Team Chat Logs

March 12, 2010

2010 2
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        
<< <   > >>

[00:03:08.778089]<dfas>is 0.12 getting close to being production ready?
[00:04:20.801095]<otaku42>moin
[01:48:35.799997]<scfe>dfas: Depends on what you require for 'production ready'. trac.edgewall.org runs 0.12dev
[01:49:17.306481]<gozerbot>trac: Ticket #4202 ([PATCH] Images are not found on WikiStart page when the URL contains an ...) closed - <http://trac.edgewall.org/ticket/4202#comment:5> || Ticket #8445 (authz_policy receives 'None' as resource) closed - <http://trac.edgewall.org/ticket/8445#comment:15> || Changeset [9345]: authz_policy: Fixed handling of empty resources (`resource`, realm or id ...
[01:49:22.205038]<gozerbot> - <http://trac.edgewall.org/changeset/9345>
[02:23:57.900627]<bootlaces>Hi Guys, questoin, is there a trac hook plugin that would post to an irc channel whenever a ticket gets modified?
[02:33:10.894490]<paissad>guys, i have a matter building locales from svn
[02:33:14.500244]<paissad>i did python python setup.py compile_catalog -f 2>&1 | wgetpaste
[02:33:30.141224]<paissad>http://pastebin.ca/1835047
[02:33:54.340774]<otaku42>bootlaces: see http://trac-hacks.org/ticket/1952 and/or http://trac-hacks.org/wiki/AnnouncerPlugin
[02:34:20.004806]<otaku42>paissad: "python python"?
[02:34:32.833500]<paissad>otaku42, only once, sorry
[02:34:46.521336]<paissad>i wrote python only once
[02:38:00.016175]<paissad>"python setup.py install --root=foo" works normally !
[02:38:28.745861]<bootlaces>otaku42: grazie :)
[02:40:34.246064]<paissad>python setup.py --help" does not show me compile_catalog -f target ...
[02:40:48.454136]<paissad>does it mean it does not exist ?
[02:42:38.438107]<paissad>nobody knows ?
[02:43:55.825423]<otaku42>paissad: i know that i also had problems with compiling the i18n catalogs for a test setup. it didn't bug me enough to investigate it further.
[02:44:19.705771]<otaku42>paissad: don't remember if the error message i got back then was the same you get now
[02:44:45.082574]<paissad>you still haven't the i18n ?
[02:44:50.760394]<otaku42>paissad: no.
[02:44:54.699771]<paissad>:/
[02:45:30.816600]<otaku42>paissad: after all it's just a test environment, for other aspects of trac. but sooner or later i will look at that problem, too (given it still exists then)
[02:45:55.668592]<paissad>ok
[02:45:57.181153]<paissad>sure
[02:47:23.625096]<otaku42>paissad: check the tickets on t.e.o, maybe it's a known issue. if it is not known, you could file a new ticket for it.
[02:47:49.422731]<paissad>Ok
[02:49:17.980162]<gozerbot>trac: Changeset [9346]: notification: Strip leading and trailing spaces from e-mail addresses. * ... - <http://trac.edgewall.org/changeset/9346>
[03:06:53.496784]<paissad>otaku42, on debian, 3 weeks ago i built from svn , revision 9216 .. compile_catalog -f succeeded ...
[03:07:18.472607]<paissad>otaku42, i just retreived the same revision, but the build fails again !
[03:07:22.121588]<paissad>o_O
[03:07:38.504654]<paissad>but it's slackware this time, not debian
[03:11:27.317099]<paissad>otaku42, what's you OS/distro ?
[03:12:21.408592]<otaku42>paissad: debian stable
[03:12:27.309366]<paissad>ok
[03:12:47.040922]<paissad>otaku42, revision 9216 should work for you i think
[03:12:56.595584]<paissad>or a bit higher too ^^
[03:13:32.833920]<otaku42>paissad: ah, ok, will give that a try. the test environment is on r9182 right now.
[03:13:41.404226]<otaku42>paissad: thanks for the hint
[03:14:23.457330]<paissad>otaku42, if you're interested ... http://deb.paissad.net/pool/main/trac/
[03:15:07.474669]<paissad>i do not use debian for my server now, i changed ... but my package worked for me .. you can either download the source & build yourself
[03:15:09.721923]<otaku42>paissad: thanks, but i'm using virtualenv and install trac and plugins inside these from subversion and/or via easy_install/pip
[03:15:23.894978]<paissad>ah ok, great
[03:19:20.772265]<gozerbot>trac: Ticket #9120 (Documentation link to trac-post-commit-hook broken) closed - <http://trac.edgewall.org/ticket/9120#comment:3> || 0.12/TracInstall edited - <http://trac.edgewall.org/wiki/0.12/TracInstall?version=15> || Ticket #8699 (Space in the end of e-mail address prevents sending notification) closed - <http://trac.edgewall.org/ticket/8699#comment:8>
[03:24:13.983962]<paissad>i think there's requirement i did not satisfied ... i did read the README file, ... but i'm confused about which json to download & install
[03:24:22.091070]<paissad>may someone tell me http://www.json.org/
[03:24:45.881624]<paissad>i see pyson, simplejson ...
[03:44:02.945166]<paissad>is it possible to install dependancies directly from easy_install ? (genshi, setuptools, pysqlite ...)
