2008-08-06T00:01:01 *** jeff__ has quit IRC 2008-08-06T00:01:12 *** jeff has joined #openils-evergreen 2008-08-06T00:04:54 okay, all done 2008-08-06T00:07:33 I think I'll build an Ubuntu 8.04 vmware image this time around, just for fun 2008-08-06T00:09:01 but for now, I'm going to go to sleep 2008-08-06T00:09:05 *** dbs has quit IRC 2008-08-06T00:15:44 *** phase_bb has quit IRC 2008-08-06T01:31:43 *** Mark__T has joined #openils-evergreen 2008-08-06T06:55:47 *** kgs has quit IRC 2008-08-06T06:55:48 *** kgs_ has joined #openils-evergreen 2008-08-06T06:55:51 *** kgs_ is now known as kgs 2008-08-06T07:05:50 *** kgs has quit IRC 2008-08-06T07:05:51 *** kgs_ has joined #openils-evergreen 2008-08-06T07:05:54 *** kgs_ is now known as kgs 2008-08-06T07:59:32 *** kgs_ has joined #openils-evergreen 2008-08-06T08:07:02 *** gmcharlt_away is now known as gmcharlt 2008-08-06T08:13:10 *** kgs has quit IRC 2008-08-06T08:18:43 *** Slazer has joined #openils-evergreen 2008-08-06T08:20:12 *** kbeswick has joined #openils-evergreen 2008-08-06T08:27:51 *** kgs__ has joined #openils-evergreen 2008-08-06T08:27:55 *** kgs__ is now known as kgs 2008-08-06T08:35:26 *** kgs_ has quit IRC 2008-08-06T09:15:51 *** dbs has joined #openils-evergreen 2008-08-06T09:16:32 so what do kbeswick and Slazer have in the works for today? 2008-08-06T09:17:18 Uhhh...I'm hitting these keys on this keyboard over and over again. *nods* 2008-08-06T09:17:47 bravo! 2008-08-06T09:20:28 Actually I ran into a problem when trying to replace a string with an XML entity... I don't think .html files use the xmlent or whatever it is called. I get a 500 internal server error. 2008-08-06T09:21:44 Slazer: if you get an internal server error, it means you definitely are using xmlent. 2008-08-06T09:21:59 good chance the xml is invalid 2008-08-06T09:22:13 but.. apache error log will give you more info 2008-08-06T09:26:08 xmllint --noout filename.html will tell you if it's well-formed XML or not 2008-08-06T09:26:12 dbs: immersed in H.C. A.T. 2008-08-06T09:26:33 "H.C. A.T." - kids these days 2008-08-06T09:27:39 kbeswick: if you have a chance, it would be awesome to 1) add the new OpenSRF Perl script berick added to autotools and 2) document what to do in the future when we add such a beast 2008-08-06T09:27:52 np b 2008-08-06T09:28:00 'H.C. A.T." - "Heavy Conversation with Ashley Thomson"? 2008-08-06T09:28:17 oh i only wish 2008-08-06T09:28:29 note that the new Perl script is only mostly functional 2008-08-06T09:28:29 Hardcore Autotools 2008-08-06T09:28:50 berick: yeah, but it should still be installed and accounted for properly by automake 2008-08-06T09:28:58 dbs: word 2008-08-06T09:29:21 * berick wonders if there would be pushback for adding template::toolkit for notification templates 2008-08-06T09:29:35 or if there is a better option 2008-08-06T09:29:37 berick: not from me ;) 2008-08-06T09:29:53 berick: no pushback here - or catalyst or whatever 2008-08-06T09:30:10 cool :) 2008-08-06T09:31:05 i'm working on example opensrf.xml settings for overdue/predue notifications... i'll send to the list shortly 2008-08-06T09:32:19 berick: I wonder if it should be overridden by an org unit setting? maybe just load up the default in a textarea and save the whole thing in the setting? 2008-08-06T09:32:33 berick: and then we can get Slazer to convert all of the entity stuff to the new templating system! 2008-08-06T09:32:35 (and there's locale to worry about too... 2008-08-06T09:33:04 eeevil: locale should be a per-org_unit and per-user thing of course 2008-08-06T09:33:19 absolutely 2008-08-06T09:33:28 but I'm thinking of overdue notice locale in particular 2008-08-06T09:33:36 right 2008-08-06T09:33:42 my stmt stands 2008-08-06T09:33:51 it does 2008-08-06T09:34:09 if we go with TT then that can be handled in-line in the template ;) 2008-08-06T09:34:18 ok, so, org and locale-specific notices 2008-08-06T09:34:42 [% IF usr.locale_preference == 'fr-CA' %] ... 