Team Chat Logs

July 4, 2010

2010 6
Mo Tu We Th Fr Sa Su
      1 2 3 4
5 6 7 8 9 10 11
12 13 14 15 16 17 18
19 20 21 22 23 24 25
26 27 28 29 30 31  
<< <   > >>

[00:20:11.219366]<evil_twin>t.e.o: Ticket #493 (Use trac with a remote Subversion repository) updated - <http://trac.edgewall.org/ticket/493#comment:167> - chris.tandiono@…
[05:16:31.568267]<evil_twin>t.e.o: Ticket #616 (escape character in the wiki) updated - <http://trac.edgewall.org/ticket/616#comment:13> - anonymous
[07:21:27.952094]<evil_twin>t.e.o: Ticket #616 (escape character in the wiki) updated - <http://trac.edgewall.org/ticket/616#comment:14> - cboos
[07:36:30.861442]<evil_twin>t.e.o: Ticket #616 (escape character in the wiki) updated - <http://trac.edgewall.org/ticket/616#comment:15> - anonymous
[07:46:32.790829]<evil_twin>t.e.o: TracSubversion edited - <http://trac.edgewall.org/wiki/TracSubversion?version=78> - cboos
[08:01:40.065352]<evil_twin>t.e.o: RelationalModel created - <http://trac.edgewall.org/wiki/RelationalModel?version=1> - tracdev@…
[08:01:45.062654]<evil_twin>t.e.o: workeffort.jpg attached to RelationalModel - <http://trac.edgewall.org/attachment/wiki/RelationalModel/workeffort.jpg> - tracdev@…
[11:47:44.209939]<evil_twin>t.e.o: TracDev/Proposals/BatchModification edited - <http://trac.edgewall.org/wiki/TracDev/Proposals/BatchModification?version=3> - meeker.brian@…
[12:19:01.841248]<prologic>Can someone help me with a bit of genshi ? http://codepad.org/cDFoOWmT <-- this isn't quite right but it's almost there
[12:19:44.503128]<prologic>I get a syntax error obviously, but what I'm trying to do is locate an <li><a> whoose href is equal to some context variable
[12:19:59.320807]<prologic>and add a css class="..." attribute to it
[12:51:19.193674]<coderanger>Time to try migrating some of my plugins to github
[12:51:25.465350]<coderanger>Keep your fingers crossed
[12:54:37.065808]<prologic>coderanger, you able to help me with a bit of genshi/xpath ?
[12:54:46.974555]<coderanger>I can try
[12:54:55.539108]*coderanger is no xpath wiz
[12:55:15.886879]<prologic>neither am i
[12:55:16.911483]<prologic>http://codepad.org/1a8Z3vrX
[12:55:30.746396]<prologic>been trying to change the class of an <li> element
[12:55:39.459618]<prologic>on here http://prologic.no-ip.org:8000/
[13:02:25.837150]<coderanger>Don't you need //body?
[13:02:58.662708]<oznt>hi guys,
[13:03:23.203498]<coderanger>prologic: I think the default is to assume /body, which won't work
[13:03:37.225857]<oznt>i'm a trac user and admin, i'm looking for what determines the file types of trac by which it determines the syntax highlighting
[13:04:01.509710]<coderanger>oznt: Assuming you are using pygments, it is done by pygments itself
[13:04:17.857293]<oznt>coderanger, is there a conf file ?
[13:04:40.605295]<coderanger>oznt: What are you trying to do?
[13:04:50.846521]<oznt>i want to tell trac to highlight files ending with .f as fortran 90 and not fortran 77
[13:05:17.519336]<oznt>my code is mixed 77/95 and i have my comments in !
[13:05:39.498880]<oznt>but trac ignores them, and does not gray them ...
[13:06:05.593077]<coderanger>hmm, first question is does pygments supports both
[13:07:57.042175]<oznt>coderanger, let me check my svn i have some files ending with f90
[13:08:00.105475]<coderanger>oznt: Fortran does not have a F77 lexer
[13:08:20.217221]<oznt>?
[13:08:47.522977]<coderanger>FortranLexer
[13:08:47.748433]<coderanger>Lexer for FORTRAN 90 code.
[13:09:44.868836]<oznt>so it can't or can do fortran 90 ?
[13:09:51.693232]<oznt>i don't understand
[13:12:14.999297]<oznt>well, files in my repo ending with f90 , are simply black...
