Tuesday, 2009-02-03

paulproteusnkinkade, Did you invite asheesh@asheesh.org ?00:01
nkinkadepaulproteus: I haven't yet.00:01
Bovinitynkinkade: interesting. when i joined "cc2008" it said i was creating a new room. there was no such message for "cc2009"00:01
nkinkadeI can't even seem to get back in the room.00:01
paulproteusAw, what a drag.00:01
paulproteus(nkinkade, thanks, that's a relief that I didn't miss such a message.)00:02
nkinkadeI guess I could just create a whole new room called "creativecommons"00:02
*** nkinkade has left #cc00:03
*** nkinkade has joined #cc00:03
nkinkadeI'm getting the feeling that Pidgin's support of XMPP might be limited.00:05
nkinkadeOr else conference.jabber.org is buggy.00:06
*** nkinkade has quit IRC00:10
*** nkinkade has joined #cc00:12
nkinkadepaulproteus: Did you by chance just receive any type of invite to asheesh@asheesh.org?00:13
paulproteusnkinkade, I did!00:13
paulproteusNow I have to figure out how to accept it.00:13
nkinkadeThings were acting weird, then Pidgin vanished and now that it's started back up it seems to work okay.00:14
Bovinityoh i got the invite.00:18
Bovinityin the web interface00:18
Bovinityi see you guys in there, i'm thinking nothing is getting through00:24
nkinkadeBovinity: You're getting the same error that Nathan Y was getting:  Alex Roberts left the room (This participant is kicked from the room because he sent an error message to another participant: service-unavailable).00:27
nkinkadeBovinity: I'm seeing your attempt to join.00:28
paulproteusI'm laughing at your attempt to join.00:28
Bovinitywho is NK-?00:28
paulproteusWell, I was.00:28
Bovinity(he was)00:28
BovinityNK- is screwing up my tab completion.,00:29
nkinkadeBovinity: I wondered that the other day.00:29
nkinkadeNicolas Kern?00:29
*** everton137 has joined #cc00:29
Bovinitynkinkade: i try sending varuous commands, and i keep getting Not Implemented  errors00:29
paulproteusSuch as00:29
BovinityAdium - using libpurple00:29
Bovinityto me it appears i joined just fine, adn can send messages fine00:30
nkinkadeBovinity: Interesting.  I was getting something similar earlier.  I could join, apparently, but not send a message or do anything.  Pretty useless.00:31
NK-yes nkinkade, how do you know my name ?00:32
nkinkadeNK-: /whois NK-00:32
NK-hah right !00:33
NK-thx for the tip ;)00:33
nkinkadeI'm NK too, by the way.00:33
Bovinityand hilarity ensues00:33
nkinkadeand tab completion goes to hell.00:33
nkinkadeBovinity: So Adium seems to join you to the room?00:34
nkinkadeIt must be a lonely room.00:34
Bovinityi see you and asheesh though00:34
Bovinityit was more lonely in cc200900:34
nkinkadeI'd say it's something with Adium, but Nathan was having the same problem with Pidgin.00:35
Bovinityadium and pidgin use the same chat library00:35
isforinsectsand Finch00:35
nkinkadeBovinity: ^^00:35
isforinsectsmmmm, command line IM00:36
nkinkadeIt's not solution, but it's current.00:36
Bovinityso it's broken then00:36
nkinkadeSeems like it might have to with Google.00:36
*** rohitj has joined #cc00:37
Bovinitycool, adium jsut crashed00:37
Bovinityi must have been booted from the chatroom00:37
nkinkadeBut interestingly the last message I see is that you joined without error, then left.00:39
nkinkade(the chat room, that is)00:39
*** tanjir has joined #cc00:40
nkinkadeIt's pretty hilarious, I know.00:41
nkinkademlinksva: What are you doing in the CC Google Apps GTalk?00:41
nkinkadeWord spreads fast.00:43
mlinksvankinkade: it just appeared00:50
mlinksvai logged out00:50
mlinksvamainly because there are ppl online i don't want to chat with :)00:51
nkinkademlinksva: Log back in and join creativecommons@conference.jabber.org00:51
