#gemrb@irc.freenode.net logs for 3 Dec 2010 (GMT)

Archive Today Yesterday Tomorrow
GemRB homepage


[02:20:41] --> pupnik_ has joined #GemRb
[02:24:04] <-- pupnik has left IRC (Ping timeout: 260 seconds)
[02:29:19] <Lightkey> edheldil: I also have directories with spaces on the DVD
[02:50:43] --- barraAway is now known as barra_home
[04:39:04] <-- barra_home has left IRC (Quit: Verlassend)
[06:14:48] --> edheldil_ has joined #GemRb
[06:24:34] <-- edheldil_ has left IRC (Read error: Operation timed out)
[06:31:24] --- pupnik_ is now known as pupnik
[07:04:59] <-- pupnik has left IRC (Quit: leaving)
[07:07:55] --> edheldil_ has joined #GemRb
[07:12:53] <-- edheldil_ has left IRC (Ping timeout: 255 seconds)
[08:11:32] --> lynxlynxlynx has joined #GemRb
[08:11:32] --- ChanServ gives channel operator status to lynxlynxlynx
[08:16:24] --> Bo_Thomsen has joined #GemRb
[08:31:39] --> lubos has joined #GemRb
[08:33:13] <edheldil> Lightkey: ok, thanks, I will fix it somehow
[08:54:58] --> SiENcE has joined #GemRb
[09:02:39] <-- |Cable| has left IRC (Remote host closed the connection)
[09:24:39] <-- SiENcE has left IRC (Ping timeout: 260 seconds)
[11:19:45] --> barra_home has joined #GemRb
[12:11:24] <CIA-5> GemRB: 03lynxlupodian * r3e83b84a78bf 10gemrb/gemrb/GUIScripts/ (bg1/GUIINV.py bg2/GUIINV.py iwd/GUIINV.py iwd2/GUIINV.py): refresh the portrait window when reassigning it, so the changes propagate
[12:12:06] <CIA-5> GemRB: 03lynxlupodian * rf4f7966818b3 10gemrb/gemrb/includes/ie_types.h: android doesn't define SIZEOF_INT
[12:12:18] <CIA-5> GemRB: 03lynxlupodian * r8a6cc2ac7463 10gemrb/gemrb/ (4 files in 3 dirs): added swab from glibc for android, who's unistd.h lacks it
[12:13:33] <barra_home> hola
[12:13:45] <barra_home> where is gemrb's git repo hosted?
[12:14:00] <lynxlynxlynx> sf
[12:15:52] <barra_home> how is it in terms of general performance?
[12:16:11] <barra_home> I remember that SVN performance at sf was rather poor at some point due the mass of projects they host
[12:16:55] <fuzzie> i haven't yet been driven to insanity by it
[12:17:04] <fuzzie> which is a vast improvement over svn
[12:17:22] <barra_home> :-)
[12:17:27] <fuzzie> but it might well be just because git doesn't really have to talk to sf that much..
[12:17:28] <barra_home> we consider to switch to git as well
[12:17:41] <barra_home> the lack of trac integration sucks though
[12:18:41] <barra_home> there is a git plugin (unofficial though) for trac, but they don't offer it at sf :-(
[12:21:29] <barra_home> what kind of bug/issue tracker do you use?
[12:21:36] <lynxlynxlynx> mostly the wiki
[12:21:41] <lynxlynxlynx> the sf one is annoying
[12:22:02] <lynxlynxlynx> don't forget there are other git hosting providers
[12:22:16] <barra_home> yep, but no free hosters with trac integration
[12:22:17] <lynxlynxlynx> gna has integration with its tracker
[12:22:28] <lynxlynxlynx> it's luckily not trac :P
[12:22:42] <barra_home> hey, what's wrong with trac, I think it's great :-)
[12:22:46] <fuzzie> this is intended for rarpg and their crazy ticket obsession? :P
[12:22:50] <fuzzie> parpg, even
[12:23:58] <tomprince> github has an issue tracker, which I would guess is integrated.
[12:24:51] <barra_home> we took a look at github
[12:25:06] <barra_home> the issue tracker is the best tracker I've ever seen, unlimited possibilities
[12:25:18] <barra_home> you can actually enter a headline for the bug and a description!