2008-08-06T09:34:53 and we'll shove date/number/currency formatting through something appropriate 2008-08-06T09:35:18 berick: right ... it's the /where/ for locale awareness that is ... tough 2008-08-06T09:35:44 eeevil: how do you mean? 2008-08-06T09:35:45 dbs: we can do that in TT to ... I'm sure there are perl mods we can pull in for formatting 2008-08-06T09:35:47 we currently use entirely separate templates per language 2008-08-06T09:36:06 eeevil: right, right, I'm assuming TT is at least as flexible as PHP 2008-08-06T09:36:44 berick: well, do we use, as LU does now, entirely different templates per language, or do we push that into one tempate using [% block %] "functions" ... which I like the idea of, personally 2008-08-06T09:36:53 the latter of which, I mean 2008-08-06T09:37:40 right 2008-08-06T09:38:18 question: who will be editing the templates? Librarians? Or will a simple "Use this block of text" interface be given to them? 2008-08-06T09:38:38 dbs: i was imagining ils admins 2008-08-06T09:39:07 unless we want to create a friendly template editor of some sort 2008-08-06T09:39:48 in that case, the latter seems reasonable - more elegant, but slightly more complex 2008-08-06T09:40:21 IMO, "friendly" can come later ... a template compiler 2008-08-06T09:40:27 * kgs observes very few things are only "slightly more complex" ;) 2008-08-06T09:40:28 my thought too 2008-08-06T09:40:29 fwiw, in theory our system today automatically prepends address info / dates in the correct locale, so that librarians simply supply a block of text 2008-08-06T09:40:52 (in practice, of course, it just uses en-US for everything... sigh) 2008-08-06T09:40:56 heh 2008-08-06T09:41:18 Slazer: what file broke for you? org_unit.html? 2008-08-06T09:41:26 admin.html 2008-08-06T09:42:03 if we allow per-org templates, then putting them into the db makes the most sense 2008-08-06T09:42:09 eg needs a code editor... 2008-08-06T09:42:14 admin.html? weird, that seems to be clean xml here. did you add the DTD declaration? 2008-08-06T09:42:54 berick: you eventually have to have per-org templates. out of the gate, maybe not 2008-08-06T09:43:04 you/we :) 2008-08-06T09:43:10 *** agJohn has quit IRC 2008-08-06T09:43:26 dbs: that was my next qestion.. do they need to be in 1.4.0 2008-08-06T09:44:38 berick: hmm, for locale purposes you mean? 2008-08-06T09:45:04 dbs: no, i mean for org-specific purposes. i'm assumging locale-awareness is a must 2008-08-06T09:45:39 so, for 1.4.0, we go with locale-aware, global TT templates (on the filesystem) 2008-08-06T09:45:45 dbs: yes but then I'm not sure if I'm doing it properly. I'm just copying what I've seen in other files and pasting it at the top of admin.html 2008-08-06T09:45:46 I think org-specific templates could come later, personally. Of course, I'm not planning on actually using 1.4 :) 2008-08-06T09:45:58 then we add org-specific templates later... 2008-08-06T09:46:00 exactly 2008-08-06T09:46:17 dbs is going to use Evergreen ME 2008-08-06T09:46:41 berick: cool. simple but functional to begin with, then enhance. 'tis the way of the agile. 2008-08-06T09:46:46 dbs: Oh I think I got it. 2008-08-06T09:48:05 berick: first we need several hundred UML diagrams though 2008-08-06T09:48:14 dbs: yes and no... it is outputting the variable, i.e.: &conify.global.admin.......; 2008-08-06T09:48:15 so, while Slazer gets rolling on the conify i18n stuff, I'm going to write a dojo.i18n resource bundle -> PO roundtripper 2008-08-06T09:48:37 Slazer: okay, so it's not going through XMLENT 2008-08-06T09:49:09 dbs: well, those will have to wait for the meeting meeting 2008-08-06T09:49:38 * dbs can't really throw stones - he's been trying to nail down the next conifer meeting date for almost a month 2008-08-06T09:50:11 dbs: So what should I do? Rename it to .xhtml or add .html to go through XMLENT? Or...something else entirely? 2008-08-06T09:50:15 Slazer: your /etc/apache2/eg_vhost.conf has a conify section like this, right? 2008-08-06T09:50:42 dbs pasted "conify section for i18n" at http://paste.lisp.org/display/64867 2008-08-06T09:51:32 dbs: Not that I see. 2008-08-06T09:51:42 Slazer: ah, well that could certainly help 2008-08-06T09:51:54 add that and force-reload apache 2008-08-06T09:52:36 Ok. 2008-08-06T10:00:48 slazer: joy? 2008-08-06T10:01:37 dbs: Nope. 2008-08-06T10:02:03 Slazer: okay, my mind-reading wasn't getting me very far 2008-08-06T10:03:20 Slazer: so what's happening? 