mlinksvai'm going to have to tomorrow, the last thing i need is another distraction this afternoon.  i should bail from IRC :)00:52
nkinkadeUnderstood.  I better bail too, because I'm donewith.info.com (for the day)00:52
paulproteusI'm done with 'im.00:53
nkinkadeThat too.00:53
paulproteus'e was a jerk.00:53
nkinkadeI can see I'm going to have to get aggressive with my donewith. registrations.00:53
paulproteusCover it all.00:54
nkinkadeBut info.donewith and im.donewith doesn't have a good ring to it.00:55
nkinkadeAnd then I'd *really* be talking to myself.00:55
nkinkadeNow that's an idea.00:55
paulproteusSeems you have to have .{co,in,net,or,ac,go,mi}.th not whatever dot th.00:56
nkinkadeBut as it is now, I can do things like im.donewith.info, your.donewith.info, hes.donewith.info.00:56
isforinsectssecond level domains as opposed to tld00:57
nkinkadeBut resist.info is guaranteed to be gone, baby gone.00:57
isforinsectsor actually, third level domains00:57
paulproteusYou could register a .tel if you hurry.00:58
paulproteusI'm hoping to get donot.tel and a few others.00:58
greg-g I want a governmental thailand domain: im.go.th  ( http://en.wikipedia.org/wiki/.th )00:58
paulproteusAlso jonpos.tel.00:58
*** Bovinity is now known as mralexsoawesome00:58
nkinkadeisforinsects: paulproteus had the idea for nath@nkinka.de.01:01
nkinkadeNow that's me and natha.nkinka.de01:01
paulproteus"En kinka dot de"01:01
nkinkademlinksva: Don't be such an extremist.  That's just crazy.01:02
*** [mharrison] has joined #cc01:07
*** [mharrison] has quit IRC01:24
*** mralexsoawesome has quit IRC01:30
*** nkinkade has left #cc01:35
*** stevel has quit IRC01:37
*** mlinksva has quit IRC02:54
*** jgay has joined #cc04:13
*** mlinksva has joined #cc04:33
*** [mharrison] has joined #cc05:00
*** mlinksva has quit IRC05:03
*** tanjir has quit IRC05:03
*** jgay has quit IRC05:23
*** SDFE has quit IRC06:14
*** ergb546uy4ry has joined #cc06:20
*** balor has joined #cc06:42
*** balor has quit IRC07:36
*** [mharrison] has quit IRC07:39
*** SDFE has joined #cc08:17
*** ergb546uy4ry has quit IRC08:17
*** TRD|Away has quit IRC08:57
*** balor has joined #cc08:58
*** greg-g has quit IRC10:49
*** UncleCJ_ has quit IRC10:49
*** klutometis has quit IRC10:49
*** greg-g has joined #cc10:49
*** klutometis has joined #cc10:49
*** UncleCJ_ has joined #cc10:49
*** UncleCJ2_ has joined #cc10:50
*** Roderick_ has joined #CC11:27
*** Roderick_ has quit IRC11:28
*** [mharrison] has joined #cc11:29
*** TRD has joined #cc12:09
*** SDFE has quit IRC12:30
*** ergb546uy4ry has joined #cc12:30
*** Un1BG-scrip has joined #cc12:37
*** Un1BG-scrip has quit IRC12:39
*** sama has joined #cc13:22
*** K`Tetch has quit IRC13:55
*** rohitj has quit IRC14:05
*** K`Tetch has joined #cc14:15
*** mlinksva has joined #cc14:24
*** K`Tetch has quit IRC14:27
*** K`Tetch has joined #cc14:49
*** nkinkade has joined #cc15:25
*** ronr has joined #cc15:37
*** ronr has left #cc15:37
*** rohitj has joined #cc15:56
*** balor has quit IRC16:00
*** astanush has joined #cc16:27
*** astanush has left #cc16:28
*** stevel has joined #cc16:57
*** ergb546uy4ry is now known as SDFE17:00
*** grob has joined #cc17:10
*** TRD has quit IRC17:11
*** grob has left #cc17:11
*** sama has quit IRC17:17
*** stevel_ has joined #cc17:20
*** TRD has joined #cc17:23
*** stevel has quit IRC17:23
*** stevel_ is now known as stevel17:23
*** jgay has joined #cc17:25
*** Bovinity has joined #cc17:26
*** stevel has quit IRC17:50
*** stevel has joined #cc18:19
nkinkadeBovinity: Any ideas why the widget looks screwy on this page: http://sciencecommons.org/test-sc-donate-form/18:31