[12:25:25] <barra_home> impressive
[12:25:53] <barra_home> I didn't know we had a reputation to be ticket obsessed fuzzie :-)
[12:26:07] <fuzzie> well, i thought there was a requirement that every commit had an associated ticket
[12:26:16] <fuzzie> but the wiki tells me i am wrong, maybe i'm thinking of something else
[12:26:20] <lynxlynxlynx> a title and a description of a bug? erm, that's part of the basics
[12:26:42] <fuzzie> lynxlynxlynx: you would think, but try finding something which lets people edit the description..
[12:26:47] <barra_home> yep, but _nothing_ beyond that lynxlynxlynx
[12:27:15] <barra_home> you can't enter a component, you can't assign the ticket to a developer, you can't state ticket dependencies
[12:27:16] <fuzzie> i quite like Google Code's issue tracker, but i doubt it integrates nicely with anything, and you are stuck with Google
[12:27:20] <lynxlynxlynx> i don't think editing the description is a good idea
[12:27:33] <barra_home> depends lynxlynxlynx
[12:27:43] <barra_home> if you're a developer, why not edit it?
[12:27:51] <lynxlynxlynx> when a silly bug is streamlined, new ones can be opened as replacements or dependencies
[12:28:01] <fuzzie> well, most existing systems confuse the idea of a 'first comment' and the description
[12:28:03] <barra_home> better than reading half a dozen comments that the ticket description is outdated and that the actual wording should be
[12:28:16] <fuzzie> so people have to read, yes, half a dozen comments, what barra said :)
[12:28:24] <lynxlynxlynx> not if you open a new bug
[12:28:29] <barra_home> sure, but why
[12:28:44] <lynxlynxlynx> because editing the description won't kill the comments
[12:28:49] <fuzzie> sure, but then you're kinda defeating the advantage of having the bug tracker in the first place :)
[12:29:05] <fuzzie> but i think it really depends on the project, anyway
[12:29:13] <lynxlynxlynx> not really, you've just made a step into making the content more sensible
[12:29:26] <barra_home> usually it works this way: you create a ticket and you just have a very vague idea how something should be implemented
[12:29:36] <fuzzie> for most things i work on, i really want 'one ticket == one actual bug'
[12:29:41] <barra_home> while digging deeper, the actual steps become clearer
[12:29:51] <barra_home> we're not only talking about bugs
[12:29:52] <lynxlynxlynx> feature requests are dreamy horrors, yes
[12:30:04] <barra_home> a ticket tracker is for feature requests and enhancements as well
[12:30:08] <fuzzie> and, feature requests can go somewhere else :P
[12:30:10] <lynxlynxlynx> but you can easily make a master bug and link the others in, atleast with bugzilla
[12:30:20] <barra_home> yep, that's useful lynxlynxlynx
[12:30:32] <barra_home> but as said: the github bug tracker doesn't offer that functionality
[12:30:35] <lynxlynxlynx> can also replace the milestone thing ;)
[12:30:46] <barra_home> no milestones for github either
[12:31:17] <lynxlynxlynx> anyway, i don't think this is much of an issue
[12:31:28] <lynxlynxlynx> a "nice to have", not a blocker
[12:31:34] <fuzzie> well, for which project? :)
[12:31:49] <fuzzie> for gemrb, this is all fairly irrelevant
[12:32:19] <lynxlynxlynx> never found much of a need for that in others either
[12:32:42] <lynxlynxlynx> one's on gna with that integration
[12:33:30] <fuzzie> but for projects which are organised entirely around the ticketing system, i think trac might be the only open-source project which seriously tries supporting that workflow?
[12:35:11] <lynxlynxlynx> i don't know, i don't follow the developments
[12:35:19] <lynxlynxlynx> there's plenty of systems out there
[12:36:21] <fuzzie> well, i've only had to deal with that kind of thing in proprietary projects with proprietary ticket systems, so i don't have much of an idea either :)
[12:36:59] <barra_home> so you rather document the tasks to tackle at the wiki?
[12:37:10] <fuzzie> barra_home: for gemrb, there is *far* too much to do
[12:37:36] <barra_home> so how do you resolve that problem?
