IRC logs of #tryton for Saturday, 2008-09-20 #tryton log beginning Sat Sep 20 00:00:01 CEST 2008
2008-09-20 01:26 <CIA-52> tryton: C?dric Krier <> default * 989:19722c2067a4 trytond/trytond/ Fix ssl socket to have the same shutdown call as normal socket for issue386
2008-09-20 01:26 <CIA-52> tryton: ced roundup * #386/server secure: [resolved] Must be fixed with changeset 19722c2067a4
2008-09-20 05:19 -!- yangoon( has joined #tryton
2008-09-20 09:21 -!- Timitos(n=Timitos@ has joined #tryton
2008-09-20 10:26 -!- cedk(n=ced@gentoo/developer/cedk) has joined #tryton
2008-09-20 12:00 <CIA-52> tryton: ced roundup * #386/server secure: [chatting] The XML-RPC with ssl has low performance, it is better to use a ssl proxy.
2008-09-20 15:30 -!- b52laptop(n=b52lapto@ has joined #tryton
2008-09-20 16:40 -!- cedk(n=ced@gentoo/developer/cedk) has joined #tryton
2008-09-20 17:46 <cedk> udono: you can find the vatnumber repository here:
2008-09-20 17:47 <cedk> udono: and the web site:
2008-09-20 17:47 <cedk> udono: I use the GPL-3 for this module
2008-09-20 17:47 <cedk> I have added some test script
2008-09-20 17:48 <cedk> udono: can you check if it is good for you? After that I will make the change in relationship module
2008-09-20 17:54 <udono> cedk: some remarks ;-)
2008-09-20 17:54 <cedk> by the way, if german people can also validate the numword module, I will be able to make a release
2008-09-20 17:55 <udono> cedk: for the I find it better to use a dictionary {"VAT Country": "VATNumber"}
2008-09-20 17:56 <cedk> udono: but I was thinking also add test for wrong vat numbers
2008-09-20 17:56 <udono> cedk: the actual numword isn't good for germans, but we dont know if its good for spain, too. So you can release when you like, and I contribute my code later
2008-09-20 17:57 <udono> cedk: yes, then use simply a second dictionary and two test routines one for right, one for wrong numbers
2008-09-20 17:58 <cedk> udono: ok, I will change
2008-09-20 17:58 <cedk> udono: for numword, I have time so if it is possible to have a right german it will be better
2008-09-20 18:01 <udono> cedk: I find it a good attitude to put contributors into the COPYRIGHT file, because every contribution is part of the release and every contributer is a legal copyright holder for his part of the release code. Another thing is, that the contributions are unter the Authorship of the contributor, and not the first author.
2008-09-20 18:03 <cedk> udono: in fact like that the COPYRIGHT is shared by every authors
2008-09-20 18:03 <udono> cedk: yes
2008-09-20 18:19 <cedk> udono: tests refactorized
2008-09-20 18:20 <udono> cedk: if there is code stolen from tryton, than its your and bechamels job alone (because of authorship) to follow up on this. Even if the stolen code isn't from yourself. So if you include all contributors as authors, then its the job for everyone of them. Another point is that you "demand" the sole authorship for an Open Source project. In fact that means - all in the context of a pure b2b software (like tryton is) - another company
2008-09-20 18:20 <udono> e code to tryton. I've seen a lot of problems and caveats on this in other OS b2b projects, like SQL-Ledger, LX-Office and even at TinyERP. Why not have a large author page like Mozilla does?!
2008-09-20 18:20 <udono> cedk: great
2008-09-20 18:21 <udono> cedk: in there is missing the vat_checker method
2008-09-20 18:23 <cedk> udono: ok, we have no problems to add people that contributes to the COPYRIGHT file
2008-09-20 18:23 <cedk> udono: but they must have provide enough code, not just one patch with single line
2008-09-20 18:25 <cedk> udono: what do you mean by vat_checker ?
2008-09-20 18:25 <udono> cedk: to have a lot of contributors is a quality characteristic of an Open Source project. I would put them in et all.
2008-09-20 18:26 <cedk> udono: I'm more for this: we add people if they ask for
2008-09-20 18:26 <udono> cedk: Thats the right way
2008-09-20 18:27 <cedk> udono: and by the way, as we work with patches people who contribute have to patch also the COPYRIGHT file
2008-09-20 18:28 <udono> cedk: this is an even better idea, so we do not need resources for controling who, where, when...
2008-09-20 18:28 <cedk> udono: and all is recorded in mecurial
2008-09-20 18:29 <udono> cedk: yes, but this is not the same then the COPYRIGHT file...