nkinkadeIt looks funny in both Firefox and Midori.18:31
nkinkadeIt's a copy of the actual widget named widget-sciencecommons.js18:31
Bovinitythere's a black bar pushing everything down18:31
nkinkadeThat's the only change.18:31
nkinkadeThat is the only things I changed about the file are the name, and the var thisScript bit.18:32
nkinkadeDid I do something else unintentionally?18:32
Bovinityhmm, the H1 in their stylesheet is overriding the H1 style for the widget, even though the widget one is set !important18:34
nkinkadeDamn.  Looks like a5 is screwed.18:34
Bovinityspecifically the margin property18:35
nkinkadeBovinity: Ah.  Thanks.  I can probably get it from there.  That was the clue I needed to get going. :-)18:35
nkinkadeIs anyone logged into a5 that can see what's up?18:35
nkinkadepaulproteus: ^^18:35
mlinksvaargh, bad time for site to be down18:47
*** oshani has joined #cc18:48
*** nathany has joined #cc18:55
nkinkademlinksva: Why's that?19:02
nkinkadeNo time is a good time, but I suspect that we were hit with the same problem we've been having, something to do with mod_wsgi and the scraper most likely.19:03
nkinkadenathany: Did you mention there was some problem with moving the scraper off of a5?19:04
nathanynkinkade: yes19:05
nathanyit has to run on cc.org due to javascript same-domain restrictions19:05
mlinksvankinkade: eric was on a phone call walking some people through licensing.  anyway, no time is a good time ... glad to see you have a lead on the problem19:05
nkinkadea5 went down again, and it was down for about 15 minutes.19:06
nkinkademlinksva: In a pinch staging can always be used.19:06
nkinkadenathany: So then is there any type of debugging in python that can be done for the scraper?19:07
nkinkadeAlso, it's a smallish thing, but the logs for the scraper have Unix timestamps instead of dates, which I can convert with date, but it's seems like it'd be more useful with regular human-readable dates.19:08
nathanynkinkade: so yes, there is debugging you can do although it's harder when the machine is belly up19:10
nathanyalso it seems like we're not sure if it's the scraper or mod_wsgi, right?19:10
nathanywrt the log, is it logging errors? (just curious)19:10
nathanyi think we're just using the standard python logger for that so i can set it up to format w/human readable date (IIRC)19:10
nathanyfinally, just curious why there were no email or text alerts re: a5; did we catch it manually before nagios?19:11
nkinkadenathany: Regarding alerts, I temporarily stopped Nagios so that it wouldn't flood us with SMS messages.19:13
nkinkadeI caught the problem before Nagios started sending alerts.19:14
nathanyah ok19:14
nkinkadeWe're not sure if it's the scraper or mod_wsgi, but the scraper is the only thing that uses mod_wsgi.19:14
nkinkadeI was somewhat shocked to see that the metadata.log file had over 50,000 entries.19:15
nkinkadeJust since it was rotated last.19:15
nathanythe deeds are arguably the most oft requested portion of the site... one line per request19:16
nkinkadeIT looks like metadata.log has normal stuff in it, not just errors.19:16
nkinkadeBut access.log is empty.19:16
nathanyno, access log has stuff in it19:17
nathanyoh nevermind19:17
nathanyi think i'm looking @ the wrong one19:17
nkinkadePart of the problem is that we don't generally know there is a problem until the machine is already belly up.19:22
nkinkadeUnless Nagios just happens to catch it because of some luck in timing and it's probes.19:22
nkinkadeOr one of us logs in.19:22
nkinkadeI suppose we could put some automated check in there that would look for warning signs and then run some debugging tests.19:22
