IRC logs of #tryton for Wednesday, 2011-11-16 #tryton log beginning Wed Nov 16 00:00:01 CET 2011
-!- yangoon_rm( has left #tryton00:00
Guest51432hi, is it possible to work with product variants in tryton ¿?00:55
cedkuoiea: grasbauer is working on a module00:58
uoieacedk: how could i try it ¿?01:00
cedkuoiea: I don't think he already has something published01:01
cedkuoiea: should ask him when he will be back01:01
uoieathanx cedk01:02
sisalpI like this openerp Aeroo module :
enlightxsisalp: did you never use it?10:28
sisalpdo you mean we have the same on Tryton ?10:29
sisalpI thought we could not add a new report without a piece of code10:30
sisalpI just modified existing ones so far, directly editing with free-office10:30
enlightxsisalp: that addon is based on the tryton reporting engine11:03
enlightxi'm the original developer, then alistek made a lot of improvements and called it aeroo (it was report_openoffice)11:04
sisalpenlightx: so what is the status on tryton ?11:17
enlightxsisalp: have a look at the default tryton's templates11:18
sisalpenlightx: the demo is about adding a report from the gui11:20
enlightxsisalp: yes, i mean that the way you create reports with aeroo is pretty much the same as for tryton11:21
sisalpenlightx: My tryton server fails to start. I'll check this point when I'll bring it up. Thank you11:22
enlightxsisalp: ok, that's another story :)11:23
sisalphi, looking for instructions about client configuration file ..11:54
sisalpI miss it in the code11:54
cedksisalp: what do you mean?11:55
sisalpwhen I invoke tryton --help it indicates a -c CONFIG, --config=CONFIG11:56
sisalpI'm looking for the format of that file11:57
cedksisalp: it is in the package11:57
cedksisalp: oops no, look at the one you got in ~/.config/tryton/x.y/tryton.conf11:58
cedksisalp: indeed, the configuration file is managed by the client GUI11:59
cedksisalp: I don't think it must be edited by hand12:00
sisalpcedk: so you don't recommand to create the config file and start tryton on it, as I use to do ?12:01
sisalpcedk: my concern is about starting several tryton guis12:02
cedksisalp: the purpose is to just use an other place12:02
cedksisalp: it is not really an issue to run many with the same config file12:03
sisalpcedk: yes I can survive ;-) but can I point different servers per default if I don't set different config files ?12:04
sisalpcedk: and also can I run different version of tryton with adequate servers list ?12:05
nicoesisalp: the connection manager stores the tryton server it knows about in (°> Nicolas Évrard12:13
nicoe( ) Liège `¯12:13
nicoeThis is the path to the file : .config/tryton/x.y/profiles.cfg12:13
sisalpcedk: nicoe: sounds clear, evereything seems thought out already. Will adapt my script. Thank you12:18
zxq9Why is this always giving me an ASCII codec error? : bgcheck_purpose = fields.Many2One('oki_base_pass.bgcheck_purpose', 'Background Check Purpose', ondelete='SET NULL')12:55
zxq9The client is telling me that I can't use unicode character 0xe3 (which doesn't make sense; I'm not using any Unicode at all anyway).12:56
zxq9The server is telling me: ERROR:sql:Wrong SQL: ALTER TABLE "oki_base_pass_oki_base_pass" ADD FOREIGN KEY ("bgcheck_purpose") REFERENCES "oki_base_pass_bgcheck_purpose" ON DELETE SET NULL12:56
nicoeI suppose the client is acting weird on the error message12:58
cedkzxq9: I guess there is a typo in foreign key or the constraint can not be applied12:58
cedkzxq9: and you should have strange coding error message12:58
nicoecheck that your error message is encoded in UTF812:59
zxq9This happens every time I try to add a Many2One field anywhere. Fresh databases too. I just can't add Many2One fields.12:59
zxq9The ASCII codec error message is what Python says if you try to use unicode, actually. I think this is a mistaken error message. The interesting one is the "wrong SQL" one, I think.13:00
cedkzxq9: the ASCII codec error is just a small issue when the client try to display the error message13:03
zxq9That makese sense. But where is the letter ã coming from, I wonder?13:06
zxq9Anyway, the real problem is this wrong SQL thing. I've checked this for typos -- and its not the only case of this problem.13:08
zxq9Ah, I think I got it.13:12
zxq9The table for the model that is being pointed to is not created yet on initial installation.13:13
zxq9If I comment out the Many2One field, do the module install, the table is created. Then if I add the Many2One table on a module update then there is no error.13:14
zxq9So you just can't put a Many2One reference to a module that is in a class coming later in the same source file.13:14
zxq9model* (not module in last sentence)13:15
cedkzxq9: of course you have wrong dependency tree13:15
zxq9No, I mean the model is a class in the same source file. Same module, same source file.13:15
cedkzxq9: same issue, the declaration order is important13:17
zxq9Does that seem a little against OOP principles?13:18
zxq9I certainly don't mind flipping the class declaration order, I'm used to it from C anyway, but I thought this was supposed to be one of the great benefits of smart-runtime interpreted OOP languages like Python?13:18