2008-09-20 18:30 <udono> cedk: but mercurial is important for infringements
2008-09-20 18:32 <udono> cedk: about vat_checker, my fault. Its a special tryton function, because of handling vatnumbers like we do, seperated in vat_country and vat_number...
2008-09-20 19:07 <yangoon> cedk: ping, ar eyou around?
2008-09-20 19:09 <udono> I ask bkuhn for handling contributors and authors...
2008-09-20 19:18 -!- bkuhn(n=bkuhn@sflc/staff/conservancy.president.bkuhn) has joined #tryton
2008-09-20 19:18 <bkuhn> udono: we can talk here if you want.
2008-09-20 19:19 <udono> bkuhn: yes, thanks
2008-09-20 19:19 <udono> bkuhn: we are talking about copyright, authors, contributors in a b2b software like tryton...
2008-09-20 19:21 <udono> bkuhn: what is the difference between them in legal aspects?
2008-09-20 19:21 <udono> bkuhn: is every contributor an author for his part of code...
2008-09-20 19:21 <bkuhn> udono: Typically, yes, everyone holds copyright on the code he/she contributes.
2008-09-20 19:22 <udono> bkuhn: is it a good way to put all contributors in a long list inside the copyright file?
2008-09-20 19:22 <bkuhn> udono: there are basically two good ways to handle what is called "copyright inventory".
2008-09-20 19:22 <bkuhn> There is the "file-by-file method" or the "top-level copyright file method".
2008-09-20 19:23 <bkuhn> Most FLOSS projects used file-by-file, which means each time a contributor adds code to a file, they add their copyright notice in just that file.
2008-09-20 19:23 <udono> bkuhn: we choose the top-level one
2008-09-20 19:23 <bkuhn> Ah, so if you do top level, each time someone contributes anything to the project, then just add their notice to the top level file COPYRIGHT.
2008-09-20 19:23 <bkuhn> Each person would have an entry like:
2008-09-20 19:23 <bkuhn> Copyright (C) YEAR1, YEAR2, ... Bradley M. Kuhn
2008-09-20 19:23 <bkuhn> Then, in each file in the project ,you should have a notice like:
2008-09-20 19:24 <bkuhn> "This file is part of the Tryton project. Copyright information can be found at the top level file in the distribution COPYRIGHT. The project is licensed under GPLv2-or-later; details on the license are in the top level file in the distribution, LICENSE"
2008-09-20 19:24 <bkuhn> then in LICENSE, you should put the the info about the license.
2008-09-20 19:25 <bkuhn> usually it will say something like the standard notice that most people put in each file (that is found in the how-to-apply section of the GPL)
2008-09-20 19:25 <bkuhn> usually, I leave the full text of the GPL in a file called GPLv2 or GPLv3
2008-09-20 19:25 <bkuhn> and LICENSE has the general license notice
2008-09-20 19:25 <bkuhn> there are other ways to do that correctly; this is just my preference.
2008-09-20 19:26 <bkuhn> The most important point is you should put SOME statement at the top of each file, even if you use top-level method
2008-09-20 19:26 <bkuhn> just so that when people get one file, they know they are missing something they must fine
2008-09-20 19:26 <bkuhn> s/fine$/find$/
2008-09-20 19:27 <udono> bkuhn: so if the project gets infringed every contributor may follow up on this?
2008-09-20 19:28 <udono> bkuhn: we put the copyright information on top of each file...
2008-09-20 19:34 <bkuhn> well, then you are choosing the "file-by-file method"
2008-09-20 19:34 <bkuhn> Note that I'm talking merely of inventory systems... systems for keeping track of who holds copyright.
2008-09-20 19:35 <bkuhn> As for infringement, that's a different matter and mostly orthogonal to the question of how you keep your iventories.
2008-09-20 19:35 <udono> bkuhn: no, we chose thoe on top method, but with remarks on top of every file, sorry for unclearness
2008-09-20 19:35 <udono> r/thoe/the/
2008-09-20 19:37 <udono> bkuhn: I meant with my question, if someone has "stolen" code from the project is it just "stolen" from the contributions made by a single contributor, or is it "stolen" from the whole project. Is it "stolen" from all contributors, and is it possible for all contributors, to clear the situation with the "thief"
2008-09-20 19:38 -!- ikks(n=igor@ has joined #tryton
2008-09-20 19:39 <bkuhn> udono: I don't much like the "stealing" analogy.
2008-09-20 19:39 <bkuhn> The question is: if someone infringes the copyright, whose copyrights are they infringing.
2008-09-20 19:39 <bkuhn> It depends on the details.
