Friday, 2010-07-02

*** robmyers_ has quit IRC00:00
*** karlushi has quit IRC00:00
*** 13WAAGRQG has quit IRC00:00
*** mralex has quit IRC00:10
*** robmyers_ has joined #cc00:11
*** karlushi has joined #cc00:11
*** 13WAAGRQG has joined #cc00:11
*** koo5 has joined #cc00:35
*** luisv has quit IRC00:44
*** akila87 has joined #cc00:56
*** akila87 has left #cc00:56
*** Bovinity has joined #cc01:10
*** robmyers_ has quit IRC01:11
*** karlushi has quit IRC01:11
*** 13WAAGRQG has quit IRC01:11
*** robmyers_ has joined #cc01:22
*** karlushi has joined #cc01:22
*** 13WAAGRQG has joined #cc01:22
*** sudiptamondal has joined #cc01:24
*** JED3 has joined #cc01:24
*** johndoigiii has joined #cc01:25
*** sudiptamondal has quit IRC01:28
*** sudiptamondal1 has joined #cc01:29
*** JED3 has quit IRC01:29
*** johndoigiii is now known as JED301:29
*** JED3 has quit IRC01:35
*** DNS777 is now known as [777]01:45
*** sudiptamondal1 has quit IRC01:45
*** timhwang has joined #cc01:48
*** sudiptamondal has joined #cc01:57
*** sudiptamondal has quit IRC02:01
*** JoiIto1 has left #cc02:24
*** sudiptamondal has joined #cc02:56
*** aurelia has quit IRC02:59
*** JED3 has joined #cc03:03
*** akila87 has joined #cc03:19
*** nkinkade has quit IRC03:30
*** sudiptamondal has quit IRC03:43
*** sudiptamondal has joined #cc03:43
*** sudiptamondal has quit IRC03:53
*** Oxygeek has joined #cc03:54
*** timhwang has quit IRC04:14
*** Bovinity has quit IRC04:40
*** Oxygeek has quit IRC05:08
*** mick_laptop has quit IRC06:24
*** mick_laptop has joined #cc06:34
*** FHaag has joined #cc07:36
*** JED3 has quit IRC07:50
*** wormsxulla has quit IRC09:23
*** akila87 has quit IRC09:26
*** akila87 has joined #cc09:34
*** [777] is now known as DNS77709:38
*** wormsxulla has joined #cc09:41
*** akila87 has left #cc09:42
*** JoiIto has joined #cc13:03
*** tvol has joined #cc13:16
*** akila87 has joined #cc13:44
*** michi_ has joined #cc14:05
*** nkinkade has joined #cc14:28
*** nkinkade has left #cc14:36
*** sudiptamondal has joined #cc14:37
*** nkinkade has joined #cc14:40
*** sudiptamondal has quit IRC14:44
*** MarkDude has quit IRC15:01
*** MarkDude has joined #cc15:11
*** aurelia has joined #cc15:18
*** mick_laptop has quit IRC15:50
*** mick_laptop has joined #cc15:50
*** mralex has joined #cc16:00
*** FHaag has quit IRC16:02
*** nyergler has joined #cc16:03
nyerglernkinkade, fyi, i got the data from amazon copied onto another external drive16:05
nyerglerneed to get some different screws to put those drives in the chassis and make the array16:05
nyerglerparoneayea, ping16:11
nkinkadenyergler: Which drive did we have that would hold all that data?16:14
nyergleri have a personal 2T external drive16:14
nyerglerdidn't want to deal with HFS+ on the time capsule16:15
paroneayeanyergler: pong16:17
paroneayeaI just saw the .au email16:17
nyerglerparoneayea, great, that's what i wanted to check in about (that and the serbia email)16:17
paroneayeanyergler: so the serbian deed should show up in both languages?16:20
paroneayeathere's special code for serbia then I assume that I must have missed16:21
paroneayeaoh wait16:21
nyerglerparoneayea, it looks like all the multi-legal-code support may have been left out16:24
nyerglerfor example,
nyerglerparoneayea, iirc we had conditionals in the old for these16:24
nyerglerand we were moving them to RDF assertions16:24
paroneayeayeah I'm seeing it in now16:24
paroneayeathe conditionals are still there, they just must not be working with the new cc.license api's attributes16:25
paroneayeaokay, I can fix this.16:25
paroneayeaworking on both now :)16:25
nyerglerparoneayea, great16:27
