chat.freenode.net #tryton log beginning Wed Dec 10 00:00:01 CET 2008 | ||
2008-12-10 00:00 <cedk> X0d_of_N0d: but as it will be generally binary document that will be attached like .doc, .jpgeg or .odt | ||
2008-12-10 00:00 <cedk> X0d_of_N0d: mercurial will also store copies instead of diff | ||
2008-12-10 00:00 <cedk> X0d_of_N0d: by the way webdav has many extentions: versioning, calendaring, groupware, ... | ||
2008-12-10 00:03 <X0d_of_N0d> cedk: The guy I work with worked on setting up calendaring. The extension to the protocal is there, but when we looked at it it wasn't really implimented very well in anything | ||
2008-12-10 00:06 <bechamel> X0d_of_N0d: another solution is just to generate a ics file that calendar app can load (like google calendar does) | ||
2008-12-10 00:06 <cedk> X0d_of_N0d: you try with wich library? | ||
2008-12-10 00:07 <X0d_of_N0d> bechamel: yeah, but that isn't really a good solution for shared calendars that need multiple users | ||
2008-12-10 00:08 <X0d_of_N0d> Aaron used a hacked up version of darwin cal server | ||
2008-12-10 00:08 <X0d_of_N0d> apple guys write some strange python | ||
2008-12-10 00:08 <X0d_of_N0d> it's all xml based... kind of does some strange stuff | ||
2008-12-10 00:08 <cedk> X0d_of_N0d: what was not working? | ||
2008-12-10 00:09 <X0d_of_N0d> But then again, doing shared calendars is actually kind of a hard thing to do, especially with the way people (mis)use them. | ||
2008-12-10 00:10 <cedk> X0d_of_N0d: we will make a backend with Tryton models, so the WEBDAV will be just a protocol to access it | ||
2008-12-10 00:10 <cedk> it seems that CALDAV is the standard for calendar | ||
2008-12-10 00:11 <cedk> iCal, Thunderbird, Sunbird, Evolution support it | ||
2008-12-10 00:11 <cedk> and also google calendar | ||
2008-12-10 00:11 <X0d_of_N0d> yeah, caldav.......... | ||
2008-12-10 00:12 <cedk> and it seems that outlook can be configured to support it | ||
2008-12-10 00:13 <X0d_of_N0d> yeah, caldav is what darwin calendar server uses.... | ||
2008-12-10 00:13 <X0d_of_N0d> so every time you change your month it queries the calendar server | ||
2008-12-10 00:13 <X0d_of_N0d> with darwin cal server it spikes your cpu up to about 100% for a second or so | ||
2008-12-10 00:13 <X0d_of_N0d> times 50 users looking at their stupid calendars... | ||
2008-12-10 00:13 <X0d_of_N0d> caldav (at least mac's version) pounds the crap out of a server | ||
2008-12-10 00:13 <CIA-54> tryton: C?dric Krier <ced@b2ck.com> default * 1103:7ff1d44306cf tryton/tryton/common/common.py: Fix override of variable args | ||
2008-12-10 00:14 <X0d_of_N0d> it's probably just a poor implimentation | ||
2008-12-10 00:14 <cedk> X0d_of_N0d: is it the server or the client that uses the cpu ? | ||
2008-12-10 00:15 <X0d_of_N0d> but I'm not entirely sure the protocal isn't partially to blame.... | ||
2008-12-10 00:15 <X0d_of_N0d> oh, and there are some problems with clients too | ||
2008-12-10 00:15 <X0d_of_N0d> ligthning and sunbird, for example, don't load the calendar until half an hour after you open it (by default) | ||
2008-12-10 00:15 <X0d_of_N0d> so someone sets up a calendar, then closes the app, then opens the calendar again and complains about how the cal is missing... | ||
2008-12-10 00:15 -!- yangoon1(n=mathiasb@p549F64AA.dip.t-dialin.net) has joined #tryton | ||
2008-12-10 00:15 <X0d_of_N0d> server | ||
2008-12-10 00:16 <X0d_of_N0d> darwin cal server uses a bunch of flat xml files, so there's probably a ton of disk use too | ||
2008-12-10 00:16 <X0d_of_N0d> lemme look at our munin of our cal server really quick | ||
2008-12-10 00:17 <X0d_of_N0d> damn... we don't have munin set up on that system | ||
2008-12-10 00:17 <X0d_of_N0d> well it's a vserver anyway.... | ||
2008-12-10 00:17 <cedk> X0d_of_N0d: what will you suggest as protocol ? | ||
2008-12-10 00:19 <X0d_of_N0d> well.....caldav is the best supported | ||
2008-12-10 00:19 <X0d_of_N0d> lol | ||
2008-12-10 00:19 <X0d_of_N0d> the other ones really just don't work | ||
2008-12-10 00:19 <X0d_of_N0d> we were going to use this implimenatation of caldav instead of apple's http://rscds.sourceforge.net/ | ||
2008-12-10 00:19 <X0d_of_N0d> but... it was broken when we went to look at it | ||
2008-12-10 00:20 <X0d_of_N0d> if you're going to impliment caldav from the ground up in tryton, you might be able to fix a lot of the problems... I really haven't looked deep into the protocal | ||
2008-12-10 00:22 <cedk> X0d_of_N0d: first it will be a big job, and I think if we start it we will look at http://github.com/orbekk/erebus/tree/master | ||
2008-12-10 00:22 <X0d_of_N0d> ACTION shrugs | ||
2008-12-10 00:23 <X0d_of_N0d> yeah... definitely a big job | ||
2008-12-10 00:24 <X0d_of_N0d> but really, as far as I can tell, there isn't anything good in the shared calendar world | ||
2008-12-10 00:25 <X0d_of_N0d> and honestly, the way most people seem to use a calendar here, they SHOULD be using a local one instead of asking us to have set up a server based one | ||
2008-12-10 00:25 <X0d_of_N0d> and it kind of depends on what you want to use it for too.... | ||
2008-12-10 00:25 <X0d_of_N0d> most people just dump in dentist's appointments and crap | ||
2008-12-10 00:26 <X0d_of_N0d> then some people drop in their own stupid stuff on the global calendar... | ||
2008-12-10 00:26 <cedk> X0d_of_N0d: on an ERP view, the main goal for me is to be able to see the calendar of other employee to see if there are available at one date | ||
2008-12-10 00:26 <X0d_of_N0d> but when we were asked to impliment it no one told us what they wanted so.... | ||
2008-12-10 00:27 <cedk> X0d_of_N0d: or create meeting for groups, etc... | ||
2008-12-10 00:27 <X0d_of_N0d> cedk: that seems like a good goal. What users tend to want is to see the work shedule (days off and such). | ||
2008-12-10 00:28 <X0d_of_N0d> cedk: one of our guys wants to be able to organize meetings... | ||
2008-12-10 00:28 <X0d_of_N0d> yeah | ||
2008-12-10 00:28 <cedk> by the way, I just try the webdav with 10000 records and it works well, despite what Fabien says on his post | ||
2008-12-10 00:28 <X0d_of_N0d> the meeting stuff would work really well if it was handled WITHIN the client, instead of the way it's handled here...which is to say it's not | ||
2008-12-10 00:28 -!- rvalyi(i=5645c57e@gateway/web/ajax/mibbit.com/x-bc57b3e4b5d6e3ad) has joined #tryton | ||
2008-12-10 00:28 <bechamel> cedk: maybe it's possible to re-read an ics file edited over webdav and resync it with the db (at least if the webdav client create unique id for new events) | ||
2008-12-10 00:29 <cedk> bechamel: ics globaly is what the caldav protocol send to the client | ||
2008-12-10 00:29 <X0d_of_N0d> I'd almost suggest that the calendar be read-only | ||
2008-12-10 00:29 <rvalyi> mortel, les devs de tryton utilisent irc! Salut les gens, un OpenERPien de passage... | ||
2008-12-10 00:29 <cedk> bechamel: but there is some optimisation to send only what the client needs | ||
2008-12-10 00:30 <bechamel> X0d_of_N0d: read-write in the gtk client + readonly ics is not bad imho | ||
2008-12-10 00:30 <cedk> bechamel: for a first step yes | ||
2008-12-10 00:30 <X0d_of_N0d> all scheduling should be done from inside the server, and all messaging can be done using the internal messaging protocal | ||
2008-12-10 00:30 <bechamel> cedk: maybe caldav is smart enough to send only new events not all | ||
2008-12-10 00:31 <cedk> rvalyi: it is community :-) | ||
2008-12-10 00:31 <cedk> bechamel: I read faster the ISO, and there is many possibility to the client to query the server | ||
2008-12-10 00:31 <cedk> like event from this date to this date | ||
2008-12-10 00:32 <X0d_of_N0d> bechamel: the main confusion that people have is that they'll set up a meeting on their group's calendar, but no one will be informed of the meeting | ||
2008-12-10 00:33 <X0d_of_N0d> bechamel: or they'll set up the meeting, then invite everyone, and then everyone has too meetings because they were invited and added it to their own calendar | ||
2008-12-10 00:33 <cedk> X0d_of_N0d: that is not the protocol fault, but the backend | ||
2008-12-10 00:34 <cedk> X0d_of_N0d: google calendar works well with that I find, but only if everybody has a google calendar | ||
2008-12-10 00:34 <bechamel> X0d_of_N0d: yes calendaring should always come with mail notifications | ||
2008-12-10 00:34 <cedk> bechamel: or request :-) | ||
2008-12-10 00:34 <X0d_of_N0d> cedk: well, or the wetware | ||
2008-12-10 00:34 <X0d_of_N0d> cedk: pbcac | ||
2008-12-10 00:34 <rvalyi> guys: if I can have my words on Tryton; I think most of the reasons why you forked were corrects (like too many features in TinyERP but not enough solid ones), too many regressions... Feedback too low.. But now I think half the th reasons might be gone away. I any case, I think you will make less working on your side than collaborating. my 0.2$ | ||
2008-12-10 00:34 <cedk> X0d_of_N0d: pbcac ??? | ||
2008-12-10 00:35 <X0d_of_N0d> cedk: problem between char and computer | ||
2008-12-10 00:35 <bechamel> rvalyi: i dont undertans, your advice is that we should collaborate ? you know fabien will never accept that | ||