2008-08-06T10:04:07 Still showing the "&conify.global.blah.blah.blah;" and the xmllint says my entity isn't defined. 2008-08-06T10:07:45 *** Mark__T has left #openils-evergreen 2008-08-06T10:14:48 Slazer: so throw me a bone here. can you paste your error log please? 2008-08-06T10:15:50 The Apache error log? 2008-08-06T10:18:36 Slazer: right 2008-08-06T10:18:46 It is.... rather big 2008-08-06T10:22:03 tail is your friend 2008-08-06T10:22:48 Slazer: okay, sorry, I'll be more explicit - paste the relevant chunk that shows the xmllent error you're talking about 2008-08-06T10:23:33 fwiw, I have XMLENT substitution working here, but it craps out after setting up the side menu with: 2008-08-06T10:24:24 dbs pasted "XMLENT causes a bit of choking in conify admin.html - hmm" at http://paste.lisp.org/display/64871 2008-08-06T10:24:57 Slazer: and you might not need that conify chunk i pasted earlier, if your eg_vhost.conf has a different conify chunk at the bottom 2008-08-06T10:25:36 you would need to add the XMLENT option to AddOutputFilter though 2008-08-06T10:32:43 dbs: I have a similar error in my error log but it has to do with a file I haven't touched. 2008-08-06T10:33:53 Rather I HAD that error... haven't gotten it in my most recent restart of apache 2008-08-06T10:34:41 *** kgs has quit IRC 2008-08-06T10:39:19 *** kgs has joined #openils-evergreen 2008-08-06T10:41:12 *** mlasater has joined #OpenILS-Evergreen 2008-08-06T11:01:36 kgs: je think we need to start publishing release notices to places like oss4lib.org and other places that have rss feeds 2008-08-06T11:02:42 dbs: I fully agree and will work on that this week, if not this day, hour, or minute. Is there a release announcement? or um, am I writing it? ;) 2008-08-06T11:03:02 dbs: you're so good at those tech summaries, of course... 2008-08-06T11:03:10 we almost have RSS feeds for our news items 2008-08-06T11:05:57 dbs: so I can do a mini-procedure, what lists should these go to? 2008-08-06T11:06:17 sorry kgs, I threw that out there and then jumped on the phone 2008-08-06T11:07:14 (prompted by jfink actually, who exclaimed "why isn't this stuff showing up in any of my rss feeds?!?" when I mentioned 1.2.3.0 was out) 2008-08-06T11:07:25 http://open-ils.org/downloads.php has a decent miker_-written summary 2008-08-06T11:08:55 I think posting an item to open-ils.org/blog, oss4lib.org, and perhaps Freshmeat would cover the most pertinent bases 2008-08-06T11:09:02 heh -- backfilling the roadmap page with some of this ;) 2008-08-06T11:09:19 thanx! 2008-08-06T11:12:07 heading back to trac to pinpoint the release date 2008-08-06T11:15:58 8/5/08 2008-08-06T11:16:03 updating wikipedia 2008-08-06T11:17:11 btw, just added a freshmeat release announcement (our last announcement was 1.2.2.0, so we were due) 2008-08-06T11:22:24 really thinking most if not all of open-ils.org needs to be wikified 2008-08-06T11:22:34 and renamed ;) 2008-08-06T11:23:08 dbs: could you shoot me your announcement so I see how you phrased it? kgs@esilibrary.com 2008-08-06T11:24:05 kgs: actually I can't atm, the freshmeat release needs to be approved by the freshmeat.net crew themselves apparently, and the content lived in a browser form only - hopefully soon though 2008-08-06T11:25:14 Ok -- I just don't know what a typical release announcement reads like. "Congratulations! It's an upgrade!" 2008-08-06T11:26:05 kgs: I just reworded the features & fixes in the download description into two paragraphs of text, with new features first, followed by fixes. 