zxq9But I might be wrong. C++ is my only real OOP experience. Everything is C and Lisp.13:18
zxq9Everything else in  my experience has been C and Lisp*13:19
cedkzxq9: it is not about OOP but Database13:20
zxq9Yes. My point is, if we're using Python, shouldn't we make the OML check table creation dependencies through all module files before placing references? Since the normal habit of a Python programmer might be to add tiny classes to the end of a source file?13:21
zxq9Or... document this exception.13:22
cedkzxq9: not possible to catch all cases13:24
cedkzxq9: this is not an exception and has nothing to do with Python13:24
cedkit is the advantage of Tryton, it is not a OOP centric framework but also a database design framework13:25
zxq9I understand what you are saying -- and you're right. My point is just that maybe the typical Python programmer's habit may conflict in some small cases (like this one).13:26
zxq9And there is no documentation that this type of error will happen.13:27
cedkzxq9: ok patch is welcome13:27
zxq9I'll think of a way to do this, though it may slow the initial table creation down a little (because it will have to crawl the code twice).13:28
zxq9Ah, and the fields.Time widget!13:29
-!- Guest10062(~oscar@ has left #tryton13:31
oxcarHi, Somebody how get all modules 2.2 version with hg clone in one command?14:07
nicoeoxcar: use the hgnested extension14:07
oxcarYes, I installed but don't know how work.14:08
oxcarYes, I installed but don't know how work it.14:08
nicoeyou just enable it in your mercurial config file14:16
nicoethe use "hg nclone" to fetch the sources14:17
oxcar@nicoe: thank you very much!14:39
-!- ciupicri(~ciupicri@unaffiliated/ciupicri) has joined #tryton14:49
zxq9Can someone point me to a good example of a well implemented search feature in a module?16:06
bechamelzxq9: a search query ?16:09
zxq9Not a query, but an implemented advanced search option. How the fields are designed, which are indexed and which are not, how the form layout works, etc.16:10
zxq9I have a need to implement a few search features and haven't really found any documentation that clears this up for me yet. But I'm sure there is a good example in the code somewhere, I just don't know where.16:11
bechamelzxq9: all of this is handled by the framework, you cannot "see" it in a module16:11
zxq9Oh, I see.16:11
sisalpwhere are documented dependencies debian/ubuntu for tryton (client) ?16:17
cedksisalp: in setup.py16:17
sisalpwhy is #        "pygtk >= 2.6", commented ?16:18
sisalponly two mandatory and two optional ?16:20
cedksisalp: pygtk doesn't install correctly through setuptools, you must use the package manager16:27
cedksisalp: and yes only 2 deps16:27
sisalpcedk: I di apt-get install python-gtk2 and got version 2.2216:28
sisalpcedk: thank you16:28
sisalpother question : where are documented the log channels mentionned in tryton --help ?16:29
cedksisalp: no where16:29
sisalpI have to debug my connection which doesn't work16:29
cedksisalp: I often think we should remove it16:29
cedksisalp: -v16:30
grasbauersomeone interessed in my approach of product_variants? should i upload it to code review?16:52
cedkgrasbauer: there was someone yesterday16:53
cedkgrasbauer: anyway, I think it is something we could have in standard if it is generic enough16:54
grasbauercedk: you know my code ;)16:54
grasbauercedk: bitbucket first or directly to review?16:55
cedkgrasbauer: review is better16:55
cedkgrasbauer: I did not read it in depth :-)16:55
zxq9How do I add fields to the list of fields searched in a tree view?16:57
zxq9I thought I saw documentation about this and how to add fields to advanced search somewhere, but can't find it now.16:58
grasbauerzxq9: in 2.0: select=1 or select=216:58
zxq9Maybe this was it.16:59
zxq9How do I enable the advanced search widgets?17:00
nicoewhich version of trytond are you using ?17:00
nicoeok we redesigned the view and there is no more an "advanced search widget"17:01
zxq9I put fields select=1 and get nothing when searching for values known in those fields.17:01
zxq9So how does advanced search work now? Or is everything "just search" with no options at all?17:02
nicoejust search17:02
nicoebut way better17:02
-!- pjstevns( has left #tryton17:02
zxq9I set fields to search=1 and cannot search for known values in those fields. Only the "name" field is being queried.17:03
zxq9Hmmm.. so we must teach users search syntax. This is a good idea for programmers, not the average idiot working at an auto shop using an ERP client.17:07
zxq9Or rather, a smart person who is smart on things other than query languages.17:07
zxq9But the feature works.17:07
zxq9So thank you.17:07
nicoezxq9: we plan to add "predefined queries" to the client17:08
nicoethose queries could be created by integrators17:08
bechamelcedk: what about adding search widget imporvements in goals for 2.4 ?17:08
cedkzxq9: FYI: google in gmail use almost the same syntax17:09