[03:44:04.234591]<paissad>??
[03:46:33.801542]<mapelkon>does trac email notification system have method to let end user to decide (case by case), whether or not ticket change should trigger email notification?
[03:46:49.532993]<paissad>mapelkon, it's ok now ^^
[03:47:22.600178]<mapelkon>paissad, yep thanks
[03:48:56.672295]<ShortWave>Hi all
[04:10:58.383729]<nascentmind>hi. How do i disable preview of attachments?
[04:11:41.732921]<otaku42>mapelkon: not trac itself, but afaik this can be achieved with AnnouncerPlugin (see trac-hacks.org)
[04:12:20.536192]<otaku42>paissad: this should happen automatically if you install trac via easy_install
[04:13:13.626303]<paissad>otaku42, but if i install trac from easy_install i won't have the latest svn revision ... or maybe i misunderstood
[04:13:40.133889]<otaku42>nascentmind: [attachment] max_size=1 might do the trick; see http://trac.edgewall.org/wiki/TracIni#attachment-section
[04:14:18.066516]<paissad>otaku42, if i install trac from svn, is there a way after to upgrade to HEAD revision without problems ?
[04:14:35.149802]<paissad>btw, easy_install clearsilver failed for me
[04:17:28.193850]<ShortWave>Ok so
[04:17:42.232105]<ShortWave>The docs are a tad thin on how you're supposed to deploy multiple projects
[04:17:58.281422]<ShortWave>Am I supposed to edit the trac.cgi file?
[04:18:25.194269]<ShortWave>Pretty sure I have the VHost setup correctly, where I think I'm getting lost is in the deploy business.
[04:19:05.635103]<otaku42>paissad: you can install trac from trunk with: easy_install http://svn.edgewall.org/repos/trac/trunk
[04:19:18.488077]<ShortWave>I have multiple svn repos and I'd like trac to use them all.
[04:19:29.088656]<otaku42>paissad: clearsilver should not be required for recent trac versions, unless you have plugins that require it.
[04:19:39.037422]<paissad>ok
[04:27:13.366849]<nascentmind>otaku42, should I just comment out render_unsafe_content?
[04:27:52.953730]<otaku42>nascentmind: commenting that out means that the default is used. you should set it to true or false, depending on what you prefer.
[04:28:23.808997]<otaku42>nascentmind: but setting it to false does not mean that attachments won't be displayed at all
[04:31:42.616800]<ShortWave>Anyone?
[04:35:11.961121]<otaku42>ShortWave: no experience on multi-env setups on my side, sorry.
[04:41:02.953539]<nascentmind>otaku42, setting max_size = 1 does not allow me to attach any file.
[04:41:15.237170]<nascentmind>otaku42, i just want the preview to be disabled.
[04:49:18.928074]<gozerbot>trac: Changeset [9347]: admin: Fixed the plugin admin panel, which would enable all components in ... - <http://trac.edgewall.org/changeset/9347>
[04:51:40.218537]<otaku42>nascentmind: hmm, indeed... now that you mention it: that setting is for the acceptable size of attachments that may be made and does not apply to which attachments get rendered or not. sorry for the confusion.
[04:52:14.183766]<otaku42>nascentmind: as it seems there is no option to disable attachment rendering. i confused that with rendering of diffs.
[04:53:50.891861]<otaku42>hey p0g0 :)
[04:54:21.875871]<p0g0>hey otaku42: good to see your typing!
[04:56:54.292098]<nascentmind>otaku42, hmm. thats bad.
[04:57:31.049889]<otaku42>p0g0: dito. how comes you're hitting #trac?
[04:58:38.156615]<ShortWave>Hrm...
[04:58:52.200962]<ShortWave>Dammit
[04:58:54.960758]<ShortWave>I'm completely lost here
[04:59:11.953253]<ShortWave>One project works fine
[04:59:19.448703]<ShortWave>others throw a weird error
[04:59:30.084816]<p0g0>otaku42: just a dillettante...
[05:06:19.087956]<otaku42>p0g0: ah, ok :)
[07:21:02.969879]<j-b>hello
[07:21:37.576743]<j-b>I have a "ProgrammingError: relation "spamfilter_bayes" does not exist" on my trac
[07:22:04.349809]<j-b>any idea what is wrong?
[07:22:25.398487]<Spec>did you install a plugin recently?
[07:22:39.837577]<j-b>nope, but this is indeed the spamfilter plugin that doesn't work
[07:22:59.021134]<j-b>and it happens when I go to /admin/spamfilter/bayes
[07:23:33.010005]<kirean>j-b: your database is missing a table?
[07:24:39.271182]<j-b>That is what I guessed... How can I fix the db
[07:24:40.490305]<j-b>?
[07:27:59.561275]<kirean>j-b: I really can't tell, but maybe a stunt like: http://trac.edgewall.org/ticket/4294#comment:6 can be worth trying?
[07:28:07.007184]<kirean>j-b: what about db backups?
[07:28:20.492219]<j-b>kirean: it never worked :)
[07:28:34.062387]<j-b>my idea is that the table were never created
[07:28:37.160150]<kirean>j-b: you need the plugin?