paulproteus(/me waves to nkinkade and poor old a5)19:23
nkinkadeThis seems like something we should give high priority to.  It's not a fluke and it's not going away, and it causes every single main CC site to go totally away for 10 or 15 minutes.19:24
paulproteusnathany, Can this web app just be deployed using FastCGI instead of mod_wsgi?19:24
nathanynkinkade: right19:24
nathanypaulproteus: i think so; it was previously deployed just using mod_rewrite19:24
paulproteusI personally think that if we deployed the same app with Chinese FastCGI (or mod_proxy via mod_rewrite) the problem would go away.19:25
nkinkadeI just don't know what else to try.  I suspect this problem is something at a lower software level, low enough that when I get in it may go over my head.19:25
paulproteusI think the problem is just in the mod_wsgi module.19:25
paulproteusWeird that low things can go over heads when non-low things won't.19:25
nkinkadeThe bottom is low and deep, so when I step into those water I don't hit the bottom until I'm already over my head.19:26
nathanyso that's a reasonable supposition and it'd be good to decouple the two19:26
paulproteusThe bottom is low and deep, so I have Barry White sing it.19:26
paulproteusAnyway, I can try my hand at reconfiguring this thing with Chinese FastCGI (since I like its process cycling feature), perhaps with a different URL than the live mod_wsgi one until we're happy it works, or e.g. nkinkade can.19:27
paulproteusI'm likely to succeed since I've used Chinese FastCGI before anyway.19:28
nkinkadepaulproteus: What do you need me to do?19:28
nathanyi'm fine with that; you may need to write a wrapper that speaks fastcgi19:28
paulproteusnkinkade, Remind me the symptoms?19:29
paulproteusIt would be nice if we could at least install debugging symbols and get the mod_wsgi guys a useful backtrace.19:30
nkinkadepaulproteus: I sent a longish email with all the data a few weeks back.19:31
paulproteusOkay, I'll read through that again.19:31
nkinkadeA single Apache process starts to grow in memory usage and CPU usage until the OOM killer starts up.19:31
nkinkadeIf that one Apache process is killed, I think it returns to normal.19:32
nathanypaulproteus: let me know if you need pointers getting the app to speak fcgi19:32
paulproteusnathany, Okay, will do.19:33
nathanyi think http://webpy.org/install should be useful but it's so poorly formatted i can't tell :)19:33
nathanypaulproteus: http://www.flickr.com/photos/clementi/3251176498/19:36
paulproteusw00t pico19:37
paulproteusAlso, um, where's the WSGI configuration for this app? I don't see it in /etc/apache2/.19:38
nathany/var/www/creativecommons.org/conf/creativecommons.org, iirc19:39
nathany(it's a shared configuration file betwixt production and staging)19:39
nathanysort of19:39
Bovinityhmm, emacs is a golden spiral.. that explains a lot. also A4.19:39
paulproteusFound it, yay.19:40
nathanyBovinity: http://wuff.me.uk/Christ%20in%20cats/Christ%20in%20cats.html19:50
Bovinity(Viz Comic Issue No.85) -- UK's version of MAD, except with more dick jokes19:51
*** SDFE has quit IRC19:51
nkinkadeThe vi graph doesn't make sense to me.  Or maybe he switches the axes for that one.19:58
Bovinityvi is always hard to use; emacs travels through time19:59
*** ergb546uy4ry has joined #cc20:01
nkinkadeI can't say I agree that vi is always easy to use.  Set somone down who may be a command line wizard and if they have never used vi they will get nothing done.20:03
nkinkadeThe command mode / entry mode things is one huge barrier.  To say nothing of a massive amount of cryptic commands to do basic things.20:04
Bovinityand years of learning20:05