[13:12:21.136617]<oznt>so i guess it can't :-(
[13:12:36.082677]<evil_twin>t.e.o: TracDev/Proposals/BatchModification edited - <http://trac.edgewall.org/wiki/TracDev/Proposals/BatchModification?version=4> - cboos
[13:14:23.220770]<coderanger>oznt: Do you have pygments installed?
[13:14:53.479524]<oznt>coderanger, i have trac 0.11 where do i find out if it is installed ?
[13:15:06.534435]<coderanger>oznt: On the About Trac page, what do you see? (pastebin)
[13:16:21.760291]<oznt>coderanger, http://paste.debian.net/79635/
[13:16:30.887078]<oznt>i don't see it there
[13:16:54.930560]<coderanger>yeah, that would do it
[13:17:04.008768]<coderanger>just install Pygments
[13:17:08.223730]<prologic>coderanger, http://codepad.org/sKwH9ljd <-- with the help of a FF Xpath extension I'm fairly sure the path is correct (at least for the test) - however the py:match replacement isn't working
[13:18:03.855889]<nielx>hola, how do I disable the translated interface of Trac without ripping out python-babel completely? We moved to Trac 0.12 this morning and we already have three foreign language bug reports
[13:18:57.353567]<oznt>coderanger, but it does not support fortran 90 ? so why even bother ? i mean files ending with f, are highlighted
[13:19:07.422398]<oznt>only not so correctly :-)
[13:19:25.261134]<coderanger>oznt: It only supports F90
[13:19:38.336341]<oznt>ok ...
[13:19:41.042293]<coderanger>oznt: And you are probably using silvercity right now for coloring which is not as flexible
[13:19:47.184525]<coderanger>oznt: Pygments is a much better library
[13:20:02.965154]<oznt>how about this ? http://trac-hacks.org/wiki/TracPygmentsPlugin they say i don't need to install it
[13:20:06.732539]<oznt>just to activate ?
[13:20:41.724749]<coderanger>oznt: Correct, you do not need that plugin for 0.11 (only 0.10)
[13:20:55.367352]<coderanger>oznt: If you install Pygments and restart your web server, Trac will pick it up
[13:21:21.206699]<oznt>so i don't need to install it... but it is working only partially or what ?
[13:22:21.668712]<coderanger>prologic: That would select the <a>, not the <li>
[13:22:31.313413]<coderanger>oznt: Just install Pygments and look again, trust me on this
[13:22:59.611942]<oznt>so i need to install pygments itself or the plugin ?
[13:23:04.839437]<coderanger>Just pygments
[13:23:06.344983]<coderanger>No plugins
[13:23:40.739907]<prologic>coderanger, yes but I want to select the <li> whoose <a> has a matching href attribute
[13:23:45.052740]<prologic>how the hell do you do that :)
[13:23:59.170290]<coderanger>prologic: You need an ancestor selector, which genshi might not support
[13:24:11.263510]<prologic>oh hmm
[13:24:13.155814]<coderanger>Yeah, it doesn't
[13:24:32.723057]<coderanger>Since it requires buffering to seek backwards in the stream
[13:24:34.139629]<oznt>coderanger, thanks, i will look into it
[13:24:49.474836]*coderanger is kind of sad at home much he knows about xpath
[13:59:01.391343]<prologic>coderanger, I ended up using CSS selectors with a bit of css in the genshi template --> http://prologic.no-ip.org:8000/
[13:59:06.156447]<prologic>at latest it works
[14:29:21.035581]<coderanger>Gah, and people wonder why I never look at patches
[14:29:30.223698]<coderanger>it is because they are utter crap most of the time
[14:31:13.755778]<coderanger>someone clearly has no idea what a set() is and yet is trying to correct my usage of them
[14:44:49.819125]<Alver>Heh
[14:45:17.367164]<Alver>coderanger: developers are doomed to suffer from that
[14:45:29.538710]<coderanger>I finally figured out what they were trying to fix
[14:45:46.590610]<Alver>Something that wasn't broken, probably?