2008-08-06T11:26:32 okey doke--perusing os4lib thanx 2008-08-06T11:32:28 *** pmurray_away is now known as pmurray 2008-08-06T11:32:28 dbs: ok... I've added the XMLENT to the AddOutputFilter in my but I'm still not getting anything. The line looks like 2008-08-06T11:32:28 AddOutputFilter INCLUDES;XMLENT .html 2008-08-06T11:32:28 Actually now though when I click on any of the links on the sides all of those pages that load in the iframe are broken and don't load properly if at all. 2008-08-06T11:33:17 Slazer: the line looks good 2008-08-06T11:33:47 * dbs taps his foot impatiently waiting for the error log paste he asked for an hour ago 2008-08-06T11:37:03 dbs: I'd paste something if I had an "[error]" line in my error log...but I don't. 2008-08-06T11:38:24 oh good, so no problems then 2008-08-06T11:38:48 still printing my entity though... and not the value. 2008-08-06T11:42:29 Well no errors until I click a link and get... 2008-08-06T11:42:31 Slazer pasted "error log" at http://paste.lisp.org/display/64885 2008-08-06T11:42:43 But I haven't touched anything but admin.html 2008-08-06T11:43:17 yeah, that's not really an error message, and you would have that in your logs from before 2008-08-06T11:45:03 I see. Hmmm... maybe I'm adding in the DTD stuff incorrectly. 2008-08-06T11:46:26 so: if you shift-reload the admin.html page, you get no errors in /var/log/apache2/error.log? 2008-08-06T11:48:36 Ohhh... I got something new. 2008-08-06T11:49:20 so all this time, you were just getting the cached version? 2008-08-06T11:49:32 ouch :) 2008-08-06T11:49:50 I though F5 was a hard reload.... 2008-08-06T11:49:57 * Slazer is embarassed 2008-08-06T11:50:33 heh - but now you'll remember for all eternity :) 2008-08-06T11:50:41 Oh yes. 2008-08-06T11:51:00 there is a method to my madness, rather than just walking 100 feet over to your office 2008-08-06T11:51:20 Heh heh. I do like your methods...yay for learning. 2008-08-06T11:53:54 dbs: all that needs to be done for the autotools/new perl script in opensrf is to add it to the dist, and have it install correctly, right? 2008-08-06T11:54:12 kbeswick: hell, you're the autotools master, you tell me! 2008-08-06T11:54:43 its a one line change 2008-08-06T11:55:23 want me to email you a patch or something? 2008-08-06T11:57:02 kbeswick: just send the patch to the list 2008-08-06T11:57:29 Slazer pasted "new errors" at http://paste.lisp.org/display/64887 2008-08-06T11:57:41 dbs: so...the first error.... it can't find lang.dtd 2008-08-06T11:57:49 there's not "en-US" in the path 2008-08-06T11:59:19 Slazer: right - because ${locale} isn't defined 2008-08-06T12:00:10 so you have a couple of options - hard code "SetEnv locale en-US" in eg_vhost.conf to force it to always be en-US; copy the SetEnvIf stuff from /opac/ in eg_vhost.conf 2008-08-06T12:01:18 the first approach would let you just get on with translating (and while you're at it, I would put the vars in a new file like conify.dtd) 2008-08-06T12:03:28 *** mlasater has quit IRC 2008-08-06T12:04:55 Hmmm, alright. 2008-08-06T12:29:07 man, professional_translators-- 2008-08-06T12:30:18 (over-generalization, but still... this is the second time we've sent out content for translation, and the second time they want to take a relatively simple translation format and return a Word file) 2008-08-06T12:32:00 dbs: did you specify that you didn't want it back as Word/TRADOS? 2008-08-06T12:32:02 berick: the major thing I stumble over in the notify_config.xml is that 14 days means ">=14 days" - but I don't have a better proposal 2008-08-06T12:32:08 that's pretty much The Standard, sadly 2008-08-06T12:32:16 sboyette: yes 2008-08-06T12:32:41 I WILL PAY YOU TO USE POEdit! 2008-08-06T12:33:29 "No, you will pay us to use Trados and return a Word file from which you can copy and paste freely." 2008-08-06T12:33:36 dbs: what's the concern, exactly? 2008-08-06T12:34:10 berick: the exact concern is that intuitively I would interpret that as "14 days, no more, no less" 2008-08-06T12:34:22 trados-- 2008-08-06T12:34:24 word-- 2008-08-06T12:35:03 dbs: originally, i had duration>14 days,, using the comma as an indicator of a range.. could do "14 days, *" or some such 2008-08-06T12:38:05 berick: I'm not sure there's really a good answer to be had in the XML format; alternative would be something like: 2008-08-06T12:38:40 duration/from, duration/to, and if no duration/to then assume to infinity 2008-08-06T12:41:04 dbs pasted "more verbose XML for circ duration range?" at http://paste.lisp.org/display/64897 2008-08-06T12:42:04 dbs: and if they want to nail it down to exactly 14 days? 14 days14 days ? 