nkinkadeI would expect Pico to have the graph that vi has.  But I've now spent way to much talking about a graph of no consequence.20:05
Bovinitypico/nano are way easier20:05
Bovinitythey even have a bar that says what hte main commands are20:06
nkinkadeI'll leave a comment on his picture and point out that it's flawed. :-)20:06
paulproteusnathany, I'll write these scraper changes as changes to trunk and then when tested merge them into production; make sense?20:12
paulproteusnathany, Yaar, a fresh buildout of the scraper's trunk doesn't run.20:18
paulproteus  /home/paulproteus/cc-work/trunk/src/cc/deedscraper/app.py(62)<module>()20:18
paulproteus-> LogResult = decorator(LogResult)20:18
paulproteus  /home/paulproteus/cc-work/trunk/eggs/decorator-3.0.0-py2.5.egg/decorator.py(113)decorator()20:18
paulproteus-> fun = FunctionMaker(caller)20:18
paulproteus> /home/paulproteus/cc-work/trunk/eggs/decorator-3.0.0-py2.5.egg/decorator.py(67)__init__()20:18
paulproteus-> assert self.name and hasattr(self, 'signature')20:18
*** balor has joined #cc20:33
paulproteusnathany, I think I have the Apache end down fine; I'm struggling with the webpy end of things.20:40
nathanypaulproteus: in a meeting, i can look in a bit -- probably around 220:41
paulproteusI think I'm making progress anyway.20:42
nathanyit's probably something incompatible in the new version of decorator20:44
paulproteusnathany, Is there some relatively straightforward way I can test the scraper? I think it's working.20:52
paulproteusI guess I'll troll access_log for references to /apps/ .20:52
paulproteus(I have a FCGI one on /appz/)20:52
nathanypaulproteus: /appz/triples?url=http://foo20:52
paulproteusnathany, BTW, since we have only one log file even though we may have multiple of these processes running (even with mod_wsgi, right?), aren't you afraid they'll clobber each other's output?20:55
paulproteus"Just a thought."20:56
nkinkadeGreat name.21:00
paulproteusOkay, Chinese FastCGI is live on /apps/ .21:03
paulproteuse.g. these two work: http://creativecommons.org/apps/triples?url=http://asheesh.org/ and http://creativecommons.org/appz/triples?url=http://asheesh.org/21:04
paulproteusappz is now mod_wsgi (left enabled for demo and comparison purposes)21:04
paulproteusNow I want to reward myself with a burrito.21:07
paulproteusEr, wait, my fcgi processes don't seem to be persistent.21:08
nkinkadepaulproteus: Do you mean that /apps is mod_wsgi and that /appz is FastCGI?21:09
nkinkadeOr did you mean what you said? :-)21:09
paulproteusnkinkade, I mean what I said.21:09
nkinkadeOkay, but say what you mean in the future.21:09
nkinkadeOr say what you don't mean, and then I'll understand you perfectly.21:10
* paulproteus fiddles a little more21:11
paulproteusThere, they're persistent now.21:12
paulproteusI was testing with the cgi-script handler instead of fcgid-script.21:12
paulproteusThis is going splendidly.21:13
*** stevel_ has joined #cc21:16
nkinkadepaulproteus: So does this mean the end of the a5 meltdowns?21:21
paulproteus"We'll see."21:21
nkinkadeYou should run your nc thing against both and see if there is any measurable speed difference between them.21:21
paulproteusab would probably be smarter.21:22
paulproteusOh, the nc thing.21:22
paulproteusI see.21:22
nkinkadeSpeaking of quoted strings ... I saw an ad on the window of a BBQ restaurant here offering "Hawaiian Beer" brewed by Kona Brewing Company.  It wasn't clear to me what "Hawaiian Beer" is.21:22
paulproteusSee the email I just sent; tty in a bit.21:23
paulproteusCould it possibly be beer brewed in Hawaii?21:23
nkinkadeYes it is.21:23
nkinkadeBut why the quotes?21:23
nkinkadeMake it seem fishy.21:23
nkinkadeSudsy too.21:23