[07:28:42.196379]<j-b>http://trac.edgewall.org/browser/plugins/0.11/spam-filter/tracspamfilter/upgrades.py
[07:28:52.818974]<j-b>kirean: it would be cool to have it
[07:29:00.744292]<Spec>doesn't the plugin automagically upgrade the trac db when you call upgrade with trac-admin?
[07:29:31.758679]<Spec>oh, that ticket says as much
[07:30:09.288618]<j-b>Database is up to date, no upgrade necessary.
[07:32:00.666227]<paissad>guys, i have this error when i try to enter the site
[07:32:03.494209]<paissad>https://trac.paissad.net/cci/Java
[07:32:28.922548]<paissad>i installed trac from easy_install
[07:33:21.550030]<paissad>& this time, i don't have the locale (fr) ... knowing that my trac environnements existed before ( in another system -> debian )
[07:33:29.046901]<paissad>& now, i'm using slackware
[07:34:19.833063]<paissad>i tried to build locale like this ( python setup.py compile_catalog -f ) but it always fail :/
[07:38:49.107209]*retracile spouts something untoward about mornings.
[07:44:20.090837]<paissad>when i run "python setup.py compile_catalog -f", i have this error
[07:44:20.974104]<paissad>http://pastebin.ca/1835560
[07:44:57.577840]<paissad>i'm just trying to install the locales, .... i already installed trac via easy_install
[07:47:39.937295]<sobersabre>hi. I have a trac environment in /home/trac/project
[07:47:56.461818]<sobersabre>and I have apache mapped the root of the project to hostname.com/project
[07:48:04.917839]<sobersabre>it works, except the icon and the header logo.
[07:48:21.354340]<sobersabre>I have put the images to /home/trac/project folder.
[07:48:33.371647]<sobersabre>how am I referring to it in trac.ini ?
[07:48:41.677950]<sobersabre>is it /project ?
[07:48:45.626443]<sobersabre>or is it something else ?
[07:49:42.856880]<osimons>paissad: think you may be missing the CLDR data
[07:49:54.660315]<paissad>osimons, o_O
[07:49:59.768056]<paissad>don't know what you mean
[07:55:21.179522]<osimons>paissad: http://babel.edgewall.org/wiki/SubversionCheckout
[07:55:32.602977]<Linuturk>wow, trac is hard to install
[07:56:42.661137]<sobersabre>guys, my trac related apache config: http://pastebin.ca/1835580
[07:56:50.396367]<sobersabre>question is inside the paste comment.
[07:59:05.174573]<osimons>paissad: releases of babel comes with the support data (like if you do "easy_install -Z Babel"), but running from svn you need to fetch it yourself
[07:59:29.901267]<paissad>ok
[08:07:34.004891]<paissad>osimons, you're a great great man ^^ .... thank you, now everything's ok , compile_catalog -f runs well
[08:07:58.746089]<osimons>paissad: your welcome :-)
[08:13:32.693829]<paissad>otaku42, follow theses instructions from the link & do your compile_catalog -f
[08:13:34.265110]<paissad>http://babel.edgewall.org/wiki/SubversionCheckout
[08:13:47.769842]<paissad>it worked for me ... hope that will for you too
[08:19:20.499408]<gozerbot>trac: Ticket #9135 (Duplicated method in pygements tests) created - <http://trac.edgewall.org/ticket/9135> || Ticket #9134 (Allow links to open in a new tab or window frame) created - <http://trac.edgewall.org/ticket/9134>
[08:22:40.168333]<retracile>Linuturk: care to phrase that as a question?
[08:24:55.433759]<Linuturk>retracile: Why is trac so hard to install on CentOS?
[08:29:43.247275]<retracile>Linuturk: What problems are you running into?
[08:35:51.195182]<Linuturk>I guess the docs are just out of date
[08:38:58.571789]<retracile>Yeah, I hear "trac is hard to install" quite a bit, but I can't get any _specific_ information so we can fix whatever the problem is. The longer this persists, the more I suspect the issue is stupid users. *grumble* Give us _useful_ feedback so we can fix the problem people!
[08:56:03.004570]<paissad>guys, i built trac with compile_catalog -f, but i still have the site in english, .. not in french
[08:56:29.300208]<paissad>i also did "trac-admin $TRAC_REPO upgrade"
[08:57:08.759413]<paissad>when i had the same repository in debian, i had the page directly in french
[08:57:09.574132]<paissad>https://trac.paissad.net/cci/m1-stage/
[08:57:28.124722]<paissad>is there a configuration i missed ?
[09:19:21.817591]<gozerbot>trac: TracInstall edited - <http://trac.edgewall.org/wiki/TracInstall?version=295>
[09:37:47.856405]<talley>hi
[10:00:58.362909]<cmc>hi
[10:01:03.544940]<cmc>:)
[10:25:22.613783]<talley>so why won't be th:AccountManager integrated in to trac .12? :)
[10:29:19.923712]<retracile>talley: planned for 0.13
[10:29:43.530776]<retracile>talley: There were enough things going into 0.12 already.
[10:29:53.800061]<talley>I see, thank you for the heads-up
[10:31:22.505258]<retracile>talley: be sure to harrass pacopablo about it as soon as 0.13dev opens up. ;)
[10:31:54.828332]<talley>allright :)
[10:49:22.382932]<gozerbot>trac: SeaChange/WhatUsersWant edited - <http://trac.edgewall.org/wiki/SeaChange/WhatUsersWant?version=114>
[11:49:24.335888]<gozerbot>trac: Ticket #9136 (MultiRepos Top level browser view + View Changes causes exception) created - <http://trac.edgewall.org/ticket/9136>
[12:18:16.527364]<rts>hi
[12:23:25.443812]<rts>uh oh cO - nothing much going on here, how about an initiative for recruiting more developers for trac?
