml by it is still too slow20:45
cedkI will try with a string concataination like I do for the html page20:46
-!- vengfulsquirrel( has joined #tryton20:53
-!- ChanServ(ChanServ@services.) has joined #tryton21:14
-!- carlos(n=carlos@ has joined #tryton22:05
-!- sharkcz_( has joined #tryton22:21
udonocedk: I will test tomorrow23:24
cedkudono: I just test with string concat. but it looks the same23:34
cedkudono: so I think it is a limitation of protocol23:35
-!- cedk_( has joined #tryton23:53

Generated by 2.11.0 by Marius Gedminas - find it at!