2008-08-06T12:42:23 that sounds reasonable 2008-08-06T12:43:03 that's more intuitive for me, anyway, but I'm not mr. typical 2008-08-06T12:45:34 berick: want me to reply to the list to see if it stimulates some discussion? 2008-08-06T12:46:39 we may need to model recurrance interval as well to get the correct round-off precision. if we generate notices at 7 days, do we check due date between 00:00:00 and 23:59:59 (what we do now) or between "runtime" and runtime+24 hours? or runtime+recurance_interval 2008-08-06T12:47:09 dbs: that would be great, thanks 2008-08-06T13:01:59 recurance and recuring, those spellings mock me to the depths of my soul 2008-08-06T13:03:04 * berick too 2008-08-06T13:16:54 *** gmcharlt has quit IRC 2008-08-06T13:37:38 *** agJohn has joined #openils-evergreen 2008-08-06T13:38:22 via /msg NickServ identify alphagjohn 2008-08-06T13:39:40 agJohn: you might want to change your password :) 2008-08-06T13:41:54 If anyone cares... 2008-08-06T13:58:12 *** agJohn has left #openils-evergreen 2008-08-06T13:58:47 *** agJohn has joined #openils-evergreen 2008-08-06T14:00:29 dbs: Word is you're the master of all things i18n-like.... Do you know if the XUL tools can support a locale that's as complicated as Arabic (bi-directional editing &c)? That is, would the 1.2.2.x staff client work with an Arabic locale? 2008-08-06T14:01:12 bidi is pushing it, I think 2008-08-06T14:01:32 well, for that matter 1.2.2.x is all en-US anyways 2008-08-06T14:02:09 Ah, good to know. Sounds like you don't think 1.4 would necessarily have resolved that either? 2008-08-06T14:03:29 our tools of choice for 1.4 (XUL + Dojo) can certainly handle bidi, but it will need someone from the arabic or hebrew community to tweak some of the layout 2008-08-06T14:03:31 *** sarabee_ has quit IRC 2008-08-06T14:04:16 Well, if these guys in Saudi Arabia want it, presumably they'd be up for tweaking. Thanks! 2008-08-06T14:06:35 agJohn: yeah. I expect to have to do basic i18n tweaking after 1.4.0 comes out and people start finding holes; bidi is another level of tweaking beyond that, but quite doable 2008-08-06T14:19:40 *** sarabee has joined #openils-evergreen 2008-08-06T15:01:55 so you know - I'm going to install yaz-3.latest on our Conifer test servers and run with that for a while. if there's no noticeable breakage, I'm going to suggest we upgrade Makefile.install accordingly 2008-08-06T15:02:20 righto 2008-08-06T15:02:30 it's been working nicely on my laptop (32-bit ubuntu 8.04), so getting a 64-bit debian etch testbed straightened out would seem to cover a lot of bases 2008-08-06T15:03:25 looks like Lenny has builds for 3.x (3.0.34-1, in particular) 2008-08-06T15:03:48 still no planned release date for Lenny, though 2008-08-06T15:14:29 *** Slazer has quit IRC 2008-08-06T15:33:53 kbeswick: dear autotools king: do other projects normally run "ldconfig" automatically as part of the install process if libs have been installed? 2008-08-06T15:34:20 I suspect that having to run it manually is unusual, but am willing to be wrong 2008-08-06T15:34:27 hmm... 2008-08-06T15:36:35 it looks like some do 2008-08-06T15:36:41 i will add it in 2008-08-06T15:38:12 *** kbeswick has quit IRC 2008-08-06T16:39:16 biz-frickin-zarre 2008-08-06T16:40:05 updated opensrf + ILS trunk on one of our conifer servers & rebuilt it, only to find that oils_cstore.so can no longer be loaded without LD_LIBRARY_PATH 2008-08-06T16:40:15 and this used to work. wth? 2008-08-06T16:40:35 dbs: rpath? 