nyerglerparoneayea, ideally we wouldn't need the conditional in the template and would just show all legalcode assertions from the RDF16:27
*** DNS777 is now known as [777]16:27
paroneayeafixing it for the old way is super-fast (license.jurisdiction -> license.jurisdiction.code)16:28
paroneayeaso why don't I do that now and then16:28
paroneayeawe can do things the smart way after I've got it back up and running16:28
paroneayeaboth bugs were related to license.jursidiction -> license.jurisdiction.code actually... :)16:28
paroneayeaI'm going to do a grep through the templates to double check that error hasn't been made again16:29
*** rjcp has joined #cc16:30
*** rjcp has joined #cc16:30
nyerglerparoneayea, great16:35
*** rjcp has quit IRC16:35
nyerglerparoneayea, i suppose this will be our first experiment with clearing the "cache" :)16:35
*** akila87 has left #cc16:35
*** akila87 has joined #cc16:36
*** rjcp has joined #cc16:37
*** rjcp has joined #cc16:37
paroneayeahey akila87... just FYI, I haven't forgotten about your plogpost.  I just have a couple of critical things to take care of first16:38
paroneayeaheh, plogpost.16:38
*** bassel has joined #cc16:39
akila87paroneayea sure :) don't worry about it. :) i just logged in automatically :)16:41
mralexwow, my irc window is littered with :)16:42
paroneayeamralex: :)16:42
mralexadium renders them as icons. it's like my screen has the measles.16:42
*** [777] is now known as DNS77716:44
*** timhwang has joined #cc16:53
nkinkade paroneayea: Where is the jurisdiction's website URL stored these days?  Or is it?16:58
nkinkadeparoneayea: As in: ./bin/jurisdiction add --lang=es_PR --uri=  pr ...  where is --uri stored?17:02
nkinkadeNEvermind.  I found it in jurisdictions.rdf.17:04
*** JED3 has joined #cc17:24
paroneayea <- As a test, I first tried wiping *just* rs's cached pages17:29
paroneayeawhich looks good17:29
paroneayeathere are a lot of them though, so I'm just going to wipe the entire cache directory I guess?17:29
paroneayeaand see how much it slows things to a crawl for a while... :)17:29
*** JED3 has quit IRC17:29
paroneayeait appears to have not destroyed the website in the process too :)17:31
paroneayeanyergler: cache wipe experiment a smooth success17:32
*** FHaag has joined #cc17:33
nyerglerparoneayea, awesome :)17:33
nyerglerparoneayea, can you reply to both serbia and australia?17:33
nyerglerthanks... am i correct that you're going to look at implementing the serbia fix "right" next?17:34
*** johndoigiii has joined #cc17:35
nkinkadenyergler: Where does the API get it's data?  From license.rdf?17:46
nyerglernkinkade, yes17:46
nyerglernkinkade, specifically, it now uses the cc.license API, just like the license engine17:47
nkinkadenyergler: So one thing that occurred to me is that we also need to retrieve the jurisdictions website URL, which is now only in jurisdictions.rdf.17:48
nkinkadeI code around that by fetching and caching that file.17:49
nyerglerit seems to me that if we're going to add an API call, let's just send everything in the packet there17:49
nkinkadeWhat I'm wondering is this: should I try to do this using the tools we already have, or should we write a new API method?17:49
nyerglernkinkade, i don't think we want to parse *all* of licenses.rdf (even if we cache it)17:50
nkinkadeI can make something that works using what we already have.  I don't think it would be too much work at all.17:50
nyerglerer, index.rdf17:50
nyerglerindex.rdf is 49M of RDF17:50
nkinkadenyergler: that wouldn't be necessary.17:50
nkinkadeI'm talking about jurisdictions.rdf, which is only 150K at the moment.17:50
*** rjcp has quit IRC17:50
nkinkadeStill not exactly a tiny file, though.17:50