nicoebechamel: +1 on adding search queries improvements17:09
zxq9cedk: FYI, a *very* low % of users even know that.17:09
cedkbechamel: but which improvements?17:09
bechamelgmail-like search box17:09
bechamelcedk: and also bokkmrks17:09
zxq9I *love* advanced query syntax, personally. I am a master of duckduckgo and google. But that is something that my mother, for example, will never understand -- and she has to use an ERP system at work every day.17:10
cedkbechamel: I find the gmail box bloated17:10
nicoecedk: +1 on the bloatitude of the gmail box17:10
bechamelcedk: I hardly see how to do it simpler17:10
nicoeit looks nice in a browser but is too much in a gtk client17:11
zxq9+1 on the term "bloatitude"17:11
cedkbechamel: bookmarks are not clear in my mind17:11
bechamelcedk: me neither :)17:11
bechamelcedk: but the bloat-box is clear17:11
cedkbechamel: so let's first define it before adding to the roadmap :-)17:11
bechamelwe can add "define the bloat-box" in the roadmap :)17:12
zxq9Does a field name have to be visible in a tree view to be queriable?17:12
cedkACTION will test search entry on my mother17:12
cedkzxq9: not visible but defined17:13
zxq9cedk: Tell her to find stuff that will require her to not just use the "name" field. You can even give her a link to the docs, but can't talk to her or coach.17:13
zxq9Ah, invisible... this is sort of workable. I just think that for a public so used to formal GUIs (which I hate, but is the expected standard) that things like "contains" or from X to Y should have a graphical selection/widget associated with it.17:15
nicoeACTION I'll try it on my girlfriend too17:15
zxq9That way a user can stab at a limited list of things on their screen to explore the features out instead of stabbing at 109+ keys on their keyboard trying to discover totally silent features.17:15
zxq9This is a problem we have had with Gnome3 design also, by the way. So I am very wary of this now.17:16
bechamelzxq9: this is the idea behind the bloat-box: to allow people discover search syntax17:17
zxq9Is there a way to turn it off after you've played with it? Or does it sit there all the time, eating screen space?17:17
zxq9I could envision a small menu that floats when a button is pressed that would build a query for you in the search line. That could be cool.17:18
zxq9Actually, that could be very cool. Why didn't I think of that 10 years ago?17:18
zxq9Oh... I never clicked anything back then and lived entirely inside vi and bash... that's probably why...17:19
bechamelzxq9: just open gmail and click on the little arrow in the search input field, you will understand17:19
zxq9I'll have to get a gmail account first.17:20
zxq9But I'll give it a whirl.17:20
nicoeACTION thinks that the bloatbox is bloated because he lives inside vim and zsh (some stuffs has changed)17:20
sisalpwhat means this message on the server Error: [('SSL routines', 'SSL23_GET_CLIENT_HELLO', 'http request')]17:20
sisalpThe client cannot connect17:21
zxq9nicoe: I have to admit, though -- vim is bloated itself, just much more quietly.17:21
cedksisalp: something wrong with SSL :-)17:22
sisalpcedk: ok, I'll turn ssl off on the server17:23
nicoezxq9: vim is a bit bloated but it hides it very well17:23
zxq9Sort of like an old girlfriend of mine.17:26
cedksisalp: perhaps bad certificate?17:27
sisalpcedk : i turned ssl off17:27
sisalpcedk; and get now : AttributeError: SimpleXMLRPCRequestHandler instance has no attribute 'path'17:28
cedksisalp: line code?17:28
sisalptrytond/protocols/", line 12517:28
cedksisalp: which Python version?17:29
sisalpPython 2.6.517:30
cedksisalp: try with the patch:
sisalpcedk: by changing localhost by the ip address for json-rpc, I got it working18:06
sisalpwith the patch in place : - - [16/Nov/2011 18:04:54] code 400, message Bad request syntax ('\x16\x03\x01\x00Z\x01\x00\x00V\x03\x01N\xc3\x18:07
cedksisalp: is it working?18:15
sisalpno it's not, still connection refused on xml-rpc, now ok on jsonrpc18:16
cedksisalp: how do you test xml-rpc?18:18
sisalpI change the port on the gtk client and wait praying18:18
cedksisalp: you was not awake during my talk?18:21
cedksisalp: tryton client speak only one protocol18:21
sisalpthe server speaks both and the client speakes only jsonrpc ?18:21
cedksisalp: yes18:22
cedksisalp: KISS18:22
sisalpyes, so if we have an application on xml-rpc, we keep both ports open18:23
cedksisalp: yes if you need it18:23
sisalpand never use xml-rpc fron the gtk client18:23
cedksisalp: yes18:24
sisalpI could have serched for some time18:24
cedksisalp: that's why you got strange error18:24
sisalpthank you18:24
cedksisalp: you should just hear at the TUL instead of talking with nicoe18:24
sisalpI did not view the total video yet18:26
cedksisalp: I saw you at the bottom near the radiator :-)18:26
sisalpthen I'll have a look at what was wrong with my ssl18:27
sisalpis there something on the video regarding ssl ,18:27
cedksisalp: no18:29
cedksisalp: but if you test it with xml-rpc, you had perhaps no ssl issue18:29
cedksisalp: bbl18:35

Generated by 2.11.0 by Marius Gedminas - find it at!