[12:24:27.177971]<rts>and also making trac more open to the public, by installing a repository that would allow patch based commits like for example hg and the likes?
[13:02:48.548098]<cmc>I'd like to upload documents in my plugin. Anyone have a *starting place*?
[13:03:39.871552]<rts>why not look at the current implementation of the attachment system and templates thereof, they provide for an upload facility
[13:04:24.857330]<rts>see also ticket/web_ui.py or wiki/web_ui.py for uploading attachments
[13:04:37.308243]<cmc>I will look at core if there aren't any available hooks. core code tends to be more intertwined than plugins
[13:04:48.209955]<cmc>thanks for the suggestion, though
[13:05:09.338162]<rts>afaik in core you will find no such notion of fileupload, good luck
[13:11:10.204786]<rts>i am currently looking into issue 5648 and i wonder how we could refactor trac/perm.py in order to accomplish what was stated in the comments by cboos
[13:11:34.674580]<rts>any hints on that?
[13:14:22.629399]<rts>afaik the behaviour of the DefaultPermissionGroupProvider should be directly integrated into the default permission store so that it will not require the ExtensionPoint(IPermissionGroupProvider)
[13:15:15.211232]<osimons>rts: i doubt the project will move from svn... however, i keep hg version quite up to date if you want to use that as a starting point for patches, mq patch series or something
[13:15:30.386027]<osimons>rts: http://bitbucket.org/osimons/trac-011/ and http://bitbucket.org/osimons/trac-trunk/
[13:16:14.938219]<rts>osimons: hi thanks for the hint and reply - i already wondered why there was no feedback at all :-/
[13:16:22.365701]<rblank>Same here, different conversion program: http://bitbucket.org/rblank/trac-trunk/
[13:16:50.929983]<rts>osimons: well, currently i am patching against trac-0.12dev
[13:17:26.527387]<osimons>ah, rblank. double effort. oh well.
[13:18:32.528741]<rblank>osimons: Not at all. It's a clone of my SVN working copy.
[13:19:54.239529]<rts>perhaps we should install this as a common basis and also provide a public commit repository from where individual patches then be incorporated into the main line? see for example linux/git repository and overall process?
[13:20:05.121444]<osimons>ok, rblank. i just use mine "pure" as basis for MQ patches - no svn commands run inside it
[13:20:42.330196]<rblank>osimons: I pull SVN into a branch, and make feature branches for my work. No mq at all (too complicated).
[13:21:29.410423]<rts>osimons: haven't we talked the other day about moving trac repository to a patch based repo?
[13:21:38.556297]<rblank>rts: It's great to see you taking an interest in working on Trac.
[13:21:56.100248]<rblank>rts: I notice you're especially interested in architecture and refactoring, correct?
[13:22:07.540777]<rts>rblank: absolutely, if find trac to be the perfect solution once it becomes more committed to, so to say
[13:22:26.978751]<osimons>rts: nope, not that i can remember... i don't mind a patch-based repos as long as it is hg :-)
[13:22:36.919464]<rts>rblank: i am, but i am also a newby on python, so i am just learning an getting ground
[13:22:38.968951]<rblank>osimons: Same here :)
[13:23:04.426250]<rts>rblank: -an +on
[13:24:13.610895]<rblank>rts: We're currently wrapping up to release a first beta of 0.12.
[13:24:45.111358]<rts>rblank: i wonder how we can make trac more web3.0 like, taking aside all of the hype of web2.0, we should make trac more scalable in all terms
[13:25:06.917296]<rblank>rts: So in the current state, architecture and refactoring are a bit left on the side, and we try to concentrate on fixing a few last issues.
[13:25:17.075976]<osimons>rts: we may well skip 3.0 and go straight for 4.0...
[13:25:22.846174]<rts>;)
[13:26:10.004106]<rblank>rts: So don't be surprised if we don't react much to tickets that require big architectural changes.
[13:26:26.168262]<rts>rblank: ok
[13:26:32.853727]<rts>rblank: well, considering architecture and all, the current architecture is quite good, afaik, but the resource system needs a big overhaul
[13:27:16.135861]<rblank>rts: We'll get to that, but after 0.12. Currently, we need to focus our efforts on 0.12.
[13:28:04.016600]<rts>rblank: is there an actual roadmap where one could see the changes to go into 0.12? i mean apart from the issues?
[13:28:33.392514]<rblank>rts: Don't think we don't value your input, on the contrary. It's just that we're focused on the immediate issues.
[13:28:49.157632]<rblank>rts: For 0.12, it's the remaining open tickets assigned to that milestone.
[13:29:56.611341]<rts>rblank: well, i started on the issues but i am unsure about which tickets to fix first, apart from that most of the tickets seem to be quite far fetched or too complicated for me to solve as they would require refactorings that would eventually break backwards compatibility or something like that
[13:30:23.383534]<rblank>rts: Yeah, that's always an issue when starting on a new project.