2008-12-10 00:36 <X0d_of_N0d> bechamel: but sometimes a boss will just want to nudge users and point out there is a meeting, but not say "you need to attend" like sending an invite does | ||
2008-12-10 00:36 <rvalyi> bechame: well for sure now it looks like your relations passed the return point... Why not take an other pseudo then ;-) | ||
2008-12-10 00:36 <X0d_of_N0d> ACTION looks at rvalyi | ||
2008-12-10 00:36 <X0d_of_N0d> 20 cents... damn inflation | ||
2008-12-10 00:37 <rvalyi> ;-) | ||
2008-12-10 00:37 <X0d_of_N0d> rvalyi: there are several basic design choices that differ between openerp and tryton | ||
2008-12-10 00:37 <bechamel> rvalyi: the return point was anterior to the fork and was one of the reason behind it. | ||
2008-12-10 00:38 <X0d_of_N0d> rvalyi: these two projects are going in different directions... this may not be obvious from the outside, but it's very obvious by looking at the code | ||
2008-12-10 00:39 <bechamel> rvalyi: and to be honest i don't think that half the reasons are aways | ||
2008-12-10 00:39 <rvalyi> Well in any case that's sad, because for sure you got a few points with Tryton. That's said between the two options, it's unlikely I bet on you. Of course not everybody can compare communites, but it will be very hard. | ||
2008-12-10 00:40 <cedk> rvalyi: I think there is enough places for many OpenSource ERP | ||
2008-12-10 00:40 <rvalyi> Technology for sure matters, but if Compiere, ERP5, Ofbiz and now Openbravo are tow fingers to collapse, that's because oss ERP's is a damn hard topic. Being on you own sounds pretty rsiky to me | ||
2008-12-10 00:41 <bechamel> rvalyi: we don't want to fight, we don't want to win anything, we just want to do something else. my best wishes is that openerp and tryton are gonna be/stay two successfull soft. Choice matters | ||
2008-12-10 00:41 <rvalyi> cedk: yeah, you are probably right. I'm pretty confident you'll end up with something; I never said you woul'd reach something. Still what I believe is that you might miss the major success while not consolidating OpenERP that would for sure benefit from your improvements | ||
2008-12-10 00:43 <rvalyi> how, and if you gus have one minute: what's the point about the webservice change: login once and then going on with a token, why does it matter? (sorry if that's a dumb question). The openERP choice sounds unusual but seems to work, what is the flaw? | ||
2008-12-10 00:44 <X0d_of_N0d> the things that make Compiere, ERP5, and Openbravo seem to be the very same things that restrict openerp... that being primarily their dificulty to code | ||
2008-12-10 00:44 <cedk> rvalyi: the password is sended in clear on every request | ||
2008-12-10 00:44 <cedk> rvalyi: and what Fabien says about the client that crashs when the server is restarted is False | ||
2008-12-10 00:44 <bechamel> rvalyi: the first reason was to speed the request, the second is to send the password less often | ||
2008-12-10 00:44 <cedk> rvalyi: the client just reconnect | ||
2008-12-10 00:45 <rvalyi> cedk: sure, but as you would end up on HTTPS, why bother? | ||
2008-12-10 00:45 <X0d_of_N0d> openerp is head and shoulders above the others, but tryton is equally more clean and clear | ||
2008-12-10 00:45 <cedk> rvalyi: second point about security is that it is not good to keep password in memory | ||
2008-12-10 00:46 <rvalyi> speed the request: because you don't have to crawl the DB to see if the user is authenticated at each request? But how much does it matter compared to network latency, I don't get it.... | ||
2008-12-10 00:46 <cedk> rvalyi: no the speed doesn't change here | ||
2008-12-10 00:46 <bechamel> rvalyi: you win also a dns request | ||
2008-12-10 00:46 <rvalyi> ? | ||
2008-12-10 00:47 <rvalyi> why? a dns request? | ||
2008-12-10 00:47 <cedk> rvalyi: next point is that with a session, we have timeout so if you let you computer idle the Tryton client will be disconnected after some time | ||
2008-12-10 00:47 <cedk> bechamel: no | ||
2008-12-10 00:47 <bechamel> cedk: ok sorry, i learn something else today | ||
2008-12-10 00:47 <cedk> bechamel: I set up a DNS cache on TinyERP netrpc | ||
2008-12-10 00:48 <cedk> rvalyi: if you look arround, session model is pretty common | ||
2008-12-10 00:48 <bechamel> rvalyi: but after all you see a problem to keep the socket ? | ||
2008-12-10 00:48 <rvalyi> cedk: "next point is that with a session, we have timeout so if you let you computer idle the Tryton client will be disconnected after some time" But an OpenERP doens't leak any server side load between two requests as it's considered disconnected, so again what's the bennefit here? | ||
2008-12-10 00:48 <cedk> rvalyi: clear text authentification is not the best | ||
2008-12-10 00:49 <cedk> rvalyi: security | ||
2008-12-10 00:49 <rvalyi> cedk: OK, I know that the session model is the way. But if it's on HTTPS, what's the difference? | ||
2008-12-10 00:49 <cedk> rvalyi: any software with sensitive data has session timeout when idle | ||
2008-12-10 00:49 <rvalyi> OK, I got it | ||
2008-12-10 00:50 <cedk> rvalyi: by the way Tryton is not on HTTPS but SSL | ||
2008-12-10 00:50 <rvalyi> it's because the client keep the password in memory, ok | ||
2008-12-10 00:50 <cedk> rvalyi: yes | ||
2008-12-10 00:50 <rvalyi> wel, whatever SSL, doens't matter to me | ||
2008-12-10 00:50 <rvalyi> dumb question then | ||
2008-12-10 00:50 <X0d_of_N0d> rvalyi: do you trust everyone on your network? more over, if you use this outside your company do you trust everyone on the internet? | ||
2008-12-10 00:50 <cedk> rvalyi: with OpenERP, if you leave the client and come one day later, you can still use it without entering the password | ||
2008-12-10 00:51 <rvalyi> X0d_of_N0d: If I'm on HTTPS / SSL, why would I need to trust everyone? It's tough to hack an HTTPS... | ||
2008-12-10 00:52 <cedk> rvalyi: I know a little about security as I work in a company that makes security network appliance | ||
2008-12-10 00:53 <cedk> s/work/worked/ | ||
2008-12-10 00:53 <X0d_of_N0d> rvalyi: you don't have to as long as the cert is valid and not self-signed | ||
2008-12-10 00:53 <rvalyi> But still , an OpenERP client could logg off after some inactivity delay. To me only the password in the GTK client memory is a real issue | ||
2008-12-10 00:53 <yangoon1> rvalyi: if someone hacks your workstation and gets the credentials, SSL doesn't matter | ||
2008-12-10 00:53 <X0d_of_N0d> rvalyi: or you, or your people, actually pay any attention to certs | ||
2008-12-10 00:54 <cedk> rvalyi: it can but it is strange to put this on the client side | ||
2008-12-10 00:54 <rvalyi> sure, it's not the common way of doing, I was just wondering how much it was a flaw | ||
2008-12-10 00:55 <cedk> X0d_of_N0d: in Tryto, you have SSL info when you put the mouse over the lock | ||
2008-12-10 00:55 <cedk> rvalyi: the main issue is really the password keept in memory | ||
2008-12-10 00:56 <rvalyi> guys, actually to me, the largest problem of OpenERP is not the technology. There are flaws for sure, but OpenERP is already hundred times better than Compiere, SAP, Openbravo, Navision and co | ||
2008-12-10 00:56 <bechamel> rvalyi: so waht is the problem ? | ||
2008-12-10 00:57 <X0d_of_N0d> rvalyi: what, to you, makes OpenERP better? | ||
2008-12-10 00:57 <cedk> and in Tryton, I think we can improve it because we let to the garbage collector the cleansing of the memory | ||
2008-12-10 00:57 <rvalyi> the biggest problem is actually the lack of polish in the feature that make you, as a consultant work a lot to fulfill a customer requirement or sell it with a determinist price. I'm cursious, do you share that analysis? If yes in which ways do you address it? | ||
2008-12-10 00:57 <X0d_of_N0d> I agree, but I'm curious what you see | ||
2008-12-10 00:59 <cedk> rvalyi: there is some guidelines that we try to follow like: | ||
2008-12-10 00:59 <rvalyi> cedk: Anyway, if Jython on day get something as good as JRuby, Then I would like to see OpenERP or Tryton run on Jython, then you'll have a TRUE garabage collector, a TRUE multithreading, a TRUE SGBD connector... | ||
2008-12-10 00:59 <cedk> - allowing as much as possible, the override of functions | ||
2008-12-10 00:59 <bechamel> rvalyi: yes of course we share it, and that why we rework all the modules, and that why fp can tell happily to everobody that tryton is miles away in terms of business coverage | ||
2008-12-10 00:59 <rvalyi> and BTW by migrating to Jython, you good benefit the MOP and code your modules in JRuby, wouldn't that be a must? | ||
2008-12-10 01:00 <mmarshall> rvalyi, what's lacking with python's garbage collector? | ||
2008-12-10 01:00 <cedk> - put docstring on function | ||
2008-12-10 01:00 <cedk> - make things as simple as possible | ||
2008-12-10 01:00 <rvalyi> Is that some generationnal GC like the JVM has? Not too sure. About the threading, I'm quite sure... | ||
2008-12-10 01:00 <cedk> - put only necessary information on core module and extend it with other modules | ||
