2008-08-20T00:04:11 miker_: AHA! 2008-08-20T00:04:43 oils_i18n_xlate is giving me grief, and I know why 2008-08-20T00:11:59 dbs pasted "Why we don't get translated values back from the DB" at http://paste.lisp.org/display/65561 2008-08-20T00:20:48 *** _bott1 has left #openils-evergreen 2008-08-20T00:21:23 *** kgs_evening has quit IRC 2008-08-20T00:24:01 So a possible fix would be to match against (fq_field = keyclass OR fq_field = keytable) 2008-08-20T00:24:18 works at psql, so I'll check that in and deal with the ramifications later 2008-08-20T00:28:30 *** dbs has quit IRC 2008-08-20T00:35:32 @later tell dbs I'm rolling back your change. the fq_field must be {class-hint}.{translated_column} not {fully qualified table} (the translation UI knows nothing of the table names, and can't set them). If the seed data doesn't match that then we'll have to change the seed data ... 2008-08-20T00:35:32 miker_: The operation succeeded. 2008-08-20T00:37:44 @later tell dbs so, your problem data would need to have fq_field = cam.value (or cam.description, for translating the description) 2008-08-20T00:37:44 miker_: The operation succeeded. 2008-08-20T00:38:20 @later tell dbs which brings up the other reason we can't use just {fully qualified table} ... we are translating more than one column on some tables 2008-08-20T00:38:20 miker_: The operation succeeded. 2008-08-20T00:41:47 @later tell dbs hrm... also, the identitiy_value is supposed to be the pkey from the row, not the string we're translating 2008-08-20T00:41:47 miker_: The operation succeeded. 2008-08-20T00:43:10 @later tell dbs we've got some disconnect in how config.i18n_core is being used... 2008-08-20T00:43:10 miker_: The operation succeeded. 2008-08-20T01:20:18 *** Mark__T has joined #openils-evergreen 2008-08-20T03:02:39 *** Mark__T has left #openils-evergreen 2008-08-20T03:03:50 *** Mark__T has joined #openils-evergreen 2008-08-20T03:05:53 *** Mark__T has joined #openils-evergreen 2008-08-20T03:17:56 *** Mark__T has joined #openils-evergreen 2008-08-20T08:10:01 *** Slazer has joined #openils-evergreen 2008-08-20T08:36:36 *** kgs_evening has joined #openils-evergreen 2008-08-20T08:40:15 *** kgs_evening is now known as kgs 2008-08-20T08:42:14 *** kbeswick has joined #openils-evergreen 2008-08-20T08:43:52 *** dbs has joined #openils-evergreen 2008-08-20T08:44:06 gah 2008-08-20T08:46:09 yeah, some disconnect... it would have been nice to have been corrected 8 months ago when I committed the db-seed parsing stuff 2008-08-20T08:47:01 and yeah, the en-US source doesn't have the object hint / property source in it, which means doing a lookup 2008-08-20T08:47:44 and primary keys can be all change-y from iteration to iteration, and that sucks 2008-08-20T08:49:14 and rolling back a change that adds functionality (even if its not in line with the intended design) and doesn't break existing functionality, without a functioning alternative, makes me grumpy 2008-08-20T08:50:57 anyways - shower time so I can cool down a bit (I was already grumpy due to anti-daycare amber; sorry for venting) 2008-08-20T08:51:01 *** dbs has quit IRC 2008-08-20T09:15:00 @later tell dbs you're right, I should have looked at the seed data. assuming on my part did what assuming does. I honestly didn't know it was even there (in an SQL form) -- that's my fault, not yours. It does break functionality, though, unfortunately :( 2008-08-20T09:15:00 miker_: The operation succeeded. 2008-08-20T09:15:23 @later tell dbs commuting ... eeevil iab 2008-08-20T09:15:23 miker_: The operation succeeded. 2008-08-20T09:52:22 @later 2008-08-20T09:52:22 kgs: Error: "later" is not a valid command. 2008-08-20T09:57:02 @help later 2008-08-20T09:57:02 dmcmorris_esi: Error: There is no command "later". 2008-08-20T09:57:21 @later tell dmcmorris_esi you're an idiot 2008-08-20T09:57:21 dmcmorris_esi: The operation succeeded. 2008-08-20T09:57:37 heh... "(09:57:29 AM) pinesol: Sent just now: you're an idiot" 2008-08-20T09:58:48 help @later 2008-08-20T09:59:17 @help 2008-08-20T09:59:17 kgs: (help [] []) -- This command gives a useful description of what does. is only necessary if the command is in more than one plugin. 