[13:30:59.659632]<rts>rblank: is communication only via the ticket system or also via this channel? if so, how about cboos entering the channel so that we can get straight to business?
[13:31:02.846001]<rblank>rts: I have started tagging "simple" issues with the "bitesized" keyword. See in the middle of http://trac.edgewall.org/wiki/HowToContribute
[13:32:01.692882]<rblank>rts: Depends on the subject. Most communication is either on tickets or on trac-dev (the mailing list). I'm on IRC when I code, cboos only rarely.
[13:32:04.627871]<rts>rblank: ok, i will filter for 0.12 and bitesized then
[13:32:31.727733]<rts>rblank: are you french? from the name i would presume so.
[13:32:46.533390]<rblank>rts: Oh, there are only a few ATM, and none on 0.12. But 0.12.1 and next-minor-0.12.x are good candidates as well.
[13:32:58.222624]<rblank>rts: Nope, swiss actually. But I do speak french
[13:33:41.686022]<rts>rblank: swiss, nice ;-)
[13:34:15.022810]<rts>rblank: although i have a different opinion of swiss policy, especially regarding 1933 - 1945 ...
[13:35:04.351118]<rts>rblank: but that is politics, and we do not have politics in trac development or do we?
[13:35:19.909551]<rblank>rts: I would hope not.
[13:36:00.502495]<rts>rblank: well then, i will further look into the existing issues targetted for 0.12.1 and next minor and see what i can fix
[13:36:08.710872]<rts>rblank: thanks for the feedback, dude
[13:37:20.074448]<rts>rblank: i wondered, though, most plugins establish their own tables for storing page data, what about a general plugin that would allow reusing the wiki table for storing versioned data?
[14:05:50.687661]<rts>rblank: never mind ;) please excuse my reference to the dark ages of time...
[14:12:31.514604]<cmc>for logging purposes, tracfullblog/core.py has a working implementation of ILegacyAttachmentPolicyDelegate and IResourceManager, both of which I've needed
[14:12:40.312480]<cmc>(to add attachments)
[14:19:23.605116]<gozerbot>trac: Changeset [9348]: db: A few `with_transaction`-related fixes. - <http://trac.edgewall.org/changeset/9348>
[14:27:07.432614]<rts>^^nice
[15:31:12.481867]<osimons>cmc: missed your attachment question earlier - check my fullblog plugin. it does the necessary resource stuff and attachment "things", using the Trac support
[15:31:43.124566]<osimons>cmc: ask if you have questions, and i'll see what i can remember...
[15:32:38.185913]<osimons>though i did much the same for Bitten recently too. no doubt i can find the changeset that implements the essence of the backend support.
[15:33:28.387531]<cmc>I actually suggested your plugin as a good example :)
[15:33:47.786854]<cmc>it's a good resource and got me to where I needed to be, thanks!
[15:33:57.416705]<osimons>cmc: yeah, saw that now. check this out too: http://bitten.edgewall.org/changeset/703/trunk
[15:34:27.100279]<cmc>i assumed I should stay away from ILegacyAttachmentPolicyDelegate
[15:35:18.038079]<osimons>50% is related to the build tool for uploads through custom protocol as part of builds, but i suppose you can easily pick out the parts that look relevant (and ignore the rest)
[15:36:42.906509]<osimons>cmc: well... if your attachments are to be resources in their own right, then you should have your own permissions and ignore the legacy delegate. for bitten and for fullblog, i want the main resouce (like blog post) to be the one controlling permissions and attachments just follow along
[15:37:04.477986]<osimons>cmc: same still used by wiki and ticket handling of attachments in trac (i think...)
[15:38:26.973548]<cmc>yeah, I'm following that strategy; wiki and tickets have a special dictionary in the core attachment.py which map permissions
[15:39:31.008196]<osimons>cmc: right - and if your perms are really plain, then things are easy. my blog permissions don't quite map to the trac standard (where you can modify_own and stuff), so i do that myself
[15:41:10.923881]<osimons>cmc: what is your plugin for? what resources?
[15:42:01.774014]<cmc>I've been extending the clients plugin pretty heavily with little sub plugins. in this case, needed to allow our admins to attach documents to clients
[15:42:47.329957]<cmc>aim is to have a mini sugarcrm by the end of the month...
[15:44:33.911850]<osimons>clients plugin. don't think i've heard of that. so... much... code....
[15:44:51.718815]<rts>i could not find it either...
[15:45:15.017100]<rts>perhaps some internal plugin of yours?
[15:45:27.163338]<cmc>http://trac-hacks.org/wiki/ClientsPlugin
[15:45:43.281464]<cmc>coling's work
[15:46:32.580558]<rts>nice, somehow google manages to not list that as the first 10 search results, wonder why
[15:46:40.722609]<cmc>haha no idea
[15:52:15.728975]<rts>so you are currently developing sort of a crm based on the ClientsPlugin, will you publish your developments or are these internal to your company?