2008-12-10 01:01 <rvalyi> "put docstring on function" -> for sure that's one thing I can't understand in OpenERP | ||
2008-12-10 01:01 <bechamel> rvalyi: i saw your idea on the forum, for me it's a nice technical challenge but from a pratical point of view, i thing cpython is good enough (even if i agree with you threads are better in java, i don(t know for the gc) | ||
2008-12-10 01:01 <X0d_of_N0d> rvalyi: I'm not a consultant, so the benefit I see is that it's easier for me to import data into openerp than any of the others. also it's easier to write modules, which is similar to what you said. | ||
2008-12-10 01:01 <rvalyi> that's the same as fp commits labeled 'improvement' and 'bugfix' without any hell crossref | ||
2008-12-10 01:02 <cedk> bechamel: not cpython but psyco and Tryton support it :-) | ||
2008-12-10 01:02 <rvalyi> yeah cpython is good enough, sure | ||
2008-12-10 01:02 <bechamel> cedk: python threads are not gonna be better with psyco | ||
2008-12-10 01:03 <rvalyi> an other point is that Openbravo just raised 12M$ by making investors dream about SUN buying them cause they run on Java | ||
2008-12-10 01:03 <cedk> bechamel: not for thread but for code execution | ||
2008-12-10 01:03 <cedk> the main issue with python is that it is mono-core I think | ||
2008-12-10 01:03 <mmarshall> rvalyi, There's information on cpython's gc here: http://arctrix.com/nas/python/gc/ | ||
2008-12-10 01:03 <rvalyi> Sun will never buy them cause it's only Oracle pl/SQL, but if you want to pay for a java based ERP, OpenERP/Tryton on Jtython is by far the easiest way | ||
2008-12-10 01:03 <bechamel> rvalyi: i was told that most of code in openbravo was plsql not java, can you confirm ? | ||
2008-12-10 01:04 <cedk> but with handle the issue by allowing to setup many instance of the server on the same host | ||
2008-12-10 01:04 <X0d_of_N0d> cedk: it's single threaded unless you fork it, yeah | ||
2008-12-10 01:04 <X0d_of_N0d> it would be possible to fork on connection though | ||
2008-12-10 01:04 <cedk> X0d_of_N0d: but we use thread and not fork | ||
2008-12-10 01:04 <X0d_of_N0d> ACTION thinks.... | ||
2008-12-10 01:04 <cedk> X0d_of_N0d: fork cost too much | ||
2008-12-10 01:05 <rvalyi> bechamel: sure, just look at the CDATA inside the XML files. 6 months ago that PL had a Compiere license header with old dead code from Compiere, back to 2001 when it was for Goodyear Germany. And look at their wiki howto, ther aren't in Java... | ||
2008-12-10 01:05 <X0d_of_N0d> cedk: ok | ||
2008-12-10 01:05 <X0d_of_N0d> cedk: ahh, ok | ||
2008-12-10 01:05 <X0d_of_N0d> ACTION hasn't really played around with threading in python | ||
2008-12-10 01:05 <cedk> X0d_of_N0d: i think the best is if you have 8 cores, you run 8 instance of Tryton | ||
2008-12-10 01:05 <mmarshall> Python has logical threads, but each logical thread shares a single 'real' thread. | ||
2008-12-10 01:06 <mmarshall> So you can only do multi-core threading in c extensions. | ||
2008-12-10 01:06 <mmarshall> (Or use multiple processes.) | ||
2008-12-10 01:06 <X0d_of_N0d> cedk: because of the way it works there's no reason you couldn't do loadbalancing with tryton on multiple systems, right? | ||
2008-12-10 01:07 <rvalyi> confirm, that's a Giant Lock, like MRI Ruby and unlike JRuby | ||
2008-12-10 01:07 <bechamel> mmarshall: do you know if tasklets or processing are a good solution ? | ||
2008-12-10 01:07 <rvalyi> well, doesn't matter too much. Having JRuby modules, a Java server, statically typed low level (not business) would just be too great... | ||
2008-12-10 01:07 <cedk> mmarshall: are you sure ? I try it and I see 2 thread with ps for the Tryton server | ||
2008-12-10 01:08 <mmarshall> bechamel, I haven't needed to use either in production yet, but they look good for some uses. | ||
2008-12-10 01:08 <cedk> rvalyi: by the way, one more things about session, is that we keep a thread running for each client so on each request the server doesn't need to create a new socket and a new thread | ||
2008-12-10 01:08 <mmarshall> cedk, It might actually be separate threads, but only one thread can run python code at a time. | ||
2008-12-10 01:09 <mmarshall> So it might as well be a single thread. | ||
2008-12-10 01:09 <rvalyi> cedk: I think the point is that Python has parralelism, not concurrency | ||
2008-12-10 01:09 <rvalyi> thats partially were the JVM and thing like JRuby and Jython later on will enter into play | ||
2008-12-10 01:09 <cedk> so ok, any way, you can run many instance as we keep cache syncronized | ||
2008-12-10 01:10 <mmarshall> So unless if you have some processing in a c extension that releases the Global Interpreter Lock (GIL) you don't get much from threading. | ||
2008-12-10 01:10 <cedk> mmarshall: except parallelism on IO access | ||
2008-12-10 01:11 <rvalyi> of course, this doens't matter for the average small company needing an ERP. But for large hosting like Google, this might be a differentiator | ||
2008-12-10 01:11 <cedk> mmarshall: like connection to the DB which is still the bottleneck of the server | ||
2008-12-10 01:11 <bechamel> rvalyi: i think that if the kernel is statically typed it will impact modules a lot (maybe completely) | ||
2008-12-10 01:11 <rvalyi> I mean for an efficient Saas | ||
2008-12-10 01:12 <bechamel> i was wondering today if its possible to use GAE with postregsql ? | ||
2008-12-10 01:12 <rvalyi> bechamel: you certainly could have the ORM, OSV and XML stuff statically typed like a Java+JRuby combo, I doubt that would hurt | ||
2008-12-10 01:12 <cedk> rvalyi: I don't think that if you want to run a SAAS, you will use one server for all your clients | ||
2008-12-10 01:12 <cedk> s/clients/customers/ | ||
2008-12-10 01:12 <mmarshall> cedk, Ok, in a case where it's waiting for a response from the db, yes, another thread would take over. (So long as the db connector is releasing the GIL like it should.) | ||
2008-12-10 01:12 <rvalyi> cedk: sure, but then computing powers starts to matters anyway | ||
2008-12-10 01:13 <bechamel> rvalyi: all business classes inherit from orm | ||
2008-12-10 01:13 <X0d_of_N0d> cedk: you can run multiple instances of tryton across multiple server though, right? | ||
2008-12-10 01:13 <cedk> mmarshall: how can I check this ? | ||
2008-12-10 01:13 <rvalyi> bechamel: what's the trouble JRuby can inherit extends Java base classes at will, just as Jython can I think | ||
2008-12-10 01:14 <cedk> X0d_of_N0d: yes, on the same host or on different hosts | ||
2008-12-10 01:14 <X0d_of_N0d> cedk: and it's possible to cluster postgres, so..... | ||
2008-12-10 01:14 <cedk> rvalyi: but what has Jython that Python doesn't ? | ||
2008-12-10 01:14 <rvalyi> I think that's the future. But in any case I think OpenERP / Tryton architecture are pretty OK and way better than competitors, so as I said, the main problem for broader adoptions aren't here | ||
2008-12-10 01:14 <cedk> X0d_of_N0d: I'm not sure for writing | ||
2008-12-10 01:14 <bechamel> rvalyi: maybe you are right i never used jython | ||
2008-12-10 01:15 <mmarshall> cedk, searching for "psycopg2 GIL" seems to imply that psycopg2 is GIL aware, so I guess you're good. | ||
2008-12-10 01:15 <cedk> mmarshall: great and psycopg1 ? | ||
2008-12-10 01:15 <mmarshall> (So long as you're using psycopg2 version 2.0.7 or later.) | ||
2008-12-10 01:15 <X0d_of_N0d> ACTION goes to look into it | ||
2008-12-10 01:15 <mmarshall> See here: http://blogs.gnome.org/jamesh/2008/04/15/psycopg2/ | ||
2008-12-10 01:16 <mmarshall> "Fixes for a few memory leaks and GIL misuses." | ||
2008-12-10 01:16 <rvalyi> cedk: today nothing. In one year it could be faster, have tru multithreading, offer Java strong libs for XML,SOAP,JMS have an easier way to optimize using Java then C, better remote monitoring, fine grained choice between static Java and Dynamic Python... | ||
2008-12-10 01:17 <cedk> rvalyi: for threading, I read that Python3 has a some good stuff for that | ||
2008-12-10 01:17 <rvalyi> cedk: so it has for optionnal static typing I think | ||
2008-12-10 01:18 <cedk> mmarshall: it seems that psycopg < 2 unlock the GIL | ||
2008-12-10 01:18 <rvalyi> but just like MRI 1.87 Ruby and 1.9 Ruby are loosing against JRuby (even with way more resources), I think an amateur interpreter has no chance against the JVM if Sun wants to push Jython as far as JRuby which looks what they are now trying | ||
2008-12-10 01:19 <rvalyi> guys, I'm curious about an other thing: what is your business model? On what money are you running for now? | ||
2008-12-10 01:20 <cedk> rvalyi: don't forget Parrot :-) | ||
2008-12-10 01:20 <mmarshall> rvalyi, Static typing was something that was talked about several years ago, but didn't make the cut. | ||
2008-12-10 01:20 <bechamel> rvalyi: "amateur interpreter", hum | ||
2008-12-10 01:20 <mmarshall> rvalyi, although you could use the new function annotations for specifying types. | ||
2008-12-10 01:20 <cedk> Python is used at Google :-) | ||
2008-12-10 01:21 <cedk> and NASA use Python | ||
2008-12-10 01:21 <mmarshall> And *I* use python! Don't forget that! ;) | ||