2008-08-06T16:41:26 berick: right, rpath - i guess we were setting that explicitly in opensrf weren't we? 2008-08-06T16:41:50 and then autotools came along and smooth-talked us to say that it would figure all of that out for us 2008-08-06T16:42:13 those smooth talkers are all the same 2008-08-06T16:42:40 when really it means "oh, I'll rpath it for you in a few certain edge cases, but I'm going to fully expect your .so's to be named lib*.so so that ldconfig can find you" 2008-08-06T16:43:50 it makes me gnash my teeth with great fury 2008-08-06T17:04:56 *** pmurray is now known as pmurray_away 2008-08-06T17:13:08 *** lisppaste6 has quit IRC 2008-08-06T17:13:22 *** lisppaste6 has joined #openils-evergreen 2008-08-06T17:23:08 kgs: oh, oss4lib-discuss! I actually meant oss4lib.org the web site (although maybe dchud has sorta-mothballed that?) 2008-08-06T17:23:20 but oss4lib-discuss is probably okay too :) 2008-08-06T17:24:24 dbs: the website didn't really want me to register... 2008-08-06T17:24:34 both are light traffic it appears 2008-08-06T17:24:38 anywhere else? 2008-08-06T17:24:40 yeah, I tried a month or so ago but didn't push it 2008-08-06T17:27:05 wow, that's kind of an inversion of the usual news associated with EG: http://blog.mlive.com/grpress/2008/08/grand_rapids_library_says_new.html 2008-08-06T17:27:14 "Grand Rapids Library says new online catalog will be easier, but limit interaction with other libraries" 2008-08-06T17:28:23 dbs: you noticed... 2008-08-06T17:29:23 kgs: yeah - the de facto "Planet Evergreen" blog aggregator is at http://www.ohloh.net/projects/evergreen/rss_articles - I don't check too often though 2008-08-06T17:31:33 it occurs to me these migrations should go on open-ils.org as well--I even have a pic of that library 2008-08-06T17:31:54 we have a working feed from ESI! http://esilibrary.com/esi/news.rss 2008-08-06T17:34:17 kgs: added to http://www.ohloh.net/projects/evergreen/rss_articles 2008-08-06T17:34:38 tip o' the hat to dmcmorris_esi 2008-08-06T17:38:32 and we really don't send all our press releases out at 8:00 AM... 2008-08-06T17:39:01 we just don't put a time on there, so the RSS just sets it to 0000GMT 2008-08-06T17:50:43 dmcmorris_esi: feedvalidator hates it a little bit, but that's probably okay :) 2008-08-06T17:52:58 looks like you could use within the description to satisfy the feedvalidator 2008-08-06T17:53:12 http://feedvalidator.org/check.cgi?url=http%3A%2F%2Fesilibrary.com%2Fesi%2Fnews.rss and http://feedvalidator.org/docs/howto/declare_namespaces.html (example at bottom) 2008-08-06T17:53:19 dbs: arg... yeah... 2008-08-06T17:53:51 but who's kidding who - i bet your feed works with any sane application, so dmcmorris++ 2008-08-06T17:54:21 dbs: thanks! let me give that a shot... 2008-08-06T17:58:37 dbs: bloglines hated it too... 2008-08-06T18:02:14 *** dbs has quit IRC 2008-08-06T18:06:26 kgs: how 'bout now? 2008-08-06T18:08:54 thanks kgs ;) 2008-08-06T18:09:13 now in production... http://www.esilibrary.com/esi/news.rss 2008-08-06T18:10:50 errrr... bloglines STILL chokin' on it... though it liked the test feed! "No feeds were found. Please verify that the website publishes an RSS feed" 2008-08-06T18:17:58 probably a cached copy... 2008-08-06T18:19:00 *** kgs__ has joined #openils-evergreen 2008-08-06T18:19:01 *** kgs has quit IRC 2008-08-06T18:19:03 *** kgs__ is now known as kgs 2008-08-06T18:39:54 *** kgs has quit IRC 2008-08-06T20:31:46 *** agJohn has quit IRC 2008-08-06T20:32:04 *** agJohn has joined #openils-evergreen 2008-08-06T20:33:24 *** agJohn has quit IRC 2008-08-06T21:11:19 *** kgs__ has joined #openils-evergreen 2008-08-06T21:11:23 *** kgs__ is now known as kgs 2008-08-06T21:15:43 *** dbs has joined #openils-evergreen 2008-08-06T21:17:49 hey miker_, semi-random question: why parser->ascii(1) rather than parser->utf8(1) for JSON::XS in OpenSRF::Utils::JSON.pm? 2008-08-06T21:19:19 * dbs reads on in JSON::XS - oh 2008-08-06T21:20:11 yup 2008-08-06T21:21:01 :) 2008-08-06T21:21:08 because berick told me to ;) 2008-08-06T21:21:15 because it's safer 2008-08-06T21:21:43 yeah - just noticed some strange happenings with some Czech chars 2008-08-06T21:22:01 ->utf8() kind of does the opposite of what I would expect 2008-08-06T21:22:06 dbs: exactly 2008-08-06T21:27:07 dbs: where were these happenings? 