2008-09-20 19:40 <bkuhn> If you have done file-by-file method (i.e., each file has individual cpyright notices at the top), then it's hard to say that they are infringing anyone's copyrights but those who have contributed to the files in question.
2008-09-20 19:40 <bkuhn> However, there are exceptions and complications for various reasons.
2008-09-20 19:40 <bkuhn> You can't know for sure until the details are in front of you.
2008-09-20 19:42 <udono> bkuhn: ok, I understand.
2008-09-20 19:45 <udono> bkuhn: Thanks a lot for the tips.
2008-09-20 19:45 <udono> s/tip/hint/
2008-09-20 19:45 <bkuhn> Sure, no problem.
2008-09-20 19:56 -!- ikks(n=igor@ has joined #tryton
2008-09-20 20:23 <cedk> yangoon: yes
2008-09-20 20:25 <yangoon> cedk: there seemed to be an installation problem, but like often, it was a cache problem
2008-09-20 20:26 <cedk> yangoon: what cache problem?
2008-09-20 20:26 <yangoon> cedk: the server has definitely to be stopped before doing an installation with -i and afterwards restarted
2008-09-20 20:27 <cedk> yangoon: yes of course
2008-09-20 20:27 <yangoon> cedk: it is not that clear to me, why it has to be done in this scenario:
2008-09-20 20:28 <cedk> yangoon: because server need to instanciate the object at startup
2008-09-20 20:28 <yangoon> cedk: I have running the server as usually as a service
2008-09-20 20:28 <cedk> yangoon: depending of the module installed
2008-09-20 20:28 <cedk> if you install module without restart the server, the server will not have the right objects
2008-09-20 20:28 <yangoon> but server has already all objects
2008-09-20 20:29 <cedk> no, it has only instanciate object from installed modules
2008-09-20 20:29 <yangoon> I will say sjhortly what I Have done
2008-09-20 20:31 <yangoon> AFAIS server inits all modules it finds in the appropraiate directory
2008-09-20 20:31 <yangoon> ok, I have the server running, I create new DB from the client
2008-09-20 20:32 <yangoon> I stop the client, run another instance of trytond with -i -d new
2008-09-20 20:32 <yangoon> Iconnect the client to the new database, and the error was there
2008-09-20 20:33 <yangoon> what I don't understand: server instance 1 had excactly same base as instance when it was started
2008-09-20 20:33 <cedk> yangoon: yes that is normal because the server instanciate only objects from installed modules
2008-09-20 20:33 <cedk> yangoon: and keep it in a pooler
2008-09-20 20:34 <cedk> yangoon: so the pooler need to be restart
2008-09-20 20:34 <yangoon> but how does it do when client connect to different DBs with different modules?
2008-09-20 20:34 <cedk> yangoon: there is a pooler by database
2008-09-20 20:35 <cedk> yangoon: one things that can be done, is to handle a signal to restart poolers
2008-09-20 20:35 <yangoon> cedk: and this pooler "knows" last state of db?
2008-09-20 20:36 <cedk> yangoon: no
2008-09-20 20:36 <yangoon> cedk: so if another instance changes something in modules installation, pooler gets confused?
2008-09-20 20:36 <cedk> yangoon: yes
2008-09-20 20:36 <cedk> yangoon: it will cost too much to track this
2008-09-20 20:37 <cedk> and that is something that must never happen in a real installation
2008-09-20 20:37 <yangoon> I will try to document this, because I think (at least) in development environments it is a mistake done quick and likely often
2008-09-20 20:38 <cedk> I can handle the SIGUSR1 to force the restart of the poolers
2008-09-20 20:39 <yangoon> and how do you know, when you have to restart poolres?
2008-09-20 20:40 <cedk> yangoon: when the server will receive SIGUSR1
2008-09-20 20:41 <yangoon> when does it get SIGUSR1? Who does send it?
2008-09-20 20:42 <cedk> yangoon: the user of course
2008-09-20 20:44 <yangoon> whether he has to restart or just to reload, where is the difference?
2008-09-20 20:45 <cedk> yangoon: it is just faster to reload the pooler than to restart
2008-09-20 20:45 <cedk> yangoon: and the user connection is not lost
2008-09-20 20:46 <yangoon> ok, then it is an advantage, but server usually starts very fast
2008-09-20 20:47 <cedk> yangoon: it will depend of the number of modules
2008-09-20 20:47 <cedk> yangoon: but of course it is just a shortcut
2008-09-20 22:23 -!- b52laptop(n=b52lapto@ has joined #tryton

Generated by 2.17.3 by Marius Gedminas - find it at!