nkinkadeBut I was thinking the netcat thing would be good because you could feed it a file full of URLs to parse.21:24
nkinkadeI didn't think ab could do a thing like that, but I could be wrong.21:24
paulproteusI agree with you re: didn't think.21:25
nkinkadeSeems that would be a useful addition to ab, to pass it a file full of URL to process, and maybe the ability for it read them all into some internal structure and then even randomize them.21:27
*** mecredis has joined #cc21:29
*** stevel has quit IRC21:31
*** nathany has quit IRC21:42
*** nathany has joined #cc21:47
*** balor has quit IRC21:48
nkinkadeBovinity: Just for future reference.  The CSS issue has to do with the fact that the home page of SC and other pages use a different header template.21:50
nkinkadeWhich is why the widget looks good on the front, but bad elsewhere.21:50
Bovinitynkinkade: and those css files were interferring?21:50
nkinkadeI suppose so, but I'm not 100% in what way.21:51
nkinkadeIt may or may not be worth looking into.  But one thing does occur to me.  I wonder if it would be generally safer to use custom divs styled appropriately rather than using well known elements and altering their attributes.21:52
nkinkadeDoing the former could reduce the possibility of CSS clashes like this, I believe?21:53
Bovinitynkinkade: in theory it shouldn't matter. but realistically, yes, we should just do that.21:54
*** [mharrison] has quit IRC21:57
*** TRD is now known as TRD|Away22:00
*** [mharrison] has joined #cc22:01
*** musty has joined #CC22:12
*** musty has left #CC22:12
paulproteusnkinkade, Now I see your "22:19
paulproteus<nkinkade> Speaking of quoted strings .."22:19
paulproteusI didn't see that before, so I didn't realize that "Hawaiian Beer" was in scare quotes.22:19
*** jgay has quit IRC22:20
nkinkadepaulproteus: Was the encoding of my quotes messed up?22:20
paulproteusNot at all, it's just that I can't read.22:21
nkinkadeAh.  I just find lots of humor in the way people quote things.22:21
paulproteusI do, "too."22:21
nkinkadeEspecially funny are the slogans on the sides of company vans.22:22
nkinkadeThings like: Dickerson's Plumbing.  We're the "best in town."22:23
paulproteusSure you are.22:23
nkinkadeOr: Sylvia's "world famous" Pastries22:25
*** TRD|Away has quit IRC22:31
*** TRD|Away has joined #cc22:36
nkinkadeBovinity: Do you think we have some official SC and ccLearn donate widgets?22:51
Bovinitynkinkade: we do not22:51
nkinkadeI think it would be as easy as modifying the background image.22:52
nkinkadeBovinity: But I mean should we?22:52
nkinkadeIf we are now going to have special contribution pages for each of them.22:52
nkinkadeWe may not need it though.22:52
nkinkadeBecause we can alter the text below the logo.22:53
*** ergb546uy4ry is now known as SDFE22:53
nkinkadeOn second though, it's simpler to just leave it as it is and only change small bits of text.22:53
ianwellernkinkade: hehe i was in a dairy queen once where a sign they had (speaking about allergies and stuff) was entirely in quotes.22:54
nkinkadeianweller: Photos like that would make for a good Flickr collection.22:54
ianwellertoo bad i never carry my camera around22:55
Bovinitynathany: can you meet quickly about some of the white whale stuff?22:55
paulproteus(Let me know if I'm needed for some reason, too, Bovinity.)22:56
Bovinityneed to quickly see if what we're looking at right now would feesibly be compatible with SMW22:57
paulproteusMaybe I'll hop in22:57
*** rohitj has quit IRC23:19
*** phreakky has joined #cc23:20
paulproteus1. http://www.mediawiki.org/wiki/Extension:Lucene-search23:35
paulproteus2. http://www.mediawiki.org/wiki/Extension:SphinxSearch23:35