2008-08-06T21:27:09 OPAC? 2008-08-06T21:27:49 dbs pasted "Czech madness during marc2bre.pl" at http://paste.lisp.org/display/64947 2008-08-06T21:30:01 source record is http://sirsiweb.laurentian.ca/uhtbin/cgisirsi/x/0/0/5?searchdata1=16-0014071{001} (which can't display it properly anyways, natch) 2008-08-06T21:30:04 hrm... dbs: can you get the unicode codepoint for the offending char? 2008-08-06T21:30:10 miker_: ayup 2008-08-06T21:30:28 did that come out of unicron? 2008-08-06T21:30:38 of course 2008-08-06T21:30:55 if so, I suspect that unicron broke it (they broke other stuff on me before) 2008-08-06T21:31:07 what disturbed me the most was seeing the \t 2008-08-06T21:31:12 on me, over my head, across my neck ... you know ;) 2008-08-06T21:31:15 heh 2008-08-06T21:31:40 that's why I suspect borken output of some sort 2008-08-06T21:32:04 dbs pasted "here's another (different one, sorry)" at http://paste.lisp.org/display/64948 2008-08-06T21:32:09 (and I suspect that the correct characters are not "s" and "e" with accents 2008-08-06T21:32:16 miker_: actually they are 2008-08-06T21:32:28 http://www.springerlink.com/content/t21783v14081534j/ 2008-08-06T21:32:36 have you checked with a ... Czeck? ;) 2008-08-06T21:32:39 (well - S and c with accents) 2008-08-06T21:33:37 the accents are "carons" 2008-08-06T21:34:07 czechs love their caron 2008-08-06T21:35:58 bah, that last example is probably our fault 2008-08-06T21:36:19 previous example scrolled away 2008-08-06T21:36:21 gree 2008-08-06T21:36:23 grrr 2008-08-06T21:39:42 dbs pasted "yay for screen's scrollback - the original messed up example" at http://paste.lisp.org/display/64949 2008-08-06T21:40:02 so the codepoints would correspond to šěanĉar, Janez. 2008-08-06T21:41:59 and a hearty dbs-- for not appending to the original lisppaste :) 2008-08-06T21:42:09 hrm... well, one thing is that they're pre-compoesed 2008-08-06T21:42:22 yep 2008-08-06T21:42:39 nothing wrong with that, is there? 2008-08-06T21:43:07 well... that depends ;) 2008-08-06T21:43:29 did marc-dumper spit out the pre-composed version? 2008-08-06T21:45:05 dbs annotated #64949 with "wikipedia on carons and precomposed chars" at http://paste.lisp.org/display/64949#1 2008-08-06T21:47:25 oh, I know it's ok for unicode ... anyway ... I'm barking up the wrong tree with that 2008-08-06T21:47:55 so ... the problem is that JSON::XS doesnt' like \x{161} for some reason 2008-08-06T21:48:09 well... \U0161 2008-08-06T21:48:27 hrm... 2008-08-06T21:48:29 hex 73, hex 65 appears to be what I'm seeing in the yaz-marcdump'ed marc8-to-utf8 marcxml 2008-08-06T21:48:40 hold on a sec there ... 2008-08-06T21:49:33 dbs: mind if I send you a patch for marc2bre in a minute? 2008-08-06T21:49:39 miker_: not in the least 2008-08-06T21:49:39 you're using trunk, correct? 2008-08-06T21:49:52 correct 2008-08-06T21:51:19 (x75, x63 was what generated š at least) 2008-08-06T21:51:45 x61, x6e for ě 2008-08-06T21:53:23 yeah ... dang ol' marc::batch :( 2008-08-06T21:53:35 opening the files in :raw IO mode 2008-08-06T21:54:15 check your gmail :) 2008-08-06T21:54:17 ah! 2008-08-06T21:54:19 let's try that... 2008-08-06T21:55:49 * dbs kills 800K record import around 90% through 2008-08-06T21:55:57 arg! 2008-08-06T21:58:14 bah, it's nothing 2008-08-06T22:01:01 not quite the right place for "my $f;" -- fixing :) 2008-08-06T22:01:46 ha! 2008-08-06T22:01:57 yeah ... sorry. I was moving stuff around ;) 2008-08-06T22:02:41 hmm - no difference 2008-08-06T22:05:02 well.. adjust the open line (for this test) to read "open $f, '<:utf8', $_;" ? 2008-08-06T22:07:40 would you believe no difference? 2008-08-06T22:07:54 w t ... 2008-08-06T22:08:31 you've got my test record in the mail if you want to try it yourself 2008-08-06T22:08:44 k... thanks 2008-08-06T22:10:00 Oh nice, I was reading vim's output wrong for those characters apparently. 2008-08-06T22:10:13 har :) 2008-08-06T22:10:46 115, Hex 73, Octal 163 < ̌> 780, Hex 030c, Octal 1414 2008-08-06T22:11:38 so u0073u030c 2008-08-06T22:11:48 har! 2008-08-06T22:12:15 so that's decomposed 2008-08-06T22:12:23 aye 2008-08-06T22:12:32 U030C is just a combining caron 2008-08-06T22:13:55 right - so is the entityization in the XML wrong, or is JSON interpreting the entityized chars wrong, or both? 2008-08-06T22:14:29 well, the entityization is spitting out the precomposed character ... 