nyerglerand then you'd make the call that generates a "BY" license to determine the version?17:51
nkinkadeSomething is going to have to parse it one way or the other.  It comes down to this: is it better to have the API parse it, or just do it in PHP.17:51
nkinkadenyergler: I already started poking around with some code changes.  The idea I was starting to write out was to just fetch jurisdictions.rdf and cache it for a predetermined amount of time.  Maybe a week or so?17:52
paroneayeaGood thing I did that grep to see if there were any more license.jurisdiction errors when it should have been license.jurisdiction.code.  There was!17:52
*** rjcp has joined #cc17:52
*** rjcp has joined #cc17:52
paroneayeathe class on the deed-main-content div17:52
nyergleri don't think that's quite the right dichotomy; i think it's more like, "how many parsers do we want to maintain, and is the incremental cost of making an HTTP call too expensive?"17:53
nkinkadenyergler: The parser I have in mind would be all of maybe 4 lines of code.17:54
nkinkadeJust load the document and fine a single element.17:54
nyerglernkinkade, and you're parsing as RDF, not XML, right?17:54
nkinkadeBut it's still 150K.17:54
nkinkadenyergler: I was thinking to parse it as XML and then fake the rest.17:54
nyerglernkinkade, no17:54
nkinkadeI wasn't sure how that would work, but figured it shouldn't be hard.17:54
nyerglerthat will fuck us eventually17:54
*** akila87 has quit IRC17:55
nkinkadeIt's an XML document, right?17:55
nyergleryes, but it's an RDF-XML document.  That means namespace labels, prefixes can change and still have the same meaning17:55
nyerglera naive XML parser won't get that right17:55
nkinkadeI don't know that there any good RDF parsers for PHP ... or at least searches didn't reveal anything totally obvious.17:55
nkinkade... or easy.17:55
*** akila87 has joined #cc17:56
nkinkadenyergler: Then I guess that means that John should just add a new API call.17:56
nkinkade... or we could just stop publishing jurisdiction site URLs.17:56
nkinkadeBecause I can already grab the current version number from the call I referenced earlier on that ticket.17:56
nyerglernkinkade, yeah, it's unlikely that we'd change things, but we have in the past when we added new terms, etc, so i don't want to set us up to fail like that17:56
paroneayeaheading to lunch, bbiab17:57
nkinkadenyergler: If we didn't bother to publish the sites URL, then we could actually eliminate the worldwide plugin altogether.17:57
nkinkadeIn fact, it would be easy enough to publish the URL in the body of the post instead of prgramatically.17:58
nkinkadeThat would certain ease things.17:58
nyerglernkinkade, i don't understand why moving the site URL to the body would eliminate the plugin; what about the license URLs?17:58
nkinkadeRight now we have to go changing XML and RDF around to effect that change, whereas it would be trivial to simply edit the post.17:58
*** akila87 has left #cc17:58
nkinkadenyergler: Because the code necessary to fetch the current version number would be all of 5 lines.17:59
*** MarkDude has quit IRC17:59
nkinkadeAnd we could just stick it in the theme.17:59
nkinkadeThere is already plenty of code in the theme instantiating classes and running class methods, and iterating through results and fetching names via the API>17:59
nkinkadenyergler: That could be simplified to: fetch from AP, grab version number, create license links.18:00
nkinkadeThat wouldn't take more than 10 lines of code, probably, and would eliminate about 150 lines of code and whole plugin.18:00
nyergleri think there's value in having this as structured information18:01