paulproteushttp://www.searchtools.com/analysis/mediawiki-search/index-mediawiki-search-stinks.html says #2 is better.23:35
*** rohitj has joined #cc23:36
*** Bovinity has quit IRC23:36
paulproteus#1 is the one I knew about.23:37
paulproteusnathany, ^^23:37
nathanythanks, paulproteus23:37
nathanywithout looking i think i like #2 since Spinx uses MySQL full text, right?23:37
paulproteusI think that's not a good enough reason to prefer one to the other.23:38
paulproteusAlso I don't yet know that that's true.23:38
paulproteusIf the user experience is way better in one, that's a good reason. Neither will be totally killer to run.23:38
nkinkadepaulproteus: Seems that admins in Pootle don't have rights to make translations on any given language.  Does that seem right to you?23:38
paulproteusLooks like Sphinx is a separate indexer, like Lucene, to me.23:38
paulproteusBut I'll stop looking for now.23:39
paulproteusnkinkade, They may have to add themselves to the language, which they can do because they're admins.23:39
paulproteusPootle separates "Can modify permissions" and "Can actually do anything in particular" as concepts.23:39
nkinkadepaulproteus: That seems like a poor "feature"23:39
paulproteusThe former is what MW calls Bureaucrats, and the latter what it calls Sysops. Pootle calls the former Admins and the latter nothing in particular.23:39
nkinkadeIt would just be nice if a site admin were already included in every possible group.23:40
paulproteusI guess, but then when a new language got created I'd have to add tehm.23:40
paulproteusAlso I think affiliates are only ever language admins, not global site admins.23:40
nkinkadepaulproteus: It shouldn't have to actually do any adding, but just a check to see if the user in an admin, if they are let them do it, whatever it is.23:40
paulproteusRight, that makes sense.23:40
nkinkadeI mean, if an admin can add themselves to a language to edit it, then it can't be a security thing ... just a hassle, sort of.23:41
nkinkadeBut I'll add myself for now. :-)23:41
nkinkadeAt least the interface makes it easy to add myself to every project and language with just a few clicks.23:42
paulproteusYou're joking, right?23:43
nkinkadeI wasn't, but I *was* wrong.23:44
nkinkadeI added all the language as language I track, but I guess I'll have to go into each language admin area and give myself admin rights, [Unicode] dammit!23:45
*** phreakky has quit IRC23:45
nkinkadepaulproteus: On second though, is there any reason I shouldn't just manually alter the PO files I'm interested in?23:53
paulproteusThat's what I generally do, I suppose. Why, what are you thinking?23:53
nkinkadeI thought maybe there was some new accounting that Pootle had to track changes, etc.23:53
nkinkadeI have to say that while Pootle is looking better that the interface in general is still wanting.23:54
paulproteusThey're Djangoifying it so the backend is less painful, and I think more UI work will follow that.23:54
nkinkadeAdministrating a language is not easy to find.23:54
nkinkade(Unless you already know where it is, and I frequently forget.)23:54
nkinkadepaulproteus: Did the location of the PO files change?23:57
paulproteusKinda, yeah.23:57
paulproteus~pootle/checkouts/ stuff23:57
paulproteusI think.23:57
paulproteusBut I think there are symlinks in place.23:57
paulproteusSince Pootle expects them in a certain place.23:57
paulproteusIt's all in ~/pootle/23:57
paulproteusEr ~pootle/23:57
nkinkadeThey aren't at /var/www/translate.creativecommons.org/po/?23:58
nkinkadeWe should clean that up, if not.23:58

Generated by irclog2html.py 2.6 by Marius Gedminas - find it at mg.pov.lt!