[12:37:39] <fuzzie> and there are often many bugs reported by end-users with a single underlying issue which might take weeks to understand
[12:38:26] <fuzzie> so we always have a large selection of obvious bugs to pick from, there is not much need to keep track of the rest..
[12:38:27] <lynxlynxlynx> and there's little need for any of the metadata that is usually provided
[12:39:17] <lynxlynxlynx> since it's on our wiki, it also slightly encourages other website contributions
[12:39:27] <lynxlynxlynx> and any issue can be expanded as much as needed
[12:39:38] <barra_home> you can file a ticket at trac as anonymous user as well
[12:39:40] <lynxlynxlynx> and it's pretty fast
[12:40:11] <lynxlynxlynx> and all the submissions can be reviewed in digests, since it is all in a vcs
[12:40:36] <lynxlynxlynx> it's similar to the "change multiple bugs at once" from bugzilla, just more diverse
[12:42:28] <fuzzie> i think in general, it is just useless for engine reimplementation projects to collect bugs (rather than a TODO list) until they are in a fairly advanced state
[12:42:33] <fuzzie> and that is really what we have, a TODO list :)
[12:44:08] <lynxlynxlynx> well, it's mixed
[12:44:34] <lynxlynxlynx> the KISSness of the approach is awesome :)
[12:44:36] <barra_home> alrighty, thanks for filling me in :-)
[12:45:12] <fuzzie> did someone comment out the CanSee in the dialog code?
[12:45:53] <fuzzie> seems like it might (incorrectly, but meh) fix some of the bugs on the list
[12:48:31] <lynxlynxlynx> doesn't look like it
[12:49:18] <edheldil> barra_home: we use trac here at work and I do not like it
[12:49:32] <barra_home> feel free to elaborate edheldil
[12:49:49] <edheldil> and *I hate* when somebody is changing description
[12:50:12] <edheldil> because than you do not know what changed. at least without extra work
[12:51:23] <lynxlynxlynx> it is changing history
[12:53:03] <edheldil> last week I forgot to install some package, because it was added to the list in the description after I have read it and I just overlooked it
[12:54:30] <barra_home> edheldil> because than you do not know what changed. at least without extra work
[12:54:48] <barra_home> that's nonsense, trac has a diff button, you click on it and you see exactly what was changed
[12:55:34] <fuzzie> yes, but that is extra confusing work if you are under time pressure for work :)
[12:56:01] <barra_home> example: http://parpg-trac.cvsdude.com/parpg/ticket/270?action=diff&version=4
[12:56:03] <fuzzie> but i think people changing work request descriptions like that is maybe more a problem with people than the tracker :P
[12:56:39] <barra_home> there is a timeline in trac, every ticket change is also recorded in the ticket itself, you click on the diff link that's displayed for every change and you get the diff
[12:56:40] <edheldil> feel free to use it, if you like it
[12:57:13] <barra_home> it depends on how the permissions are set up edheldil, usually only devs can change the descriptions
[12:57:23] <barra_home> will unregistered users can just add comments
[12:58:03] <edheldil> on a positive note, unlike our other request tracker, that leaks memory like rusty barrel water
[12:58:14] <edheldil> I mean it's small and simple
[12:59:14] <barra_home> yep, I think it's pretty straight forward, yet powerful as it can be extended by plugins
[12:59:42] <barra_home> anyway, didn't want to preach or be a missionary, I was more looking for feedback what devs dislike about it
[13:00:10] <edheldil> I dislike it as an admin ;-)
[13:00:58] <edheldil> Lightkey: I made a new version of the script, it now copes with spaces in names correctly ;-)
[13:02:33] <barra_home> fair enough edheldil :-) thanks for your insight
[13:03:17] <edheldil> well, I am obviously biased :)
[13:04:52] <barra_home> nothing wrong with that, so am I
[13:13:24] <lynxlynxlynx> did you investigate redmine?
[13:15:59] <edheldil> is not it a webmail? Or is it TT as well?
[13:16:18] <lynxlynxlynx> project managment suite
[13:16:20] <barra_home> we took a look at it lynxlynxlynx
[13:16:30] <lynxlynxlynx> fuzzie: the two in BeginDialog?