[14:45:58.845939]<coderanger>No, its a reasonable fix, but they did it oddly
[14:46:11.906081]<Alver>Oh. Well, better than the former. :)
[14:46:16.448293]<coderanger>for Postgres I need to cast to an int
[14:46:21.613356]<coderanger>instead of a string
[14:46:31.389019]<coderanger>silly typesafe DBs
[14:46:54.258616]<Alver>I did write some kind of extension to the authz stuff, but I think my style of coding is beyond what you'd call utter crap
[14:47:54.163141]<Alver>(though the trac permissions system doesn't really allow a clean way of doing things, I'm afraid)
[14:48:04.682356]<coderanger>yeah, it didn't help that the patch is some crazy format the both patch and git don't speak
[14:48:10.060362]<coderanger>and had literal tabs in it
[14:48:13.964102]<Alver>heh
[14:51:52.597192]<Alver>coderanger: has there ever been given thought to the idea of rewriting the trac auth stuff?
[14:52:05.917425]<coderanger>Alver: Define "rewriting"
[14:52:38.425525]<Alver>coderanger: kicking out the whole "permissions are users are groups" soup that seems to be the current situation
[14:52:55.526891]<Alver>or I could be totally mis-interpreting it, but I don't think I am
[14:53:14.301459]<coderanger>The groups are perms stuff isn't likely to change
[14:53:20.855228]<coderanger>since it is simple and not generally a problem
[14:53:21.317269]<Alver>Mmm. Bugger.
[14:53:33.093606]<coderanger>the lack of a directory system is more likely to change
[14:53:41.190023]<coderanger>What problems have you had with that?
[14:53:42.577912]<Alver>No, but it makes it rather nasty to get a clean view on what permissions someone has
[14:53:59.075838]<Alver>You have to recursively loop over everything you have until nothing new is discovered
[14:54:08.755623]<coderanger>Yes, how else would you do groups?
[14:54:25.481630]<coderanger>Unless you require that groups can't contain other groups, it must be recursive
[14:54:51.222361]<coderanger>so splitting groups and perms just means you remove the last iteration
[14:55:07.825421]<Alver>it means you remove an iteration per permission you have
[14:55:28.818851]<Alver>because now my code has to check if WIKI_VIEW is a group or not
[14:55:46.246215]<coderanger>No, anything that is all caps is defined as a permision
[14:56:03.074927]<Alver>Mmmmh. Right, I can restrict it a bit then
[14:56:36.103357]<coderanger>The unrolling does one iteration per level of group you have, more or less
[14:57:07.558505]<Alver>Is there a prefab function to get the cleaned up list of permissions of a user, btw?
[14:57:21.545163]<Alver>There almost has to be :)
[14:57:37.486567]*Alver probably just reinvented the wheel... badly
[14:59:05.099634]<coderanger>yeah, there is stuff in the perms module
[14:59:22.214889]<Alver>I'll look into it then
[14:59:54.540446]<coderanger>I would show you but t.e.o is dead
[14:59:56.945823]<Alver>I just edited an existing patch for authz to make ACLs work with the builtin groups, instead of yet another kind of permission system
[15:00:12.069264]<Alver>Hehe I can manage. Knowing it exists is already a relief
[15:00:33.632035]<coderanger>http://github.com/edgewall/trac/blob/trunk/trac/perm.py#L351
[15:02:38.198478]<Alver>Mmm... that would clean things up quite a bit
[15:03:03.284878]<Alver>Thanks
[15:04:02.381806]<Alver>Oh, by the way - that CacheSystemPlugin seems to be yours; is it still relevant to use it?
[15:04:27.868135]<Alver>Looking around a bit for ways to keep trac performing well, and this plugin caught my attention
[15:04:48.108702]<coderanger>Alver: I don't even know what that does anymore
[15:05:39.007262]<coderanger>Alver: There is a 0% chance that still works
[15:05:49.414607]<Alver>Hah :)
[15:05:56.995892]<Alver>http://trac-hacks.org/wiki/CacheSystemPlugin <<
[15:06:12.144880]<Alver>2003... blast from the past
[15:06:18.974177]<coderanger>Very 0.10 specific
[15:06:20.649471]<Alver>Or 2008
[15:06:30.699866]<Alver>I don't understand USA date formats... heh
[15:06:41.675637]<Alver>Okay, so I shouldn't bother trying it on 0.12?
[15:06:46.322699]<coderanger>It wouldn't work
[15:07:08.280510]<coderanger>Also, for 0.11 and up the slow part isn't the wiki rendering, it is usually running the template itself
[15:07:15.716054]<coderanger>(and Genshi has few provisions for caching)
[15:08:04.516042]<Alver>I must say that once you understand how it works (and it took me a while), it turns out rather easy to template trac
[19:31:53.799982]<exarkun>Changes made by my trac post-commit hook are randomly dropped
[19:31:56.602128]<exarkun>What's up with that?