nyergleri think this is an easy add to the API18:01
nkinkadeThere is also value in simplicity.18:01
nyergleradding it to the API makes the worldwide plugin simpler18:01
nkinkadeThe current setup is ridiculously complicated to achieve a ridiculously simple task.18:02
nyergleri concur; but if:18:02
nkinkadeBy the way, what structure are you referring to?18:02
nyerglera) we already have to update the RDF for cc.license @ the API when we launch18:02
nyerglerb) we're building some tooling with fabric to make that easier18:03
nyergleri don't see that using an API method is any more complex/difficult to maintain18:03
nyergleri'm referring to storing it in the jurisdiction RDF18:03
nkinkadenyergler: If there were an API call that returned 1) the jurisdiction site URL 2) And a list of all the current licenses .... then we wouldn't need the worldwide plugin.18:04
nkinkadeIt could disappear into about 15 lines of code in the theme.18:04
nyerglerthat's the sort of call i'd like to implement; i'm quite annoyed right now that you keep conflating the place we store the data and how we structure the code18:04
nyerglerlet's implement the API call so we know where we're storing the information and then we can look at how the code work18:05
nyergleri'll update the ticket18:05
nkinkadenyergler: I don't understand "conflate" in this context.  But all I'm saying is that worldwide plugin would no longer be necessary because the amount of code required to handle it all would be trivial, and could just go right in the theme.  Eliminating a kludgy plugin.18:06
nyerglersorry, just seems like there are two issues: 1) how we store/access the information (in RDF? in the post? local? via HTTP?), and 2) where the PHP code lives18:07
nyergleri concur that eliminating a kludgy plugin is awesome18:08
nyerglerlet's write a jurisdiction info API call and eliminate the plugin18:08
nkinkadenyergler: I'm happy to keep this all in the RDF, and to just make a new API call.  That seems useful, and not just for us.18:08
nyerglercomplete agreement :)18:08
nkinkadeThat sounds like next step.  Implement that API call, then we can look at the code.18:09
nyerglersorry to be so prickly18:09
*** rjcp has quit IRC18:10
nkinkadeYour pricklyness didn't come through in text very much.  Sorry if I wasn't explaining myself well, or was just being slow.18:11
nyerglerno worries; irc is a low-emotional-bandwidth transport :)18:12
*** rjcp has joined #cc18:22
*** rjcp has joined #cc18:22
nyerglernkinkade, what's the easiest format for you to handle in response from the API?18:26
nyerglerXML? JSON?18:26
nkinkadenyergler: I'd say JSON would be the easiest because it wouldn't require any parsing.18:26
nkinkadeBut XML is easy enough, if it's better to go with that for consistency with the rest of the API or something.18:27
nyerglernkinkade, i don't understand how it wouldn't require *any* parsing... or does PHP just turn it into an array for you?18:27
nkinkadenyergler: Yeah, PHP has built in functions to parse the JSON.18:27
nkinkadeSo, yes, it just spits out a nice array that I can walk.18:28
*** timhwang has quit IRC18:28
nyerglernkinkade, i just created a new issue for the API call18:29
nkinkadeThanks.  This actually seems like a useful call, our use case notwithstanding.18:29
*** rjcp has quit IRC18:29
*** jgay has joined #cc18:32
*** mlinksva has joined #cc18:35
nyerglernkinkade, i concur... and i'd like to keep us moving towards eating our own dogfood18:36
*** MarkDude has joined #cc19:13
*** michi__ has joined #cc19:50
*** johndoigiii has quit IRC20:04
*** jgay_ has joined #cc20:04
paroneayeaoh noes :(20:27