2008-08-06T22:14:59 gah 2008-08-06T22:14:59 ok 2008-08-06T22:15:27 add a second param to the entityize call (stock marc2bre) ... 'D' (tell it to use NFD) 2008-08-06T22:15:33 then .. ingest likes it 2008-08-06T22:15:50 DANG IT ALL 2008-08-06T22:16:18 har de har 2008-08-06T22:16:49 seems only to be a problem with certain character ranges, though 2008-08-06T22:17:03 precomposed e-acute is fine... 2008-08-06T22:17:14 would you look at that, nice combining char entities :) 2008-08-06T22:17:35 so, JSON::XS is not fond of certain NFC codepoints 2008-08-06T22:17:36 so... sounds like the makings of a good set of ingest tests 2008-08-06T22:17:36 boooo 2008-08-06T22:17:42 haha 2008-08-06T22:18:06 * dbs looks around for a potential tester 2008-08-06T22:18:14 squints at his cat... hmm... 2008-08-06T22:18:40 ok ... so, I'll check that into head and 1.2[.3] 2008-08-06T22:19:08 yeah - i'm shocked that the armenian records (with their mix of armenian and cyrillic) seemed to ingest fine in comparison 2008-08-06T22:19:39 miker_++ 2008-08-06T22:20:13 dbs: thank you sooooo much for your input on that rss junk! ;) 2008-08-06T22:20:15 miker_-- #should'a cought that crap long ago 2008-08-06T22:21:33 dmcmorris_esi: heh - painful experience at the hands of trying to fix up some of EG's feeds :) 2008-08-06T22:22:38 dbs: actually, just committing to trunk, since it's just a JSON::XS thing 2008-08-06T22:23:20 miker_: groovy 2008-08-06T22:23:51 done 2008-08-06T22:24:14 now if I could just get vbox to quit dying 2008-08-06T22:25:09 miker_: it's open source, man, go to town (or do what I do, and just hang out on the channel and pose seemingly innocent questions...) 2008-08-06T22:25:20 * dbs has his cover on #openils-evergreen blown 2008-08-06T22:25:50 ha! 2008-08-06T22:30:58 miker_: should I assume that marc2are needs the same love? 2008-08-06T22:31:28 done did it :) 2008-08-06T22:31:37 svn up, capt'n 2008-08-06T22:31:45 and org_tree_html_options.pl to be sure? 2008-08-06T22:31:56 hrm... 2008-08-06T22:33:10 well... that's just generating html for display in the slimpac, but I suppose it wouldn't hurt 2008-08-06T22:33:28 that entityize doesn't know about NFD, though 2008-08-06T22:33:46 aren't they all locally defined anyway? 2008-08-06T22:34:07 they are 2008-08-06T22:34:27 I'm not saying no :) 2008-08-06T22:34:39 s/NFC/NFD/ 2008-08-06T22:34:42 ship it! 2008-08-06T22:34:46 heh 2008-08-06T22:35:52 done 2008-08-06T22:37:16 Now I can't wait until some corner case fails only in NFD() 2008-08-06T22:37:28 @roulette 2008-08-06T22:37:28 dbs: *click* 2008-08-06T22:38:03 yeah ... 2008-08-06T22:38:20 perl 5.10 will decide that only NFC e` is acceptable 2008-08-06T22:47:46 mmmmm perl 5.10 :) 2008-08-06T22:53:49 oh, btw, it may just be that people are using crappy xml libs and that's what's breaking the m:f:x build 2008-08-06T22:54:20 we got it going on sboyette's desktop with no apparent problems (sboyette: still true?) 2008-08-06T22:56:50 yep 2008-08-06T22:56:50 crappy as in expat? or as in old versions of libxml2? 2008-08-06T22:56:58 there's crappy and then there's crappy. 2008-08-06T22:57:03 it was totally a CPAN problem, seemingly caused by the bad naming 2008-08-06T22:57:54 well .. yeah ... there's that too 2008-08-06T22:58:06 which I guess was the root problem 2008-08-06T22:58:29 'cpan MARC::File::XML' fixed it 2008-08-06T22:59:24 that's too hard. 2008-08-06T22:59:50 sboyette: so, was /that/ what got past the missing namspacesupport package stuff? 2008-08-06T23:00:43 anyway, yeah. the package go released at some point as MARC::XML ... but that was all wrong. should have been MARC::File::XML 2008-08-06T23:01:11 miker_: yes, because it let M::F::X install cleanly 2008-08-06T23:04:03 * dbs notes that it's M:F:X in Makefile.install, curses anyone who tries to install with anything else 2008-08-06T23:04:15 what? there are other library systems? pshaw. 2008-08-06T23:04:32 i wasn't installing EG; i was working on some of the support tools 2008-08-06T23:40:30 *** scottmcd has quit IRC 2008-08-06T23:47:54 *** dbs has quit IRC