[13:16:33] <barra_home> seems like a ruby-based trac clone
[13:16:44] <barra_home> and a pretty good one to say that :-)
[13:17:07] <barra_home> we couldn't find git integration for redmine though, hostedredmine.com just offers SVN integration
[13:18:08] <lynxlynxlynx> you could install it yourself
[13:29:24] <barra_home> sure, I'm not too keen on setting up a root server and administering it though
[13:29:41] <barra_home> would eat up quite some time that I could spend otherwhise on the project
[13:30:03] <lynxlynxlynx> no need to set up a server, just set up redmine
[13:30:52] <barra_home> wouldn't it be hosted on a server? :-)
[13:37:39] <edheldil> we used to had in previous company a TT system which assigned you to the ticket if you looked at it
[13:37:53] <edheldil> s/had/have/
[13:39:07] <barra_home> haha
[13:39:19] <barra_home> great excuse for not using the ticket system
[13:39:37] <barra_home> I wanted to have a look at the ticket but I was afraid of having to take it over ...
[13:42:35] <lynxlynxlynx> sure, but you don't have to manage that host
[13:42:40] <edheldil> better still, it happily passed on any html/js that helpdesk copied from customer emails to it, so once we had a ticket that could not be seen, because it immediately redirected you to another page when viewed
[13:43:01] <lynxlynxlynx> all the big project managment players give you enough web space to do anything
[13:43:21] <lynxlynxlynx> hehe
[13:46:30] <barra_home> are we talking about free for open source hosting lynxlynxlynx ?
[13:46:42] <lynxlynxlynx> sure
[13:47:12] <lynxlynxlynx> sf and gna come to mind, since i mostly used them
[13:47:56] <barra_home> good point, we might switch to sf in the long run in case we run out of space
[13:48:10] <barra_home> hopefully they'll provide trac integration for git at that point
[13:49:28] <lynxlynxlynx> if it is just a plugin, submit a ticket err bug
[13:49:39] <lynxlynxlynx> just wishing for it won't change anything
[13:50:11] <lynxlynxlynx> i mixed my thoughs: if it is a plugin, it should be doable pretty quickly for them
[13:50:25] <barra_home> there is a feature request for that at sf
[13:50:33] <barra_home> they haven't implemented it yet though
[13:50:47] <lynxlynxlynx> vote for it, comment, annoy
[13:50:54] <barra_home> that's the plan :-)
[13:51:10] <lynxlynxlynx> that's how things get done everywhere :)
[13:51:20] <barra_home> wow, they even offer hg support at sf nowadays
[13:51:54] <barra_home> btw.: what was the main reason for chosing git over other dvcss?
[13:52:05] <lynxlynxlynx> it's the best? :P
[13:52:10] <lynxlynxlynx> we all know it too
[13:52:18] <fuzzie> and we were all already using it, with git-svn
[13:52:25] <lynxlynxlynx> avenger was the only one that needed a migration
[13:53:34] <barra_home> alrighty, fair enough, I've been reading that one lately: http://www.python.org/dev/peps/pep-0374/
[13:53:40] <barra_home> I found it pretty interesting
[13:53:57] <barra_home> but I can see that good points can be made for each different dvcs
[13:54:45] <fuzzie> media (sound, textures, etc) is a completely different story, also
[13:55:00] <barra_home> do you store it outside the git repo?
[13:55:21] <lynxlynxlynx> it's also still evolving fast, so i wouldn't trust a two year analysis too much
[13:55:36] <lynxlynxlynx> we don't, but we also don't have much binary data
[13:55:40] <barra_home> to be fair lynxlynxlynx: the docs has been consistenly updated
[13:55:49] <barra_home> * doc
[13:57:23] <fuzzie> but both hg and git seem good
[13:58:31] <lynxlynxlynx> Git lacks a value for the 700 commits scenario as it does not seem to allow checking out a repository at a specific revision. <-- cough
[14:00:20] <lynxlynxlynx> only one of the end choice criteria was technical, so that part is not that useful
[14:00:32] <barra_home> I'm sure there are more failures in there :-)
[14:00:54] <fuzzie> i couldn't find a way to actually check out a whole repos at a specific revision (as opposed to something like recent history with --depth), at a glance
[14:01:03] <barra_home> but python is no small hobby project, so I trust them that they actually invested some time into proper research
[14:01:37] <fuzzie> the sensible thing to use for a big game project is presumably Perforce, but you will certainly get no free hosting for that :)
[14:02:01] <barra_home> yep, heard about that as well fuzzie, they usually recommend it at gamedev.net
[14:03:11] <lynxlynxlynx> ouch
[14:03:20] <lynxlynxlynx> we switched from p4 to git
[14:03:31] <lynxlynxlynx> in 2006
[14:03:41] <lynxlynxlynx> that should illustrate things :)
[14:04:07] <lynxlynxlynx> hg, svn, svk and bzr were also considered
[14:12:18] <barra_home> p4 = perforce 4?