paroneayea"bad file descriptor" and hda errors in my dmesg20:28
paroneayeababel.core.UnknownLocaleError: unknown locale 'oci_ES'20:40
paroneayeanyergler: 'oci_ES'?20:40
*** FHaag has left #cc20:41
nyerglerparoneayea, i have no idea; nkinkade?20:41
nyergleris that one of Spain's many languages?20:41
paroneayeamsgid "lang.oci_ES"20:41
paroneayeamsgstr "Aranese"20:41
nkinkadenyergler: Yeah, it's another of Spain's langauges.20:41
paroneayeawell babel dislikes it20:41
nyerglerparoneayea, does babel have a locale for Aranese (not country specific) -- so "oci"?20:42
*** tvol has quit IRC20:43
nkinkadeparoneayea: It's a valid ISO 639-2 lang code:
nkinkadeBut it's named "Occitan" there.20:43
nkinkadeApparently in20:44
nkinkadeSpain it's known as Aranese in the Aran Valley or something like that.20:44
nkinkadeIgnasi from CC Spain has me add it, but it remains at 0% translation, so I suspect it's safe to ditch it for now.20:45
nkinkadeAnd for your disk errors: smartmontools ?????20:49
paroneayeacool, I'll look at that20:50
paroneayeait turns out the error and the dmesg message were unrelated20:50
paroneayeathe file descriptor error happened with trying to open file() 'rw'20:51
paroneayeanot sure why that was a problem with babel20:51
*** MarkDude has quit IRC20:52
paroneayeamaybe I'm being foolish about something basic right now.20:52
paroneayeawell babel didn't like oci_ES and it didn't like ast_ES, so I just ignored them both.20:56
paroneayeaapparently it doesn't like all of spain's silly languages20:56
*** JED3 has joined #cc20:58
paroneayeathose are all of the jurisdictions with translations that said "a creative commons license" instead of the right license20:59
paroneayeaand the offending strings20:59
paroneayealots of translations there offending20:59
nyerglergreat. fuzzify those fuckers20:59
nyerglerand i'd check the spain languages by hand quickly since they may have the same issue21:00
paroneayeakk, will do21:00
paroneayeaI'm not sure if opening the Catalog with read_po in babel w/o specifying the locale would create a problem on write?21:01
*** aurelia has quit IRC21:01
paroneayeabecause otherwise I could just have babel do it.21:01
nyerglerparoneayea, should not, but should be easy to figure out -- just do a git status :)21:01
paroneayeatru dat :)21:01
paroneayeadidn't look like it created a problem but then again they didn't have the issue so my script didn't try writing the file21:03
* paroneayea makes a note to file a bug with Babel21:03
paroneayea wonder why babel's dropping "python-format" on those items?21:06
paroneayeaI'm not even touching them21:06
paroneayeawhat does that mean, anyway?21:07
*** aurelia has joined #cc21:07
paroneayeankinkade / nyergler: ^^??^^21:08
nkinkadeparoneayea: It would appear that was added by gettext:
paroneayeankinkade: :O :O :O21:18
paroneayeathen why is babel, a python library, deleting this valuable flag!21:19
nkinkadeBut I still don't know precisely what it's supposed to mean.  I don't know that much about PO files.21:19
nkinkadeI think it may mean that part of the string may be code that shouldn't be altered by the translator.21:21
nkinkade... like our {%variable} tokens, or whatever we use.21:21
*** aurelia has quit IRC21:22
nkinkadeparoneayea: This is helpful too:
paroneayeankinkade: thx21:23
nyerglerparoneayea, figure it out?21:28
nyergler(sorry, wasn't paying attention)21:28
paroneayeanyergler: haven't figured out *why* it's being stripped yet21:43
paroneayeathough I've figured out what it is21:43
paroneayeaabout to hackily set it back anyway21:43
nyerglerparoneayea, i noticed that the version of babel seems to be older than before21:43