2008-12-10 01:21 <rvalyi> (oh one last thing about Jython: JRuby got there because some smary guys like Tim Bray at Sun decided that Rails would be the killer app the JVM would need. If those guys see OpenERP / Tryton IS the killer Jython app they need, that could push Jython ahdead. For now they are slowly porting Django. Of course they don't have Nutter in the team, but they aren't to slow either) | ||
2008-12-10 01:22 <cedk> http://www.python.org/about/success/ | ||
2008-12-10 01:22 <bechamel> rvalyi: for the business model, it's roughly the same than tiny, execpt then we don't want to charge partners if they want to remove the logo :), | ||
2008-12-10 01:23 <rvalyi> bechamel: but Tiny is making a some money out of their partners, so how could you afford not doing so? | ||
2008-12-10 01:23 <cedk> rvalyi: not completly the same, we provide devs and support | ||
2008-12-10 01:23 <rvalyi> ok, how many customers for now? | ||
2008-12-10 01:23 <rvalyi> how many employees? Is taht a company? | ||
2008-12-10 01:23 <bechamel> rvalyi: you need support ? | ||
2008-12-10 01:23 <rvalyi> I'm selling support ;-) | ||
2008-12-10 01:24 <rvalyi> at wwww.Smile.fr | ||
2008-12-10 01:24 <cedk> rvalyi: Tryton is not only B2CK, there is other people and companies | ||
2008-12-10 01:24 <rvalyi> ok | ||
2008-12-10 01:25 <cedk> http://www.tryton.org/services.html | ||
2008-12-10 01:25 <cedk> for companies | ||
2008-12-10 01:25 <cedk> people can be found in the contributor list or on the chan :-) | ||
2008-12-10 01:25 <rvalyi> I find it Ok OpenERP charging partners, except that it's too expensive for developping countries, they should have a way to address this | ||
2008-12-10 01:26 <mmarshall> Is there anyone using Tryton in the US? | ||
2008-12-10 01:26 <cedk> mmarshall: X0d_of_N0d is from the US | ||
2008-12-10 01:26 <rvalyi> shame on you | ||
2008-12-10 01:26 <rvalyi> sorry, was to easy ;-) | ||
2008-12-10 01:26 <mmarshall> lol | ||
2008-12-10 01:27 <rvalyi> and now that US voted right, there is no more reasons... | ||
2008-12-10 01:28 <cedk> rvalyi: Please don't speak politics on this chan | ||
2008-12-10 01:30 <rvalyi> yeah linuxfr is meant for it ;-) ( to deal with oss ERP's polictics) | ||
2008-12-10 01:30 <mmarshall> So, is there a module available to rearrange the address forms? | ||
2008-12-10 01:31 <cedk> mmarshall: not yet, but the address form can be customized easily | ||
2008-12-10 01:31 <cedk> mmarshall: what do you need? | ||
2008-12-10 01:31 <mmarshall> ZIP after city, mostly. | ||
2008-12-10 01:32 <mmarshall> And use TX instead of Texas. | ||
2008-12-10 01:32 <cedk> mmarshall: look at modules/party/address.py line 42 | ||
2008-12-10 01:32 <mmarshall> ok | ||
2008-12-10 01:32 <cedk> mmarshall: there is a function that is used in each report to display the address | ||
2008-12-10 01:33 <cedk> mmarshall: so you can create a module that override this function and change the order | ||
2008-12-10 01:33 <rvalyi> an issue we ran into today with internationalization is that OpenERP doesn't seem to translate report headers (probably because the setlang is not reached yet), how stand Tryton here? | ||
2008-12-10 01:34 <cedk> rvalyi: the header is translated | ||
2008-12-10 01:34 <cedk> rvalyi: but you have it on the report in one language | ||
2008-12-10 01:34 <rvalyi> on other issue is that for a country selling to France and US for instance, they can't afford just translating dates because thir salesman will make error when reading 09/12/2008 vs 12-09-2008, so we should find a way to print the month explicitely | ||
2008-12-10 01:35 <rvalyi> ok | ||
2008-12-10 01:36 <cedk> rvalyi: in Tryton, you can override the format_lang function and display the date with the format that you want | ||
2008-12-10 01:36 <rvalyi> cedk: I'm curious: what abou the rest of the OpenERP belgium team (not fp), did they come after or before you, how good are they (honnestly?) | ||
2008-12-10 01:37 <cedk> rvalyi: don't think that I will answer this question :-) | ||
2008-12-10 01:37 <rvalyi> cedk: Python has no open class like Ruby, right? So I can't subclass the Hash base class to send a nice message for instance is a key isn't found, correct? | ||
2008-12-10 01:37 <bechamel> rvalyi: ask fp :) | ||
2008-12-10 01:37 <rvalyi> wow... | ||
2008-12-10 01:38 <cedk> rvalyi: but somes was there when we leave and others not | ||
2008-12-10 01:39 <cedk> rvalyi: don't know about "open class" | ||
2008-12-10 01:39 <cedk> rvalyi: but you can replace any class of any module | ||
2008-12-10 01:39 <cedk> rvalyi: if it is in python | ||
2008-12-10 01:39 <cedk> I think | ||
2008-12-10 01:39 <rvalyi> I think it's hard to find people able to developp on OpenERP / Tryton, I mean to be autonomous. We certainly have difficulties around this at Smile to reach our full spedd | ||
2008-12-10 01:40 <bechamel> rvalyi: in python you can do "instance.old_method = lambda *a: return None", but i don't know if it answer your question | ||
2008-12-10 01:40 <rvalyi> cedk: even base classes backed by C? | ||
2008-12-10 01:40 <cedk> rvalyi: I don't know | ||
2008-12-10 01:40 <cedk> never try | ||
2008-12-10 01:40 <rvalyi> wil try | ||
2008-12-10 01:41 <mmarshall> rvalyi, this is called "monkey patching" and is generally frowned upon in python. | ||
2008-12-10 01:41 <rvalyi> rails does that a lot and smartly I think | ||
2008-12-10 01:41 <rvalyi> (sometimes a bit too much of course) | ||
2008-12-10 01:42 <mmarshall> It's mostly a matter of taste, but all the monkey patching is something I really don't like about rails. | ||
2008-12-10 01:42 <rvalyi> I python had no len(my_fucking_table_wich_is_an_object) I would almost love it | ||
2008-12-10 01:42 <bechamel> rvalyi: there are a lot of openerp dev at Smile ? | ||
2008-12-10 01:43 <rvalyi> bechamel: do you mean enabled devs? | ||
2008-12-10 01:43 <cedk> so it is late here, good night! | ||
2008-12-10 01:43 <rvalyi> joker, just like for you about Tiny | ||
2008-12-10 01:43 <bechamel> rvalyi: both | ||
2008-12-10 01:43 <mmarshall> cedk, thanks for pointing me to the address stuff. | ||
2008-12-10 01:43 <bechamel> rvalyi: ;) | ||
2008-12-10 01:43 <rvalyi> buy, cice talk | ||
2008-12-10 01:44 <rvalyi> err, bye | ||
2008-12-10 01:45 <rvalyi> OK, then Tryton is not don by true hardcore devs coding overnight, that's known already | ||
2008-12-10 01:45 <bechamel> rvalyi: otoh it's 2 in the morning .. | ||
2008-12-10 01:46 <rvalyi> and I've that fucking sale direct delivery module to developp... | ||
2008-12-10 01:46 <bechamel> rvalyi: validate packing when the sale is done ? | ||
2008-12-10 01:47 <rvalyi> a bit more: looking at virtual stocks and product supplier to choose a direct delivery, | ||
2008-12-10 01:48 <bechamel> rvalyi: direct means: from the supplier direct to the customer ? | ||
2008-12-10 01:48 <rvalyi> then changing the sale into make to order, cross linking the so_line and the po_line over the MRP, pasting the shipping address into the po, changing the destination of the picking... | ||
2008-12-10 01:48 <rvalyi> yes it means | ||
2008-12-10 01:49 <rvalyi> sale_supplier_direct_delivery in trunk extra_addons | ||
2008-12-10 01:49 <rvalyi> more soon | ||
2008-12-10 01:49 <bechamel> rvalyi: not an easy dev i agree | ||
2008-12-10 01:50 <rvalyi> well, I was still learing the framework, now I think I master it at some point | ||
2008-12-10 01:51 -!- vengfulsquirrel(n=ian@c-67-170-212-242.hsd1.ca.comcast.net) has joined #tryton | ||
2008-12-10 01:51 <CIA-54> tryton: Bertrand Chenal <bch@b2ck.com> default * 1341:fd9a4bf8bab4 trytond/trytond/ir/cron.py: Fix cron messages: fallback to default langage if missing on the user | ||
2008-12-10 01:52 <rvalyi> an other thing: you server side ordering + boolean conditions between srearch fields is definitely a good thing too I would like to see in OpenERP | ||
2008-12-10 01:54 <bechamel> rvalyi: doesn't fp tell in the forum that it will be added for the 5.0 ? | ||
2008-12-10 01:55 <vengfulsquirrel> What is tryton's relationship with openerp ? | ||
2008-12-10 01:56 <bechamel> vengfulsquirrel: http://www.tryton.org/documentation/faq.html | ||
2008-12-10 01:57 <vengfulsquirrel> Yeah I read that already. | ||
2008-12-10 01:57 <yangoon1> bye all | ||
2008-12-10 01:58 <vengfulsquirrel> I guess I haven't gotten openerp to really install correctly yet, so I don't understand the "less cutting edge features" part. | ||
2008-12-10 01:58 <vengfulsquirrel> I'm pretty amazing people exist here though. | ||
2008-12-10 01:58 <vengfulsquirrel> *amazed | ||
2008-12-10 01:58 <vengfulsquirrel> Will improvement be shared between the two code bases? I guess that was my original question's intent. | ||
2008-12-10 01:59 <vengfulsquirrel> *improvements | ||
2008-12-10 01:59 <bechamel> vengfulsquirrel: maybe if you manage to install tryton it will means that "solid solution" is better on this point | ||
2008-12-10 02:00 <bechamel> vengfulsquirrel: there is already code moving in both directions | ||
2008-12-10 02:00 <mmarshall> In the screenshot of the menu there is a "Timesheet Management" item. Which module includes that? | ||