[14:17:56] <fuzzie> lynxlynxlynx: you really see the difference when you have 20gb of media in the repo
[14:18:09] <fuzzie> but it is a horrible system :P
[14:18:23] <tomprince> This may be my bias, but I get the impression that the default dvcs is git.
[14:21:09] <barra_home> depends on how you interprete default tomprince
[14:21:17] <barra_home> I agree that it seems to be the most common right now
[14:21:31] <barra_home> but there are some high profile projects that either use bzr or hg as well
[14:22:13] <fuzzie> my view of bzr is somewhat tainted by the fact that i tried using it 4 or so times, and every time i couldn't because there was yet another new not-backwards-compatible repos format
[14:22:14] <barra_home> SDL uses hg e.g.
[14:22:46] <barra_home> http://en.wikipedia.org/wiki/Mercurial#Projects_using_Mercurial
[14:31:38] <lynxlynxlynx> everyone makes mistakes j/k
[14:39:42] <barra_home> whatever floats their boat, I guess :-)
[14:54:27] <edheldil> fuzzie: git checkout <hash> , no?
[14:56:02] <tomprince> I think they are probably talking about cloning.
[14:58:44] <fuzzie> yes. otherwise the test in question is useless.
[15:58:00] <-- barra_home has left IRC (Quit: Verlassend)
[16:01:08] --> barra_home has joined #GemRb
[16:09:26] --> |Cable| has joined #GemRb
[16:09:39] <-- |Cable| has left IRC (Client Quit)
[16:09:59] --> |Cable| has joined #GemRb
[16:24:12] <-- Bo_Thomsen has left IRC (Quit: Leaving.)
[16:54:37] <-- lubos has left IRC (Quit: Leaving.)
[18:21:32] --> SiENcE has joined #GemRb
[18:28:41] --> barra_away has joined #GemRb
[18:32:20] <-- barra_home has left IRC (Ping timeout: 245 seconds)
[18:43:30] <-- tomprince has left IRC (Ping timeout: 265 seconds)
[18:43:30] <-- SiENcE has left IRC (Read error: Connection reset by peer)
[19:14:50] --> SiENcE has joined #GemRb
[20:04:56] --> tomprince has joined #GemRb
[21:15:14] <-- SiENcE has left IRC (Quit: cya @all)
[21:49:31] --> edheldil_ has joined #GemRb
[23:02:45] <Lightkey> edheldil_: btw, ever heard of md5deep? ^^
[23:03:55] <Lightkey> I usually use "md5deep -rel *"
[23:07:16] <-- edheldil_ has left IRC (Ping timeout: 276 seconds)
[23:08:11] <Lightkey> ......
[23:20:30] --> edheldil_ has joined #GemRb
[23:23:10] <Lightkey> edheldil: you did that on purpose
[23:33:18] <-- edheldil_ has left IRC (Ping timeout: 250 seconds)
[23:40:43] <lynxlynxlynx> uh, well tell me when you two sort it out
[23:41:12] <lynxlynxlynx> i'd like to know the demo-unique files if there are any
[23:41:52] <Lightkey> if you wait a min
[23:45:56] <lynxlynxlynx> it's up
[23:46:03] <Lightkey> hehe
[23:46:11] <Lightkey> had to move around files first
[23:46:15] <lynxlynxlynx> i'll be back in a few hours
[23:46:24] <Lightkey> and cache/ is probably not pristine
[23:46:40] <-- lynxlynxlynx has left IRC (Remote host closed the connection)
[23:47:13] <Lightkey> guess it can wait till he asks again, then