nyerglermaybe that's a 1.0 feature?21:44
nyerglercan't remember why we built our own version at one point21:44
paroneayeathe babel version I was using was the alleged "current" version21:45
paroneayeabut I will try now with the one in
paulproteusruh roh21:45
paulproteusparoneayea: I don't think my patches related to "python-format" ever.21:46
paroneayeapaulproteus: still your version is in the 1.0 series21:46
paroneayeathe one easy_install pulls is 0.9x21:46
paulproteus is a conversation about those patches from a while ago21:47
paulproteusBabel has pretty simple code, especially if you read the patches I wrote ages ago,21:47
paulproteus links to my patches and upstream's commit log messages for them.21:47
paroneayeapaulproteus: ah ok, thx21:48
paulproteus is more chat log (actually from earlier this month)21:48
paroneayeanew babel still removes python-format though21:49
paroneayea(new being ~paulproteus/eggs/)21:49
paroneayeaseems peculiar21:49
paulproteusYeah, removing things it doesn't know about is typical Babel behavior when round-tripping a PO file through it.21:49
*** mlinksva has quit IRC21:51
paroneayeahey mlinksva was in #cc!21:52
paroneayeaI was even going to joke that the log you showed me had mlinksva in it back then21:52
paroneayeanyergler: hm, question21:57
paroneayeashould I have been marking those strings as fuzzy in i18n/i18n/, i18n/po/, or both21:58
paroneayeaI am guessing maybe I should have done it in po21:59
paroneayeaand then run po2cc21:59
paroneayeabut I am not entirely sure21:59
nyerglerparoneayea, probably both; one of those directories holds "po style" and the other holds "cc style" strings21:59
nyerglerparoneayea, actually what you said sounds right21:59
nyerglerpo2cc should carry the fuzzy over21:59
paroneayeawell that is not what I did :)21:59
paroneayeaI only did it in i18n22:00
nyerglerthat'll get overwritten22:00
paroneayeaI'll do it in po then22:00
paroneayeaand then run po2cc.22:00
paroneayeahm :\22:02
paroneayeadoing it in po/ seems to do nothing.22:02
*** mlinksva has joined #cc22:11
*** bassel has quit IRC22:13
*** mlinksva has quit IRC22:16
*** jgay_ has quit IRC22:17
*** bassel has joined #cc22:29
paroneayeankinkade / nyergler: one more thing before I check this in.
paroneayeais it okay that babel is adding Plural-Forms there?  (I'm not sure why it would be a problem)22:31
paroneayea looks like it would be useful even to have that added22:32
nyerglerparoneayea, yeah, that's fine22:32
paroneayeagreat :)22:32
nyerglerlooks like it added human-readable language header, too22:32
nyerglerthat's cool22:32
nkinkadeI'll go with whatever nyergler said. :-)22:32
paroneayeaone more question22:37
paroneayeawhen I'm doing an svn up in ~pootle/checkouts/$project/cc-i18n-trunk and I get a conflict, what should I do?22:37
paroneayearesolve the diff manually I assume22:38
paroneayeadamnit it opened the file in nano.22:39
*** JoiIto has quit IRC22:40
*** mick_laptop has quit IRC22:40
*** robmyers_ has quit IRC22:40
*** karlushi has quit IRC22:40
*** 13WAAGRQG has quit IRC22:40
*** JoiIto has joined #cc22:41
*** mick_laptop has joined #cc22:41
*** robmyers_ has joined #cc22:41
*** karlushi has joined #cc22:41
*** 13WAAGRQG has joined #cc22:41
paroneayeakk done... still gotta send that email to CCi but I'll do that next week.  Later all22:42
nyerglerparoneayea, later, thanks for your work today22:54
*** bassel has quit IRC23:02
*** KarlinTrouble is now known as karlcow23:14
*** nyergler has quit IRC23:15
*** mralex has quit IRC23:21
*** jgay has quit IRC23:24
*** wormsxulla has quit IRC23:28
*** wormsxulla has joined #cc23:48

Generated by 2.6 by Marius Gedminas - find it at!