2008-12-10 02:01 <vengfulsquirrel> the setup.py is cleaned up that is _NICE_ | ||
2008-12-10 02:02 <rvalyi> ok, buy guys, late here | ||
2008-12-10 02:02 <bechamel> mmarshall: the screenshot comes from the timesheet module, it's availble with hg. the module is not complete so we choose to not package it with the 1.O | ||
2008-12-10 02:02 <rvalyi> err, bye | ||
2008-12-10 02:02 <bechamel> rvalyi: yes, very late, bye | ||
2008-12-10 02:02 <mmarshall> bechamel, ok, thanks. | ||
2008-12-10 02:02 <bechamel> rvalyi: btw interesting conversation | ||
2008-12-10 02:02 <rvalyi> I think I heard that one about not completely complete modules... | ||
2008-12-10 02:03 -!- rvalyi(i=5645c57e@gateway/web/ajax/mibbit.com/x-bc57b3e4b5d6e3ad) has left #tryton | ||
2008-12-10 02:04 <bechamel> vengfulsquirrel: I'm gonna sleep too, if you have questions there is also the mailing list | ||
2008-12-10 02:04 <vengfulsquirrel> Thanks, you guys are mostly in Europe? | ||
2008-12-10 02:04 <vengfulsquirrel> ha that's my last question | ||
2008-12-10 02:04 <bechamel> vengfulsquirrel: yes i think so | ||
2008-12-10 02:05 <bechamel> maybe mmarshall ? | ||
2008-12-10 02:05 <vengfulsquirrel> and then I'm going to rebuild my slice on slicehost for the 10th time, but try trytond | ||
2008-12-10 02:07 <CIA-54> tryton: Bertrand Chenal <bch@b2ck.com> default * 145:674901b91943 stock_supply/purchase_request.py: Isolate request creation to ease inheritance | ||
2008-12-10 03:41 -!- ikks(n=igor@190.12.153.202) has joined #tryton | ||
2008-12-10 04:08 -!- mmarshall_(n=mmarshal@adsl-99-186-213-109.dsl.rcsntx.sbcglobal.net) has joined #tryton | ||
2008-12-10 04:36 -!- vengfulsquirrel(n=ian@c-67-170-212-242.hsd1.ca.comcast.net) has joined #tryton | ||
2008-12-10 05:18 -!- yangoon(n=mathiasb@p549F5597.dip.t-dialin.net) has joined #tryton | ||
2008-12-10 07:15 -!- nicoe(n=nicoe@ip-80-236-193-104.dsl.scarlet.be) has joined #tryton | ||
2008-12-10 07:45 -!- sharkcz(n=dan@plz1-v-4-17.static.adsl.vol.cz) has joined #tryton | ||
2008-12-10 08:08 -!- LordVan(n=lordvan@gentoo/developer/LordVan) has joined #tryton | ||
2008-12-10 08:19 -!- Guest62659(n=ged@77.109.114.68.adsl.dyn.edpnet.net) has joined #tryton | ||
2008-12-10 08:28 -!- Timitos(n=Timitos@88.217.184.172) has joined #tryton | ||
2008-12-10 09:38 -!- vengfulsquirrel(n=ian@c-67-170-212-242.hsd1.ca.comcast.net) has joined #tryton | ||
2008-12-10 09:41 -!- Gedd(n=ged@ip-80-236-193-104.dsl.scarlet.be) has joined #tryton | ||
2008-12-10 09:54 -!- marmu(n=marmu@p54AD4878.dip.t-dialin.net) has joined #tryton | ||
2008-12-10 10:07 <CIA-54> tryton: leyoda roundup * #655/Tryton Client Installation under OSX 10.5.X: [new] After the setup.py installation, the Tryton Client can't start with this error : File "/Library/Python/2.5/site-packages/tryton-1.1.0-py2. ... | ||
2008-12-10 10:09 -!- cedk(n=ced@gentoo/developer/cedk) has joined #tryton | ||
2008-12-10 10:28 <CIA-54> tryton: C?dric Krier <ced@b2ck.com> default * 1104:a0e0a8c03e1d tryton/setup.py: Fix path in setup.py for tryton-icon.png for issue655 | ||
2008-12-10 10:28 <CIA-54> tryton: ced roundup * #655/Tryton Client Installation under OSX 10.5.X: [resolved] Fix with changeset a0e0a8c03e1d | ||
2008-12-10 11:17 -!- LordVan(n=lordvan@gentoo/developer/LordVan) has joined #tryton | ||
2008-12-10 12:11 -!- ikks(n=igor@190.12.153.202) has joined #tryton | ||
2008-12-10 12:21 -!- bechamel(n=user@host-213-213-252-112.brutele.be) has joined #tryton | ||
2008-12-10 12:25 -!- ecir3(n=brice@sednaco19320-gw.clients.easynet.fr) has joined #tryton | ||
2008-12-10 12:40 <CIA-54> tryton: C?dric Krier <ced@b2ck.com> default * 1342:740c6541c943 trytond/TODO: Add todo for pre-fetch in webdav | ||
2008-12-10 13:04 <vengfulsquirrel> I tried adding the gentoo overlay but when I try to emerge it says "masked by: missing keyword". | ||
2008-12-10 13:05 <cedk> vengfulsquirrel: every ebuild in overlay are ~arch | ||
2008-12-10 13:06 <cedk> vengfulsquirrel: so you must enter a keyword for the packages you want to install in /etc/portage/package.keywords | ||
2008-12-10 13:07 <vengfulsquirrel> hmm | ||
2008-12-10 13:07 <vengfulsquirrel> As in ~amd64 correct? | ||
2008-12-10 13:09 <vengfulsquirrel> Sorry if I sound stupid but I've never used overlays I just wanted to make sure I wasn't missing something obvious. | ||
2008-12-10 13:10 <cedk> vengfulsquirrel: I just put ~x86 in the ebuild but if you can confirm me that it works on amd64, I will put it | ||
2008-12-10 13:11 <vengfulsquirrel> Yeah putting ~amd64 does not work in the package.keywords but you mean more like does "trytond" work in amd64 right ? | ||
2008-12-10 13:12 <cedk> vengfulsquirrel: I think you must also add it in package.unmask | ||
2008-12-10 13:13 <vengfulsquirrel> Putting x86 works. | ||
2008-12-10 13:14 <cedk> vengfulsquirrel: tell me if all works wel on amd64 | ||
2008-12-10 13:15 <vengfulsquirrel> If what works? | ||
2008-12-10 13:16 <vengfulsquirrel> I've installed the server via mercurial and http://hg.tryton.org/trytond/. | ||
2008-12-10 13:16 <cedk> vengfulsquirrel: all the Tryton package | ||
2008-12-10 13:16 <vengfulsquirrel> I have not installed the client. | ||
2008-12-10 13:16 <vengfulsquirrel> I also don't think the modules are installed. | ||
2008-12-10 13:16 <vengfulsquirrel> I haven't really figured out how to get them to install. | ||
2008-12-10 13:16 <cedk> vengfulsquirrel: if you succed to run Tryton on amd64, I will add amd64 to the arch of ebuilds in the overlay | ||
2008-12-10 13:17 <vengfulsquirrel> *Installed the server but connected to it remotely from an x86 client machine. | ||
2008-12-10 13:17 <cedk> vengfulsquirrel: so for now, you can confirm that the server run on amd64 | ||
2008-12-10 13:17 <vengfulsquirrel> So the server(trytond) runs on amd64 without the modules installed. | ||
2008-12-10 13:17 <vengfulsquirrel> How do I install the modules? | ||
2008-12-10 13:18 <vengfulsquirrel> These modules trytond/tryond/modules. | ||
2008-12-10 13:19 <cedk> vengfulsquirrel: from mecurial? | ||
2008-12-10 13:20 <bechamel> cedk: i think that it would be usefull to talk about the script tryton-dev.sh on the dowload page, because its a problem for a lot of people | ||
2008-12-10 13:21 <bechamel> vengfulsquirrel: this script allow to do fetch client server and modules at once | ||
2008-12-10 13:22 <vengfulsquirrel> Well I was going from this: http://code.google.com/p/tryton/wiki/InstallationMercurial | ||
2008-12-10 13:22 <bechamel> vengfulsquirrel: http://www.b2ck.com/~ced/tryton-dev.sh | ||
2008-12-10 13:22 <vengfulsquirrel> Oh yeah so I've gotten the modules already into their directory. | ||
2008-12-10 13:23 <cedk> bechamel: can you add it to the wiki page | ||
2008-12-10 13:23 <bechamel> vengfulsquirrel: this should also work, for modules you can copy-paste the list of "hg clone http://..." on your terminal and will fetch all of thiese | ||
2008-12-10 13:23 <vengfulsquirrel> But only AFTER I ran python setup.py install | ||
2008-12-10 13:23 <vengfulsquirrel> How can I tell they are installing when I run python setup.py install ? | ||
2008-12-10 13:24 <bechamel> cedk: where is the "edit" link on the wiki ? | ||
2008-12-10 13:25 <yangoon> bechamel are you logged in? | ||
2008-12-10 13:25 <bechamel> yangoon: yes but maybe with the wrong user | ||
2008-12-10 13:26 -!- LordVan(n=lordvan@gentoo/developer/LordVan) has joined #tryton | ||
2008-12-10 13:26 <cedk> vengfulsquirrel: you don't need to install package | ||
2008-12-10 13:26 <cedk> vengfulsquirrel: if you want just test it, Tryton works with the sources without install | ||
2008-12-10 13:27 <cedk> vengfulsquirrel: but you can also install it if you want, but you need to do it with root | ||
2008-12-10 13:27 <vengfulsquirrel> Yeah I'm in a virtual server | ||
2008-12-10 13:27 <vengfulsquirrel> So I'm root already | ||
2008-12-10 13:27 <cedk> vengfulsquirrel: as it will put modules in /usr/lib/pythonx.x/site-packages | ||
2008-12-10 13:27 <vengfulsquirrel> and I've already installed the server ... the modules just don't seem to be going with it | ||
2008-12-10 13:27 <cedk> vengfulsquirrel: each module has his own setup.py | ||
2008-12-10 13:28 <vengfulsquirrel> ha oh jesus | ||
2008-12-10 13:28 <vengfulsquirrel> that's the problem | ||
2008-12-10 13:30 <cedk> vengfulsquirrel: but there is the script that extract all in one directory and you don't need to run the setup.py | ||
2008-12-10 13:31 <vengfulsquirrel> yeah but maybe mention that in the INSTALL of the trytond | ||
2008-12-10 13:31 <cedk> vengfulsquirrel: or you can use easy_install from pypi | ||
2008-12-10 13:31 <cedk> vengfulsquirrel: or the gentoo-overlay | ||
2008-12-10 13:47 <CIA-54> tryton: C?dric Krier <ced@b2ck.com> default * 1105:a30eee88ce85 tryton/tryton/gui/window/view_form/view/form_gtk/interface.py: | ||
2008-12-10 13:47 <CIA-54> tryton: Fix color on form widget | ||
2008-12-10 13:47 <CIA-54> tryton: Now the color from default style is stored on first use | ||
2008-12-10 13:47 <CIA-54> tryton: C?dric Krier <ced@b2ck.com> default * 1106:85a79a5d457b tryton/tryton/gui/window/view_form/view/form_gtk/one2many.py: Don't set model modified when adding record to dialog screen | ||