2008-08-20T09:59:35 @help later tell 2008-08-20T09:59:35 dmcmorris_esi: (later tell ) -- Tells the next time is in seen. can contain wildcard characters, and the first matching nick will be given the note. 2008-08-20T10:12:13 * eeevil returns 2008-08-20T10:19:51 *** Mark__T has left #openils-evergreen 2008-08-20T10:36:29 *** dbs has joined #openils-evergreen 2008-08-20T10:42:28 hrm - all the unit tests ran clean... 2008-08-20T10:42:59 I kill myself 2008-08-20T10:46:56 so, for getting the class hint / properties into place, I guess we could add another parameter to oils_i18n_gettext() to hold that 2008-08-20T11:25:07 so I'm plugging ahead with a revised oils_i18n_gettext, but to do the primary key thing, all the defaults are going to have to be changed to explicit IDs 2008-08-20T11:25:48 and we'll have to repeat the ID as another parameter for oils_i18n_gettext() to have a trustworthy marker. ah well. 2008-08-20T11:25:55 grunt work, thy name is dbs 2008-08-20T11:28:51 dbs++ 2008-08-20T12:08:59 kbeswick: severals days back, you menteioned that you get a python error when you try to make-install ILS. i continue to get "zipimport.ZipImportError: bad local file header ..." errors.. same as yours? 2008-08-20T12:09:12 yeah 2008-08-20T12:09:28 until i re-run make install 2008-08-20T12:09:44 yeah, it doesn't alway happen to me, either 2008-08-20T12:10:31 I've never been able to figure out why that happens, myself. 2008-08-20T12:16:48 apparently setuptools isn't 100% in this area. sigh. 2008-08-20T12:17:39 sort of the opposite problem: http://www.gossamer-threads.com/lists/trac/users/37647 2008-08-20T12:21:18 miker_: so would you be opposed to a two-stage evolution (probably over the course of the next day or two) of the i18n infrastructure? 2008-08-20T12:21:33 stage 1: grab class hint and property instead of just table name 2008-08-20T12:21:48 stage 2: use explicit primary keys instead of the identifying string 2008-08-20T12:26:32 *** scottmcd has quit IRC 2008-08-20T12:27:17 on a different tangent, I take back everything I ever said about licensing and suggest that we follow the advice of the nice, knowledgable people here: http://www.softwarefreedom.org/resources/2007/gpl-non-gpl-collaboration.html 2008-08-20T12:27:52 (to wit: one top-level contributors file, one top-level license file, and inline license / copyright headers only where you're including something that deviates from the norm) 2008-08-20T12:35:02 ashgfhhagfhakjhfg 2008-08-20T12:35:04 asmodai: did you ever figure out how to dodge the setuptools "bad local file header in zipimport" bullet? 2008-08-20T12:35:12 sorry .. returning from meeting 2008-08-20T12:35:15 brain bleeding 2008-08-20T12:36:21 dbs: Never saw that message tbh 2008-08-20T12:36:29 asmodai: LIES! 2008-08-20T12:36:38 Yes, indeed. 2008-08-20T12:36:43 Every day, straight into your face. 2008-08-20T12:36:47 http://paste.lisp.org/display/27826 2008-08-20T12:37:11 that's peculiar 2008-08-20T12:37:18 btw, why not upgrade to 0.5? 2008-08-20T12:37:53 asmodai: we're running into the same error message with EG - nothing to do with genshi :) 2008-08-20T12:38:14 *** scottmcd has joined #openils-evergreen 2008-08-20T12:38:15 dbs: I'm not opposed to evolution at all, in any direction. the whole i18n issue is my fault, I know ... charging ahead without explaining what I'm doing 2008-08-20T12:38:53 asmodai: just thought that you might remember if/how you resolved that problem with genshi, and that it might be applicable to us 2008-08-20T12:39:50 eeevil: I don't think it's your fault - most of it wouldn't exist without you 2008-08-20T12:41:03 dbs: I think it's important that we can identify not only the field, but also the specific row that a translation relates to ... I'm far from an expert, but I think context will be important esp in non-latin languages, and (in my mind) that means the specific row ... does that make sense? 