[15:57:25.689717]<cmc>originally, I modified the core clients plugin a lot. There's no way I'd be able to get coling to accept all of that, though, so I've started breaking it out into separate plugins
[15:58:35.241790]<cmc>the hard part is that there needs to be *some* changes to the core plugin (such as using a client id instead of a name), so atm, I'm thinking of making an "extendedclients" plugin and slowly merging all of it's features in to clients core, assuming coling will accept
[15:58:42.927980]<rts>^^the same with me, i started figuring out that trac needed a lot more patches ;) until remy blank pointed me to the correct direction of making all of these plugins
[15:58:48.614086]<cmc>but yeah, it'll all go to trac-hacks
[15:59:09.417346]<rts>^^great
[15:59:49.113870]<cmc>yeah the plugin system is pretty nice
[16:00:05.342773]<rts>but why use a client id instead of a name, is the client id numeric? why not use unique names for the clients such as foobar.com
[16:01:06.045268]<cmc>id is numeric, yes. also, it makes it much easier to account for changing client names and *long* client names
[16:01:23.107722]<cmc>plus there's the whole indexing and joining thing
[16:01:23.606268]<rts>besides that, you could always install a table for mapping numeric ids to names in the clients table
[16:01:59.588529]<cmc>that's true, but would require that names be unique
[16:02:19.856986]<rts>domain names are unique ;-)
[16:02:37.108997]<cmc>;)
[16:02:50.210456]<cmc>clients for us are much more generic
[16:03:31.761061]<cmc>every now and then there will be a conflict; we can't have everything crap out when that happens
[16:03:59.582678]<rts>personally, i resolved to using domain names for clients, and when it comes to projects, i just subdomain them
[16:04:53.628056]<cmc>not a bad idea by any means
[16:04:59.002928]<rts>^^
[16:05:04.060171]<cmc>:)
[16:05:13.396857]<cmc>time for me to go though. good talking with you both
[16:05:34.051838]<rts>gn8 or whatever time it is you are located in ;)
[16:09:43.994662]<rts>so many peops on this channel and nobody talks, why?
[16:14:43.363418]<talley>http://www.useit.com/alertbox/participation_inequality.html
[16:15:47.876845]<rts>lol
[16:17:28.463338]<talley>well I've made my first contribution on t-h, but I'm more of a lurker
[16:18:22.492169]<rts>oh me too, especially on guild wars watching the teeny talk ;)
[16:18:35.209324]<rts>which contribution did you make?
[16:18:44.532235]<talley>only a patch so far
[16:18:55.998688]<rts>link?
[16:19:02.486042]<talley>nothing big http://trac-hacks.org/ticket/6798
[16:20:43.326669]<talley>while I was searching for the solution I did develop a liking to python and/or trac :)
[16:22:47.619972]<rts>nice, i did so, also ;)
[16:28:40.867544]<rts>well, i just flicked through the code base of the announcer plugin, but i wonder why you would want to restrict access to WIKI_VIEW only?
[16:29:19.413312]<talley>our group is using a private trac instance, with AccountManager, without anonymous access
[16:29:44.377781]<talley>Announcer is a very good plugin, but it was posssible to leak out wiki pages
[16:30:04.672268]<rts>mmmkay...
[16:30:37.663762]<rts>how about restricting access to the role authenticated-user instead of requiring WIKI_VIEW?
[16:31:08.746615]<talley>I don't follow
[16:31:31.326472]<rts>i dont actually know what the announcer plugin is all about, but i reckon that is not actually wiki related
[16:31:36.536716]<talley>problem is: if you are not logged in, you still have access to /prefs/, you can set an email adddress
[16:31:54.852551]<talley>Announcer will send email to you if a matching wiki page is modofied
[16:31:57.590861]<talley>modified*
[16:32:04.270776]<rts>oh, well, i see
[16:32:26.091765]<talley>so, you did set your email address, set a * (wildcard) watch
[16:32:36.088202]<rts>that way, the patch seems very sound to me
[16:32:45.256760]<rts>why not join trac development?
[16:32:47.696929]<talley>you can't see the wiki (because anonymous has no permissions), but you will still get wiki change announcements
[16:33:35.672318]<talley>it would feel good to help trac, I did ask coderanger_ approx. 2 years ago if I could donate
[16:33:58.051780]<rts>ah, forget coderanger, ask remy blank ;)
[16:33:58.694074]<talley>problem is I'm very busy to learn python on a level whcih is sufficent for the project
[16:34:08.578857]<rts>me too
[16:35:26.251228]<talley>I mean, too busy
[16:35:35.263387]<talley>not busy with learning it :/
[16:35:39.594820]<rts>oh well, too bad
[16:36:38.043032]<rts>but then again, the meta concept of trac.core is quite compelling don't you think?
[16:37:46.838769]<talley>not enough information to answer :)
[16:38:08.874801]<rts>%/
[16:38:26.133726]<talley>I was pleased with what I saw during my adventure to fix that leaking problem
[16:38:45.111426]<rts>;)
[16:39:38.274959]<rts>i just thought about the realm that the announcer plugin establishes. perhaps it would be best to introduce additional permissions to the system , for example ANNOUNCER_VIEW, ANNOUNCER_EDIT and so on?