2008-12-10 13:52 <vengfulsquirrel> I can't use the overlay until amd64 is set though and with easy_install I still have to do it on a module by module basis. | ||
2008-12-10 13:53 <cedk> vengfulsquirrel: no dependencies are resolved with easy_install | ||
2008-12-10 13:53 <cedk> vengfulsquirrel: so if you install trytond_account_invoice, it will install also trytond_account, ... | ||
2008-12-10 13:54 <vengfulsquirrel> ohh | ||
2008-12-10 13:55 <cedk> vengfulsquirrel: and I think it must be possible to force install with overlay | ||
2008-12-10 13:56 <vengfulsquirrel> Yeah unmask didn't seem to work | ||
2008-12-10 13:56 <vengfulsquirrel> but the server runs on amd64 | ||
2008-12-10 13:56 <vengfulsquirrel> I just haven't checked the modules | ||
2008-12-10 13:56 <vengfulsquirrel> and I haven't checked the server in complete thoroughness besides connecting to it with the client. | ||
2008-12-10 13:57 <vengfulsquirrel> Is there a init script anywhere for trytond or will it be trivial to create? | ||
2008-12-10 13:57 <vengfulsquirrel> *an | ||
2008-12-10 13:58 <cedk> LordVan: ping | ||
2008-12-10 13:59 <cedk> vengfulsquirrel: init script depend of each distribution, so the ebuild provide an init script for Gentoo | ||
2008-12-10 13:59 <cedk> vengfulsquirrel: there is packages in progress for Debian and RedHat | ||
2008-12-10 14:01 <vengfulsquirrel> Okay well thanks for all the help, its pretty late here like 5 AM, so I'll try to force install it tomorrow using the overlay and tell you how it goes. The code base looks a lot cleaner though, I'm excited. | ||
2008-12-10 14:02 <cedk> vengfulsquirrel: yes can try with: ACCEPT_KEYWORDS=~x86 emerge trytond | ||
2008-12-10 14:02 <vengfulsquirrel> Yeah it works with x86 it just won't emerge with amd64. | ||
2008-12-10 14:02 <cedk> vengfulsquirrel: but becarefull, it can break the install | ||
2008-12-10 14:03 <cedk> vengfulsquirrel: I mean, you will be able to install it if you set ACCEPT_KEYWORDS to ~x86 | ||
2008-12-10 14:03 <cedk> vengfulsquirrel: on amd64 | ||
2008-12-10 14:03 <LordVan> cedk, yes? | ||
2008-12-10 14:03 <cedk> vengfulsquirrel: but it is not a recommanded way | ||
2008-12-10 14:03 <cedk> LordVan: it was for the missing keywords | ||
2008-12-10 14:04 <cedk> LordVan: but I think I find a solution | ||
2008-12-10 14:04 <cedk> LordVan: vengfulsquirrel can not install tryton with the overlay because it miss ~amd64 | ||
2008-12-10 14:05 <LordVan> i know that | ||
2008-12-10 14:05 <LordVan> i put the packages in /etc/portage/packages.keywords/tryton.keywords :D | ||
2008-12-10 14:05 <LordVan> ^_^ | ||
2008-12-10 14:06 <vengfulsquirrel> Is there a known problem with amd64? | ||
2008-12-10 14:06 <LordVan> works4me (tm) | ||
2008-12-10 14:06 <LordVan> :d | ||
2008-12-10 14:06 <LordVan> it is python so there shouldn't be any issues | ||
2008-12-10 14:06 <LordVan> i think he means if u emerge any C dependencies | ||
2008-12-10 14:06 <cedk> LordVan: ha ok, I will add amd64 keywords so | ||
2008-12-10 14:07 <LordVan> cedk, mind u i didnt' try all modules, cuz i don't know how to use them :D but the app and server run fine and i added some addresses,.. ;) | ||
2008-12-10 14:07 <LordVan> i'm quite ok to be your amd64 tester for now if u want me to .. gotta mail me when u update though ;) | ||
2008-12-10 14:08 <cedk> LordVan: module are just in python | ||
2008-12-10 14:08 <LordVan> exactly | ||
2008-12-10 14:08 <LordVan> :) | ||
2008-12-10 14:09 <vengfulsquirrel> Okay well I'll try the overlay tomorrow, thanks for the help. Good night | ||
2008-12-10 14:09 <LordVan> i do have some general size related issues, but i gotta try it on another machine as my laptop got (a) a funny resolution, (b) weirdish dpi i think cuz i got a small screen | ||
2008-12-10 14:10 <cedk> LordVan: you mean that it is too small? | ||
2008-12-10 14:15 <LordVan> 12" :D | ||
2008-12-10 14:15 <LordVan> the dialogue box isn't big enough for all the elements to fit .. they overlap | ||
2008-12-10 14:15 <LordVan> want a screenshot? | ||
2008-12-10 14:16 <cedk> LordVan: why not? | ||
2008-12-10 14:16 <cedk> LordVan: but 12" is small | ||
2008-12-10 14:16 <CIA-54> tryton: C?dric Krier <ced@b2ck.com> default * 138:218620e17a53 gentoo-overlay/app-office/tryton/ (Manifest tryton-1.0.0.ebuild tryton-1.0.1.ebuild): Add amd64 keyword | ||
2008-12-10 14:16 <CIA-54> tryton: C?dric Krier <ced@b2ck.com> default * 139:708b6357fdb9 gentoo-overlay/app-office/trytond/ (Manifest trytond-1.0.0.ebuild trytond-1.0.1-r1.ebuild): Add amd64 keyword | ||
2008-12-10 14:16 <CIA-54> tryton: C?dric Krier <ced@b2ck.com> default * 140:a360c09d4ff2 gentoo-overlay/app-tryton/account/ (Manifest account-1.0.0.ebuild account-1.0.1.ebuild): Add amd64 keyword | ||
2008-12-10 14:16 <CIA-54> tryton: C?dric Krier <ced@b2ck.com> default * 141:940453fe00c5 gentoo-overlay/app-tryton/account_invoice/ (3 files): Add amd64 keyword | ||
2008-12-10 14:16 <CIA-54> tryton: C?dric Krier <ced@b2ck.com> default * 142:e0ce426e16f3 gentoo-overlay/app-tryton/account_product/ (3 files): Add amd64 keyword | ||
2008-12-10 14:16 <CIA-54> tryton: C?dric Krier <ced@b2ck.com> default * 143:6d58815c9494 gentoo-overlay/app-tryton/analytic_account/ (3 files): Add amd64 keyword | ||
2008-12-10 14:16 <CIA-54> tryton: C?dric Krier <ced@b2ck.com> default * 144:a97a2cc0f316 gentoo-overlay/app-tryton/analytic_invoice/ (3 files): Add amd64 keyword | ||
2008-12-10 14:16 <CIA-54> tryton: C?dric Krier <ced@b2ck.com> default * 145:0bae4e575a03 gentoo-overlay/app-tryton/analytic_purchase/ (3 files): Add amd64 keyword | ||
2008-12-10 14:16 <CIA-54> tryton: C?dric Krier <ced@b2ck.com> default * 146:5216fa784865 gentoo-overlay/app-tryton/analytic_sale/ (3 files): Add amd64 keyword | ||
2008-12-10 14:16 <CIA-54> tryton: C?dric Krier <ced@b2ck.com> default * 147:632e4c31ae09 gentoo-overlay/app-tryton/company/ (Manifest company-1.0.0.ebuild company-1.0.1.ebuild): Add amd64 keyword | ||
2008-12-10 14:16 <CIA-54> tryton: C?dric Krier <ced@b2ck.com> default * 148:7690f8f45322 gentoo-overlay/app-tryton/country/ (Manifest country-1.0.0.ebuild country-1.0.1.ebuild): Add amd64 keyword | ||
2008-12-10 14:16 <CIA-54> tryton: C?dric Krier <ced@b2ck.com> default * 149:973cfcea9e23 gentoo-overlay/app-tryton/currency/ (Manifest currency-1.0.0.ebuild currency-1.0.1.ebuild): Add amd64 keyword | ||
2008-12-10 14:17 <CIA-54> tryton: C?dric Krier <ced@b2ck.com> default * 150:b10a18ee3523 gentoo-overlay/app-tryton/google_maps/ (Manifest google_maps-1.0.0.ebuild google_maps-1.0.1.ebuild): Add amd64 keyword | ||
2008-12-10 14:17 <CIA-54> tryton: C?dric Krier <ced@b2ck.com> default * 151:be1747ad076d gentoo-overlay/app-tryton/party/ (Manifest party-1.0.0.ebuild party-1.0.1.ebuild): Add amd64 keyword | ||
2008-12-10 14:17 <CIA-54> tryton: C?dric Krier <ced@b2ck.com> default * 152:94a6078fc175 gentoo-overlay/app-tryton/product/ (Manifest product-1.0.0.ebuild product-1.0.1.ebuild): Add amd64 keyword | ||
2008-12-10 14:17 <CIA-54> tryton: C?dric Krier <ced@b2ck.com> default * 153:b40f099c8743 gentoo-overlay/app-tryton/purchase/ (Manifest purchase-1.0.0.ebuild purchase-1.0.1.ebuild): Add amd64 keyword | ||
2008-12-10 14:17 <CIA-54> tryton: C?dric Krier <ced@b2ck.com> default * 154:5baee9415c4d gentoo-overlay/app-tryton/sale/ (Manifest sale-1.0.0.ebuild sale-1.0.1.ebuild): Add amd64 keyword | ||
2008-12-10 14:17 <CIA-54> tryton: C?dric Krier <ced@b2ck.com> default * 155:50e79fa6e782 gentoo-overlay/app-tryton/stock/ (Manifest stock-1.0.0.ebuild stock-1.0.1.ebuild): Add amd64 keyword | ||
2008-12-10 14:17 <CIA-54> tryton: C?dric Krier <ced@b2ck.com> default * 156:472e4fb7f508 gentoo-overlay/app-tryton/stock_supply/ (Manifest stock_supply-1.0.0.ebuild stock_supply-1.0.1.ebuild): Add amd64 keyword | ||
2008-12-10 14:17 <CIA-54> tryton: C?dric Krier <ced@b2ck.com> default * 157:eae852579d77 gentoo-overlay/dev-python/relatorio/ (4 files): Add amd64 keyword | ||
2008-12-10 14:17 <CIA-54> tryton: C?dric Krier <ced@b2ck.com> default * 158:19b37aeae97f gentoo-overlay/dev-python/vatnumber/ (Manifest vatnumber-0.1.ebuild vatnumber-0.2.ebuild): Add amd64 keyword | ||
2008-12-10 14:31 <LordVan> cedk, http://www.lordvan.com/pics/tryton_createDB_toosmalldialogue.png | ||
2008-12-10 14:38 <CIA-54> tryton: C?dric Krier <ced@b2ck.com> default * 1107:379402a32fdd tryton/tryton/gui/window/dbcreate.py: Remove set_size_request on dbcreate dialog | ||
2008-12-10 14:38 <cedk> LordVan: this changeset must fix | ||
2008-12-10 14:45 <LordVan> cedk, i'll check it out at some point .. since it's nothing critical ;) if i forget feel free to remind me (i got the release installed atm ;)) | ||
2008-12-10 14:47 <cedk> LordVan: I will wait a little before putting this changeset in the 1.0 branch | ||
2008-12-10 14:58 <LordVan> sure :) | ||
2008-12-10 15:39 -!- ikks(n=igor@201.244.188.98) has joined #tryton | ||
2008-12-10 15:45 -!- rvalyi(i=58aaece0@gateway/web/ajax/mibbit.com/x-608e5ed6fa1880c5) has joined #tryton | ||
2008-12-10 16:19 <LordVan> cyas | ||