2008-08-20T12:41:23 dbs: ha! I can build a bunch of it and still screw it up ;) 2008-08-20T12:41:43 or, screw up the plans of those trying to work on it with me 2008-08-20T12:43:52 eeevil: well, yes and no 2008-08-20T12:44:01 I mean, you can't argue with perfection 2008-08-20T12:44:52 but in the hy-AM translation, there are 686 values in config.i18n_core; 684 are distinct - but the translations are identical for the repeated identity_values 2008-08-20T12:45:50 but sure - making oils_i18n_gettext() a 4-argument marker that accepts primary key value, en-US string, class hint, and property is fine with me 2008-08-20T12:46:16 because of course it will avoid any possible problems 2008-08-20T12:46:46 can you join #python-genshi? :) 2008-08-20T12:48:17 asmodai: I can 2008-08-20T13:17:22 *** kados has quit IRC 2008-08-20T13:26:33 the #setuptools guys claim that the problem is fixed in setuptools dev since 2006, but that there just hasn't been a release (so most of us are stuck with 0.6c8) 2008-08-20T13:27:23 quote - "To get the svn version, you can do "easy_install setuptools==dev"" - so we could forcibly upgrade people to the SVN version of setuptools (gah) 2008-08-20T13:29:58 *** scottmcd has quit IRC 2008-08-20T13:48:09 that's an impressive release cycle 2008-08-20T13:48:18 even emacs does better than that these days 2008-08-20T13:52:59 I'm going to go cook 2 hotpockets and chew in dbs' ear 2008-08-20T13:53:20 I've got some sloppy chili, right back at ya eeevil 2008-08-20T14:03:33 *** scottmcd has joined #openils-evergreen 2008-08-20T14:22:36 http://open-ils.org/dokuwiki/doku.php?id=faqs:evergreen_roadmap 2008-08-20T14:41:24 *** phase_bb has quit IRC 2008-08-20T15:06:12 http://browsing.justdiscourse.com/2008/08/08/thoughts-on-xulrunner-part-one-why-xulrunner/ 2008-08-20T15:37:12 intl.accept_languages 2008-08-20T15:42:24 *** Slazer has quit IRC 2008-08-20T15:43:44 berick: thoughts about roadmap? 2008-08-20T15:46:21 we've been talking about it on the call we're still on 2008-08-20T15:47:07 *** kbeswick has quit IRC 2008-08-20T15:48:13 ah 2008-08-20T15:48:34 fill me in later on any comments about it? 2008-08-20T15:49:01 roger roger 2008-08-20T16:50:03 *** kados has joined #openils-evergreen 2008-08-20T16:51:55 phasefx: heh - I used to work at big blue with the first commenter on that post you linked 2008-08-20T16:53:28 :) 2008-08-20T17:14:30 *** dbs has quit IRC 2008-08-20T20:31:06 *** phase_bb has joined #openils-evergreen 2008-08-20T20:43:47 *** dbs has joined #openils-evergreen 2008-08-20T21:08:19 man. are we sure the 'dojo not defined' exception is only coming from RemoteRequest.js on the client, and not from something else? 2008-08-20T21:11:02 yep, that's where it is 2008-08-20T21:11:14 jeez. my iq must be hovering in the high 80s 2008-08-20T21:11:24 just above mine 2008-08-20T21:23:00 INT=8 2008-08-20T21:31:48 *** sarabee_ has joined #openils-evergreen 2008-08-20T21:31:54 *** sarabee has quit IRC 2008-08-20T21:48:02 *** _bott1 has joined #openils-evergreen 2008-08-20T21:54:38 *** scottmcd has quit IRC 2008-08-20T21:54:38 *** sboyette has quit IRC 2008-08-20T21:54:38 *** _bott1 has quit IRC 2008-08-20T21:54:42 *** scottmcd has joined #openils-evergreen 2008-08-20T21:59:40 *** sboyette has joined #openils-evergreen 2008-08-20T22:01:46 still hitting "locale not defined" errors but I think that's coming from outside of RemoteRequest.js 2008-08-20T23:47:21 on the i18n side... I have db strings coming through, using the primary key and all. w00t. 2008-08-20T23:49:07 dbs++ 2008-08-20T23:49:14 wow ... that was fast as all get-out 2008-08-20T23:50:24 miker_: i'll need your sharp eyes to double-check my commit, of course 2008-08-20T23:50:52 * miker_ goes now so as to avoid more ass-ification of self 2008-08-20T23:53:39 no no, I'm serious - massive touching of the database seed values in particular, and there might be a more elegant way 2008-08-20T23:54:09 of handling oils_i18n_gettext() than my hamfisted "give it two separate signatures" approach 2008-08-20T23:54:22 then again, it is awfully close to midnight :) 2008-08-20T23:54:45 oh, I'm serious too ... I fell down before ... it's irking me :P 2008-08-20T23:56:32 almost there .... 2008-08-20T23:56:39 ha wrong channel 2008-08-20T23:58:29 heh