[16:40:05.126590]<talley>I wanted to keep changes to a minimal
[16:40:33.031021]<talley>I had 5 different plans, for example disabling the whole /prefs/ page for anonymous
[16:41:04.879415]<talley>but then, our instance shall be opened one day and I would have to unhack it, or rehack when trac is updated
[16:41:19.117195]<rts>would be easily done by implementing a request filter (IRequestFilter) and preventing access to that page
[16:41:51.669912]<talley>I think this was the best solution: if you have WIKI_VIEW permission, you should be able to get change announcements
[16:43:07.125689]<talley>our ticket system is closed too, so for now we are using the first patch, the second patch is just an example of fine-grained access control
[16:43:20.950630]<rts>note that request filters are evaluated by trac.web.main before that request handlers get a chance of processing the request
[16:44:50.563350]<talley>yes, this might be a good solution too, but would require a new plugin if you want to persist functionality between trac updates
[16:45:00.374023]<rts>so in case of the anonymous user you would simply req.redirect to an error page of your own or simply raise an HTTP 404 error exception or something like that
[16:45:36.247459]<talley>for our need it would've been an overkill to develop a plugin, "fixing" Announcer was simpler
[16:45:58.073130]<rts>;) the plugin would consist of approx 5 lines or so
[16:46:17.389184]<rts>apart from the overhead of defining the class and the imports
[16:46:35.949487]<talley>yes, but consider this: a good-behaving plugin shall enforce security by evaulating permissions
[16:46:52.431933]<rts>for granted
[16:48:25.254668]<talley>but aside from this, I must mention
[16:48:29.005108]<rts>but then again, the problem lies within the announcer plugin - and not by merely enforcing a WIKI_VIEW permission, so the fix should be rather at the web_ui of the plugin than in the prefs system
[16:48:32.798533]<talley>doki_pen: I really like the Announcer plugin! :)
[16:48:47.049214]<rts>;)
[16:48:58.589147]<rts>is he online?
[16:49:15.102864]<talley>I have no idea
[16:51:04.858852]<talley>I don't really understand what you said
[16:51:19.850083]<talley>in this case, where exactly?
[16:51:35.305536]<talley>btw I think the fix should be in the resolvers/*
[16:51:54.406066]<rts>oh i just took a closer look at the code base, and i found no web_ui ;)
[16:52:20.355894]<talley>but it's problematic with the specified.py, which lets you specify an independent email address
[16:52:41.991336]<talley>if it's shown on prefs/announcer/, you can work-around all permission control
[16:53:25.459271]<talley>so, I thought the best way is to eliminate the /prefs/announcer page for those without WIKI_VIEW
[16:54:01.076985]<rts>and i see that is well documented, i like that
[16:55:09.794542]<talley>thanks. that's a standard for programmers.
[16:55:27.873463]<talley>well, should be :)
[16:55:37.499903]<rts>i still have to figure out how the plugin actually establishes the /prefs/announcer realm, i have a hard time finding it out ;)
[16:56:17.958646]<rts>aah i see IPreferencePanelProvider
[16:57:36.930203]<rts>tough, never went into that part of trac btw
[17:00:50.535495]<rts>nice plugin, learned much ;)
[17:02:54.713496]<talley>yes, very straight-forward
[17:03:07.573569]<talley>it only took around 20 minutes to figure it out
[17:03:11.823973]<talley>and fix it
[17:03:26.154185]<rts>yeah the code base is very well done
[17:04:25.313982]<talley>btw how does gozerbot work?
[17:04:48.383132]<talley>is it a supybot with IrcAnnouncer? don't think so
[17:05:19.394832]<rts>dunno, i do not use irc very often
[17:05:52.518324]<talley>oh it's a special case. gozerbot.org
[17:06:28.250679]<dunk_>yeah ;]
[17:06:34.409250]<dunk_>i made the thing
[17:06:44.119930]<talley>it looks handy
[17:07:02.879242]<dunk_>depends on what you want to do with it ;]
[17:07:37.513495]<talley>I'm looking for a way to announce timeline changes on irc
[17:07:49.826807]<talley>so far this is the most promising
[17:08:08.948793]<dunk_>yeah you need the rss plugin
[17:08:23.549308]<talley>hmm
[17:08:29.213748]<dunk_>well the bots plugin
[17:08:29.992676]<dunk_>;]
[17:08:39.760610]<talley>and is it able to authenticate with trac, with AccountManager?
[17:08:45.261375]<dunk_>no
[17:08:47.252141]<talley>:(
[17:08:55.864695]<talley>where can I put my feature request
[17:08:56.745721]<talley>:)
[17:09:26.798295]<dunk_>dunno i'm not a trac dev
[17:09:46.371598]<talley>anyway I'll look into it tomorrow
[17:09:59.469084]<dunk_>it has a bugtracker plugin though
[17:10:04.371315]<dunk_>!bug 500
[17:10:06.396017]<evil_twin>description="I was switching between the timeline, and browse views, and I switch between them before my browser had finished downloading the updated timeline. Essentially the broswer had terminated one http request an initiated another one. .. id=500 .. owner=jonas .. reporter=eoin dunne .. summary=trac causes subversion data base to lock... .. http://trac.edgewall.org/ticket/500
[17:10:25.400054]<dunk_>but no ticket submission through the bot
[17:10:55.447749]<talley>for my needs it is sufficent, I only want an announcer on IRC
[17:11:02.922358]<dunk_>talley if you want to try out gozerbot best is to use the mercurial repo see http://gozerbot.org/0.9.2/
[17:11:09.917686]<dunk_>ah ok that it can do
[17:11:52.946676]<talley>only thing is, our trac instance is private and closed for anonymous access (not even _VIEW permissions)
[17:12:39.654749]<dunk_>you can pass creds on in the rss url but not sure if you want that
[17:13:00.308259]<dunk_>maybe make a seperate account for the viewer or so
[17:13:04.198085]<talley>really? didn't know that
[17:13:06.171607]<rts>talley: why not implement a site.html template that would lock out users when not authenticated?