2008-12-10 16:56 <CIA-54> tryton: C?dric Krier <ced@b2ck.com> default * 397:7e0e4fd7d180 stock/packing.py: Fix location domain for internal packing | ||
2008-12-10 17:25 <CIA-54> tryton: bertrand.chenal@b2ck.com * r372 /wiki/InstallationMercurial.wiki: Edited wiki page through web user interface. | ||
2008-12-10 17:25 <CIA-54> tryton: bertrand.chenal@b2ck.com * r373 /wiki/InstallationMercurial.wiki: Edited wiki page through web user interface. | ||
2008-12-10 17:25 <CIA-54> tryton: bertrand.chenal@b2ck.com * r374 /wiki/InstallationMercurial.wiki: Edited wiki page through web user interface. | ||
2008-12-10 17:43 <CIA-54> tryton: Bertrand Chenal <bch@b2ck.com> default * 330:c8b3b2de43af account/INSTALL: Typo | ||
2008-12-10 17:43 <CIA-54> tryton: Bertrand Chenal <bch@b2ck.com> default * 20:337004630a2b account_be/INSTALL: Typo | ||
2008-12-10 17:43 <CIA-54> tryton: Bertrand Chenal <bch@b2ck.com> default * 219:cc8d0708ac99 account_invoice/INSTALL: Typo | ||
2008-12-10 17:43 <CIA-54> tryton: Bertrand Chenal <bch@b2ck.com> default * 22:0cb2b4eb1b72 account_product/INSTALL: Typo | ||
2008-12-10 17:43 <CIA-54> tryton: Bertrand Chenal <bch@b2ck.com> default * 80:696210263179 account_statement/INSTALL: Typo | ||
2008-12-10 17:43 <CIA-54> tryton: Bertrand Chenal <bch@b2ck.com> default * 54:5d64fa4b5247 analytic_account/INSTALL: Typo | ||
2008-12-10 17:43 <CIA-54> tryton: Bertrand Chenal <bch@b2ck.com> default * 34:2e9a16dfae02 analytic_invoice/INSTALL: Typo | ||
2008-12-10 17:43 <CIA-54> tryton: Bertrand Chenal <bch@b2ck.com> default * 35:afd3e8822d85 analytic_purchase/INSTALL: Typo | ||
2008-12-10 17:43 <CIA-54> tryton: Bertrand Chenal <bch@b2ck.com> default * 23:6d1697edb370 analytic_sale/INSTALL: Typo | ||
2008-12-10 17:43 <CIA-54> tryton: Bertrand Chenal <bch@b2ck.com> default * 111:98ffdbc73219 company/INSTALL: Typo | ||
2008-12-10 17:43 <CIA-54> tryton: Bertrand Chenal <bch@b2ck.com> default * 19:3d7a899df989 country/INSTALL: Typo | ||
2008-12-10 17:43 <CIA-54> tryton: Bertrand Chenal <bch@b2ck.com> default * 49:9c0af67b8f76 currency/INSTALL: Typo | ||
2008-12-10 17:43 <CIA-54> tryton: Bertrand Chenal <bch@b2ck.com> default * 30:5941de80edd8 google_maps/INSTALL: Typo | ||
2008-12-10 17:43 <CIA-54> tryton: Bertrand Chenal <bch@b2ck.com> default * 213:96167211e3dd party/INSTALL: Typo | ||
2008-12-10 17:43 <CIA-54> tryton: Bertrand Chenal <bch@b2ck.com> default * 118:8aa4835665ad product/INSTALL: Typo | ||
2008-12-10 17:43 <CIA-54> tryton: Bertrand Chenal <bch@b2ck.com> default * 38:d910cc2a9c69 project/INSTALL: Typo | ||
2008-12-10 17:43 <CIA-54> tryton: Bertrand Chenal <bch@b2ck.com> default * 21:80b5b6d07b33 project_revenue/INSTALL: Typo | ||
2008-12-10 17:43 <CIA-54> tryton: Bertrand Chenal <bch@b2ck.com> default * 177:8630b21fbae6 purchase/INSTALL: Typo | ||
2008-12-10 17:43 <CIA-54> tryton: Bertrand Chenal <bch@b2ck.com> default * 80:36264695727b sale/INSTALL: Typo | ||
2008-12-10 18:20 <cedk> bechamel: OpenERP uses now psycopg2 :-) | ||
2008-12-10 18:20 <cedk> bechamel: but they make the switch during stabilisation of the code | ||
2008-12-10 18:21 <cedk> I really don't understand how they stabilize the code, they make more developpement between the rc than before | ||
2008-12-10 18:22 <bechamel> cedk: it's not the usual way for RC ? | ||
2008-12-10 18:22 <cedk> bechamel: now you fix the bug only not new developpement | ||
2008-12-10 18:22 <bechamel> ACTION was kidding | ||
2008-12-10 18:24 <cedk> do you remember that Fabien arguments on the forum that they don't want to implement the locale customization | ||
2008-12-10 18:24 <cedk> and http://bazaar.launchpad.net/%7Eopenerp/openobject-server/trunk/revision/1362 | ||
2008-12-10 18:24 <cedk> they do it :-) | ||
2008-12-10 18:25 <bechamel> cedk: maybe he didn't understood the issue when he post on the forum | ||
2008-12-10 18:25 <cedk> the forum post: http://openerp.com/forum/topic8584.html | ||
2008-12-10 18:26 <cedk> this is the real commit: http://bazaar.launchpad.net/%7Eopenerp/openobject-server/trunk/revision/1355.1.7 | ||
2008-12-10 18:28 <cedk> bechamel: but he says "I think it's better to keep using standards" | ||
2008-12-10 18:28 <cedk> but they don't use it to format the date in the client interface | ||
2008-12-10 18:29 <cedk> so you can have customization on reports but not in client interface :-) | ||
2008-12-10 18:39 -!- Timitos(n=Timitos@88.217.184.172) has joined #tryton | ||
2008-12-10 19:03 <CIA-54> tryton: cedric.krier@b2ck.com * r375 /wiki/InstallationMercurial.wiki: Edited wiki page through web user interface. | ||
2008-12-10 19:03 <CIA-54> tryton: cedric.krier@b2ck.com * r376 /wiki/InstallationMercurial.wiki: Edited wiki page through web user interface. | ||
2008-12-10 19:06 <CIA-54> tryton: matb roundup * #650/Items reappearing untranslated in new database: [resolved] Seems to be solved. | ||
2008-12-10 19:12 <CIA-54> tryton: matb roundup * #656/Translation of title in packing_out.odt: [new] For the title 'Packing List' I get two separate items to translate "List" "Liste" "False" "stock.packing.out" "Deutsch" "ODT" "0" "stock" ... | ||
2008-12-10 19:46 <CIA-54> tryton: Bertrand Chenal <bch@b2ck.com> default * 146:f167fce3f4b9 stock_supply/INSTALL: Typo | ||
2008-12-10 19:49 -!- udono(n=udono@dynamic-unidsl-85-197-16-94.westend.de) has joined #tryton | ||
2008-12-10 19:55 <CIA-54> tryton: Bertrand Chenal <bch@b2ck.com> default * 178:bb3011580be9 purchase/purchase.xml: Added menuitem | ||
2008-12-10 19:55 <CIA-54> tryton: Bertrand Chenal <bch@b2ck.com> default * 81:c2d695f12c33 sale/sale.xml: Added menuitem | ||
2008-12-10 20:05 -!- X0d_of_N0d_(i=C-C_C-X@gateway/tor/x-0cd4048ff35690b8) has joined #tryton | ||
2008-12-10 20:10 -!- mmarshall(n=mmarshal@adsl-99-186-213-109.dsl.rcsntx.sbcglobal.net) has joined #tryton | ||
2008-12-10 20:26 -!- leyoda(n=leyoda@pic33-1-87-88-179-8.dsl.club-internet.fr) has joined #tryton | ||
2008-12-10 20:31 <CIA-54> tryton: matb roundup * #657/Additional fields of inherited report also created for parent module: [new] Not absolutely sure, if this is a bug, but after several days of try and error I think it is one: The attached invoice.xml belongs to a cus ... | ||
2008-12-10 20:35 <X0d_of_N0d_> res.company under tinyerp would be company.company under tryton? | ||
2008-12-10 20:35 <X0d_of_N0d_> damn, posted it to the wrong channel a min a go | ||
2008-12-10 20:35 <X0d_of_N0d_> lol | ||
2008-12-10 20:53 <CIA-54> tryton: matb roundup * #658/Make translation of title of invoice report independent from invoice.type: [new] see http://groups.google.com/group/tryton/t/facd1b774c45d1fe | ||
2008-12-10 21:22 <Timitos> X0d_of_N0d_: yes company.company is like res.company under tinyerp. but i havenĀ“t looked at the differences in detail yet. | ||
2008-12-10 21:23 <X0d_of_N0d_> I was going to try to port the ldap_users module over | ||
2008-12-10 21:23 <X0d_of_N0d_> it looks like some things have changed in the authentication system too | ||
2008-12-10 21:24 <Timitos> X0d_of_N0d_: i think you should talk with udono. he is sitting on this too | ||
2008-12-10 21:24 <Timitos> X0d_of_N0d_: yes i think there were bigger changes in this area | ||
2008-12-10 21:24 <X0d_of_N0d_> ok | ||
2008-12-10 21:25 <X0d_of_N0d_> when does udono usually show up? | ||
2008-12-10 21:27 <Timitos> this is different. but most time he is online from about 8-18 GMT+1. sometimes later too. but today he seems to be away | ||
2008-12-10 21:28 <Timitos> X0d_of_N0d_: i will inform him about your work. he is my partner | ||
2008-12-10 21:28 -!- vengfulsquirrel(n=ian@c-67-170-212-242.hsd1.ca.comcast.net) has joined #tryton | ||
2008-12-10 21:28 <X0d_of_N0d_> cool, yeah, I modded the terp one a bit to make it work with our server, I'd love to see what he's doing with his code | ||
2008-12-10 21:29 <X0d_of_N0d_> Timitos: thanks | ||
2008-12-10 21:29 <Timitos> X0d_of_N0d_: you are welcome | ||
2008-12-10 21:30 <Timitos> X0d_of_N0d_: he is just in the beginning so you should expect too much for now | ||
2008-12-10 21:31 <X0d_of_N0d_> Timitos: If he knows tryton the ldap stuff is easy, I'd be glad to help | ||
2008-12-10 21:31 <Timitos> X0d_of_N0d_: this would be great! thx | ||
2008-12-10 21:32 <X0d_of_N0d_> Timitos: np | ||
2008-12-10 21:35 <X0d_of_N0d_> so.... are the passwords for tinyerp's users stored in clear text in the db? It looks like tryton stores a sha hash | ||
2008-12-10 21:35 <X0d_of_N0d_> huh... they are | ||
2008-12-10 21:36 <X0d_of_N0d_> huh | ||
2008-12-10 21:37 <Timitos> :-) a nice detail | ||
2008-12-10 21:41 <CIA-54> tryton: Timitos roundup * #604/NoneType-Exception if pressing Ctrl-C in console: @htgoebel could you please test again? i think the issue is solved. | ||
2008-12-10 21:44 -!- Gedd(n=ged@233.86-242-81.adsl-dyn.isp.belgacom.be) has joined #tryton | ||
2008-12-10 21:46 -!- nside(n=nside@206-248-163-68.dsl.teksavvy.com) has joined #tryton | ||
2008-12-10 21:46 <X0d_of_N0d_> How does Tryton protect against sql injection? I'm digging through the tinyerp code right now trying to figure that crap out | ||
2008-12-10 21:47 <X0d_of_N0d_> it seems like you should be able to inject sql... or is that cleaned up by psycopg? | ||
2008-12-10 21:47 <ikks> psycopg does not clean sql injection afaik | ||
2008-12-10 21:49 <X0d_of_N0d_> huh.... | ||
2008-12-10 21:54 <X0d_of_N0d_> ACTION heads off to get some lunch | ||
2008-12-10 21:55 <bechamel> ikks: psycopg escape automaticaly each params for somethink like "cursor.execute('select %s from %s where %s = 1', params)" | ||
2008-12-10 21:56 <bechamel> ikks: X0d_of_N0d_ : the problem is when the params are used to construct the string containing the query | ||
2008-12-10 21:59 <bechamel> example, line 1340 in orm.py of openerp-server: cr.execute('delete from wkf_instance where res_type=%s and res_id in ('+','.join(map(str, ids))+')', (self._name, )) | ||
2008-12-10 22:01 <bechamel> self._name will be escaped (but it's not a problem because self._name comes from the code, but ids might containt user data, hence the risk of injection | ||
2008-12-10 22:09 -!- mmarshall(n=mmarshal@99.186.213.109) has joined #tryton | ||
2008-12-10 22:10 -!- Gedd(n=ged@77.109.116.111) has joined #tryton | ||
2008-12-10 22:25 <CIA-54> tryton: C?dric Krier <ced@b2ck.com> default * 1343:b524c5b3d4c5 trytond/trytond/osv/orm.py: Fix guidelines | ||
2008-12-10 22:25 <CIA-54> tryton: C?dric Krier <ced@b2ck.com> default * 1344:488247aad1d0 trytond/ (5 files in 3 dirs): Allow to inherit views from an other model | ||
2008-12-10 22:25 <CIA-54> tryton: C?dric Krier <ced@b2ck.com> default * 1345:fea842a62cd0 trytond/trytond/ir/sequence.xml: Use inherit view from sequence for sequence strict | ||
2008-12-10 22:25 <CIA-54> tryton: ced roundup * #649/Unable to create new employee: [resolved] Fix with changeset 5f7988bc8333 | ||
2008-12-10 22:26 <CIA-54> tryton: C?dric Krier <ced@b2ck.com> default * 112:5f7988bc8333 company/company.xml: Use inherit view of party for company and employee for issue649 | ||
2008-12-10 22:26 <CIA-54> tryton: C?dric Krier <ced@b2ck.com> default * 113:e1b4895c6c78 company/company.xml: Add select for company on employee | ||
2008-12-10 22:26 <CIA-54> tryton: C?dric Krier <ced@b2ck.com> default * 114:807867c27f2f company/: merge | ||
2008-12-10 22:36 <X0d_of_N0d_> bechamel: ok, that's kind of what I was thinking | ||
2008-12-10 22:36 <CIA-54> tryton: C?dric Krier <ced@b2ck.com> default * 398:3a064677a655 stock/packing_out.odt: Fix "Packing List" split in translation for issue656 | ||
2008-12-10 22:37 <CIA-54> tryton: ced roundup * #656/Translation of title in packing_out.odt: [resolved] Fix with changeset 3a064677a655 | ||
2008-12-10 22:37 <X0d_of_N0d_> bechamel: it still seems like the auth stuff should query for the user, get the password (or in tryton's case the hash), and then compare the password against the pw/hash | ||
2008-12-10 22:37 <X0d_of_N0d_> but if psycopg is taking care of everything then it's probably fine | ||
2008-12-10 22:38 <cedk> X0d_of_N0d_: this is made in trytond/security.py | ||
2008-12-10 22:38 <cedk> X0d_of_N0d_: login or check are call on each request of the user | ||
2008-12-10 22:40 <X0d_of_N0d_> cedk: right, but it's done in one query... | ||
2008-12-10 22:40 <X0d_of_N0d_> SELECT id FROM res_user WHERE login = %s and password = %s and active | ||
2008-12-10 22:40 <X0d_of_N0d_> instead of SELECT id FROM res_user ' <newline/> | 'WHERE login = %s and password = %s and active | ||
2008-12-10 22:40 <X0d_of_N0d_> err | ||
2008-12-10 22:41 <X0d_of_N0d_> SELECT id,passwd FROM res_user WHERE login = %s and active | ||
2008-12-10 22:41 <cedk> X0d_of_N0d_: what is the difference ? | ||
2008-12-10 22:41 <X0d_of_N0d_> then passwd==passwd_sha | ||
2008-12-10 22:43 <X0d_of_N0d_> if psycopg somehow failed to correctly clean the sql the first one would still let you in | ||
2008-12-10 22:43 <X0d_of_N0d_> however even if psycopg failed, the second one would not let you in | ||
2008-12-10 22:43 <X0d_of_N0d_> unless you had the right password anyway | ||
2008-12-10 22:44 <cedk> X0d_of_N0d_: even with the second as you use %s for the user name | ||
2008-12-10 22:44 <cedk> s/user name/login/ | ||
2008-12-10 22:45 <X0d_of_N0d_> cedk: right, but that means that the query would respond with hte pw and id of all the users | ||
2008-12-10 22:46 <X0d_of_N0d_> then the sql.fetchone would give the first id and pw | ||
2008-12-10 22:46 <X0d_of_N0d_> then you'd compare that pw to the pw returned by the query | ||
2008-12-10 22:46 <X0d_of_N0d_> and you'd dump the user unless the password matched | ||
2008-12-10 22:47 <X0d_of_N0d_> whereas with the other way you could just inject login=admin and passord=1 or login=admin and get the admin user, then it'd drop you in | ||
2008-12-10 22:47 <X0d_of_N0d_> of course this is all in the hypothetical instance that psycopg somehow failed... but... | ||
2008-12-10 22:50 <cedk> X0d_of_N0d_: any way, it will be difficult to provide a string that once "sha"ed will produce a sql clause that will failed the query | ||
2008-12-10 22:50 <X0d_of_N0d_> I'm just saying the current way relies on psycopg to work, where returning the password and manually checking it would not ever be exploitable... | ||
2008-12-10 22:51 <X0d_of_N0d_> cedk: right, but it would be easy to provide a login that would allow injection.... | ||
2008-12-10 22:52 <cedk> X0d_of_N0d_: if you can make sql injection, it is possible to inject " login = 'admin' and (update res_user set password = 'foo')" | ||
2008-12-10 22:52 <X0d_of_N0d_> hum... | ||
2008-12-10 22:52 <X0d_of_N0d_> yeah, maybe not | ||
2008-12-10 22:53 <X0d_of_N0d_> cedk: yeah, fair enough | ||
2008-12-10 22:53 <cedk> X0d_of_N0d_: this will work " login = 'admin'; update res_user set password = 'foo'" | ||
2008-12-10 22:53 <cedk> X0d_of_N0d_: with two tries | ||
2008-12-10 22:54 <cedk> X0d_of_N0d_: once you have SQL injection, you can make anythings | ||
2008-12-10 22:54 <cedk> X0d_of_N0d_: and I would prefer to let psycopg handle this as there is more user that will check psycopg than our code | ||
2008-12-10 22:56 <X0d_of_N0d_> cedk: yeah, that makes sense | ||
2008-12-10 22:57 <X0d_of_N0d_> cedk: but I will say that using hashes is definitely an improvment over the way tinyerp works, I definitely have to give you props man | ||
2008-12-10 23:03 <cedk> X0d_of_N0d_: and session :-) | ||
2008-12-10 23:06 <X0d_of_N0d_> cedk: oh yeah, and the session time out... and the fact that there isn't a ton of duplicate code in login and check | ||
2008-12-10 23:06 <X0d_of_N0d_> etc... | ||
2008-12-10 23:08 -!- juanfer(n=juanfer@201.244.188.98) has joined #tryton | ||
2008-12-10 23:15 <X0d_of_N0d_> does tryton have any of it's own documentation on creating modules, or should one refer to the tinyerp docs? | ||
2008-12-10 23:21 <bechamel> X0d_of_N0d_: there is no tutorials yet, but you will find some background here http://www.tryton.org/doc/branches/1.0/trytond/doc/index.html | ||
2008-12-10 23:26 <X0d_of_N0d_> bechamel: thanks | ||
2008-12-10 23:42 <CIA-54> tryton: X0d_of_N0d roundup * #659/AttributeError: 'bool' object has no attribute 'split': [new] Traceback (most recent call last): File "/trytond/netsvc.py", line 282, in run res = method(*msg[2:]) File "/trytond/web_service/obj ... | ||
2008-12-10 23:42 <X0d_of_N0d_> is it possible to create new forms in tryton? | ||
2008-12-10 23:43 <bechamel> X0d_of_N0d_: yes it is, forms are defined in the xml files | ||
2008-12-10 23:44 <X0d_of_N0d_> I mean through the gui, like in tinyerp | ||
2008-12-10 23:47 <CIA-54> tryton: X0d_of_N0d roundup * #659/AttributeError: 'bool' object has no attribute 'split': [chatting] I tried to create a new form through the gui Form->New, it looks like complete name isn't editable, or is it filled in with the value f ... | ||
2008-12-10 23:49 <bechamel> X0d_of_N0d_: i've never tried to create them from the gui | ||
2008-12-10 23:50 <X0d_of_N0d_> bechamel: I'm reading through the tinyerp book and it mentioned that it could be done | ||
2008-12-10 23:50 <X0d_of_N0d_> bechamel: so I tried it on tinyerp then tried it on tryton | ||
2008-12-10 23:53 <bechamel> X0d_of_N0d_: i just tried to create a new view and it seems t owork | ||
2008-12-10 23:53 <X0d_of_N0d_> lemme try it with the latest svn | ||
2008-12-10 23:53 <X0d_of_N0d_> errr hg | ||
2008-12-10 23:55 <bechamel> X0d_of_N0d_: but i don't see how to change which view is linked to which action | ||
2008-12-10 23:56 <X0d_of_N0d_> wait, what did you do to make it work? | ||
2008-12-10 23:56 <X0d_of_N0d_> Form-> new | ||
2008-12-10 23:57 <X0d_of_N0d_> type in Menu: asdfwhatever | ||
2008-12-10 23:57 <X0d_of_N0d_> save | ||
2008-12-10 23:57 <X0d_of_N0d_> backtrace | ||
2008-12-10 23:57 <bechamel> X0d_of_N0d_: I took a view, I duplicated it and modified the dupicate and give it a bigger priority | ||
2008-12-10 23:58 <bechamel> X0d_of_N0d_: now when i open a party list (not from the menu but when i open a field pointing to a party from another form) i can see the new view | ||
2008-12-10 23:59 <bechamel> X0d_of_N0d_: menuitem are linked to action and action are linked to views so creating a new one doesnt change anything when opening forms from the menu |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!