[17:13:10.178119]<talley>&user=&pass=?
[17:13:14.315796]<dunk_>yes
[17:13:34.964494]<dunk_>not sure if that works for trac though
[17:13:37.195536]<dunk_>hmm
[17:13:49.432517]<dunk_>brb
[17:20:57.303464]<talley>can't embed creds in rss url :/
[17:21:08.216753]<talley>can't figure it out how to*
[17:22:59.388406]<dunk_>http://user:pass@host or so
[17:23:12.152951]<talley>oooh. http auth
[17:23:33.742814]<dunk_>well isnt that what the rss feed uses ?
[17:23:35.322486]<talley>don't think it would work with AccountManager's form-based login
[17:27:47.521093]<rts>dunno what server you are using ,but under apache you could setup different locations with different authentication schemes?!
[17:28:24.561155]<talley>yeah but it would interfere with AccountManager
[17:29:01.709075]<rts>damn application level authentication and authorization ;9
[17:29:35.666139]<talley>right now I'm looking at the wiki page rename ticket
[17:29:46.648054]<talley>I might backport it to our .11.x
[17:30:20.199834]<rts>pls provide link
[17:31:25.030294]<talley>http://trac.edgewall.org/ticket/1106
[17:32:23.311251]<rts>thx - there already exists a plugin on trac-hacks, in case you did not know
[17:32:56.338103]<talley>now that you've mentioned it, yes I remember glancing at it
[17:34:44.966122]<rts>btw. i am currently developing a more or less universal redirection system, which will hook into the existing change listeners and install redirections upon name/id change
[17:35:43.532946]<talley>sounds handy
[17:35:52.513903]<rts>and it would also require a new implementation of the rename plugin, afaik - i have not looked into the code of the existing rename plugin
[17:36:27.245985]<talley>will you provide a Create redirection next to Create wiki page button?
[17:36:50.879137]<talley>or a macro thingy
[17:37:14.573993]<rts>difficult, as that would require patching the original templates i presume, or is there another way to do this?
[17:38:06.462234]<talley>I only know there is a wiki/ folder, nothing more :)
[17:38:35.417805]<talley>but my guess would be that it might be sufficent to patch wiki/web_ui
[17:39:31.271985]<talley>but you know other wiki engines, right? like wikimedia(wikipedia) for example
[17:39:44.878595]<talley>you can toughtfully set up redirections
[17:40:23.041026]<rts>i looked into mediawiki, but it is somehow retarded, lest it is also php *g
[17:40:37.754470]<talley>though* of course
[17:41:28.987297]<rts>^^ of course
[17:42:14.872660]<rts>actually patching wiki/web_ui does not suffice *g (pun intended and taken)
[17:43:07.787759]<talley>:)
[17:45:01.503175]<rts>well, yes i thought about establishing the /redirection realm for manually setting up redirections, everything else would be automatic
[17:46:17.260788]<rts>and, redirections can even be cross-realm, meaning that for example one plugin defines /wiki/foo/bar to be a restricted area of the wiki so that it should be redirected to /foobar
[17:47:57.331794]<rts>that way plugins may reuse the versioning caps of the wiki system for their own purposes while providing their own specialized views
[17:50:19.241643]<rts>the plugin is currently done by approx 60% , i still need to implement the genshi templates and the *sigh* web_ui part
[17:55:11.238208]<rts>still on?
[17:57:39.534814]<talley>yes
[18:15:20.803625]<rts>and what do you think about the proposed/already implemented features of the redirection plugin, or rather the RedirectionSystemPlugin as i dubbed it?
[18:18:22.693054]<talley>I think I'm staying up too late
[18:19:09.651240]<talley>other than that, it looks like the mentioned ticket will obsolate your original implementation
[18:19:30.246572]<talley>but redirection functionality would be very cool, so you should still work on this
[18:21:34.092192]<rts>^^me too, it is 3:19 am now and i am getting dizzy...
[18:22:06.921225]<rts>which ticket, sorry i must have missed the linnk
[18:22:41.329657]<rts>oh i should replace the batteries, it gets btw. i am currently developing a more or less universal redirection system, which will hook into the existing change listeners and install redirections upon name/id changebtw. i am currently developing a more or less universal redirection system, which will hook into the existing change listeners and install redirections upon name/id change
[18:22:50.927615]<rts>aargh
[18:24:25.381451]<talley>[02:31] < talley> http://trac.edgewall.org/ticket/1106
[18:24:50.137326]<rts>thx for the link
[18:28:40.545269]<rts>*sigh* - but i will pursue on this since it seems to be the better solution...
[18:29:10.653740]<rts>btw. i am done for tonight, it is early 3:27 am here
[18:29:27.895735]<rts>cu and thx for the good chat, gn8
[19:43:31.829831]<oxez>Hi. Running trac 0.11.6 with mod_python 3.3.1, apache 2.2.14. When starting apache with trac, memory usage goes up by 130-150MB. Without trac it's fine.. I cannot seem to put the finger on what's causing this. Any ideas?