2010-08-09T07:04:18 *** Dyrcona has joined #evergreen 2010-08-09T07:32:17 *** shopkins has joined #evergreen 2010-08-09T07:33:59 *** mck9 has joined #evergreen 2010-08-09T08:17:18 *** gdunbar has joined #evergreen 2010-08-09T08:29:31 *** dbs has quit IRC 2010-08-09T08:47:00 *** cerpy has joined #evergreen 2010-08-09T08:54:34 * Dyrcona just checking if someone sent me an @later. 2010-08-09T09:00:09 *** dbs has joined #evergreen 2010-08-09T09:03:13 Dyrcona: auths should be loaded before bibs if the bibs you're loading use $0 subfields in your bibs to link controlled fields to the authority.record_entry.id of the corresponding auth records 2010-08-09T09:03:25 g' morning, dbs. can you point me at your lenny trunk vbox image? 2010-08-09T09:03:41 The 2010-07-16 extravaganza? Certainly 2010-08-09T09:03:47 yeah, thanks 2010-08-09T09:03:54 dbs: tyvm. I'm just getting to that bit now. You're timing is impeccable. 2010-08-09T09:04:51 berick: http://evergreen-ils.org/~denials/LennyTrunk_2010_07_16.zip 2010-08-09T09:05:36 thanks 2010-08-09T09:06:21 cerpy: two possible approaches for enlarging your VM window with that image: 1) ssh from your workstation into the image (that way you control the size of your terminal window) 2) start X on the image (and if it's not installed - can't remember - install a desktop & X) 2010-08-09T09:07:29 dbs: thanks - I stopped playing with virtual box and started using good old putty. 2010-08-09T09:07:38 *** rjackson-isl has joined #evergreen 2010-08-09T09:07:52 cerpy: right on, approach #1 for the lower-memory win :) 2010-08-09T09:08:30 dbs: I'm trying out the matching of 648 bib and 148 auth records. 2010-08-09T09:15:49 dbs: authority record imported - didn't I need to do something to activate the indexing of that specific record? 2010-08-09T09:16:44 cerpy: Depends on whether the image was updated to incorporate the changes I committed for in-database ingest (indexing) of authority records 2010-08-09T09:17:49 *** leed has joined #evergreen 2010-08-09T09:17:50 dbs: I think it is in the image which means it's automated - as far as I remember. 2010-08-09T09:28:00 network or server failure in my Staff Client. Is there anyway to get the message out of the way other than killing the app in windows task manager? 2010-08-09T09:29:10 Check the box that says "Ignore this error" and click Cancel I think 2010-08-09T09:31:02 *** bshum has joined #evergreen 2010-08-09T09:36:51 cerpy: and yep, 2010-07-16 includes the first cut at in-db authority ingest; there have been some bug fixes since (courtesy Dyrcona and others) but you should be good to go 2010-08-09T09:38:32 *** jenny has joined #evergreen 2010-08-09T09:40:54 Dyrcona: I would be very surprised if your bibs actually contained $0 subfields for authority control, but I'm easily surprised. 2010-08-09T09:41:21 I'd be surprised, too, but I'm going to load them first, just in case. 2010-08-09T09:41:52 There do appear to be links in the current ILS, but I'm not certain they done via MARC. 2010-08-09T09:43:04 dbs: 17310 -- the upgrade script, I believe the INSERT INTO asset.call_number_class needs to happen directly after the table is created. otherwise, the alter table on asset.call_number DEFAULT 1 fails 2010-08-09T09:43:17 trying to use a nonexistent value (DEFAULT 1) 2010-08-09T09:43:33 or something to that effect 2010-08-09T09:43:33 *** yboston has joined #evergreen 2010-08-09T09:44:12 Hrm, that sounds reasonable 2010-08-09T09:44:22 I can fixify 2010-08-09T09:45:10 awesome, thanks 2010-08-09T09:45:39 I guess I thought the deferrable status would hold off until the end of the transaction, but maybe there's no deferring a constraints check at DDL time for ADD COLUMN. Or something. 2010-08-09T09:45:54 You're seeing the upgrade fail? 2010-08-09T09:48:33 berick pasted "upgrade fail for dbs" at http://paste.lisp.org/display/113335 2010-08-09T09:49:16 heh, certainly sounds like it :) committed the fix. i think. 2010-08-09T09:49:26 dbs++ 2010-08-09T09:49:41 testing 2010-08-09T09:49:42 Obviously I focus on clean installs rather than the upgrade approach 2010-08-09T09:50:45 excellent, cuz i'm doing the opposite ;) 2010-08-09T09:51:53 We are a distributed testing platform 2010-08-09T09:52:05 heh. 2010-08-09T10:03:31 *** bshum has quit IRC 2010-08-09T10:05:49 dbs: looks like missing additions to the auditor table for asset.call_number. needed in the same order as the base table's column additions 2010-08-09T10:06:00 (only needed for upgrade script, of course) 2010-08-09T10:07:02 and needs defaults for the not nulls 2010-08-09T10:08:07 interesting that it needs a default, since it gets its values from asset.call_number, which provides a default 2010-08-09T10:08:22 I'm pretty sure it doesn't need a DEFAULT 2010-08-09T10:08:58 berick pasted "error 2 for dbs" at http://paste.lisp.org/display/113336 2010-08-09T10:09:05 not on an empty db 2010-08-09T10:09:25 * dbs resigns 2010-08-09T10:09:35 * phasefx hugs dbs 2010-08-09T10:10:19 *** bshum has joined #evergreen 2010-08-09T10:11:36 weird. but auditor.actor_org_unit_history doesn't define default values for fiscal_calendar 2010-08-09T10:12:52 ok, asset.cn is updated with a default value, /then/ the auditor table is tweaked, and does not get the default value 2010-08-09T10:13:08 since the value is not coming from the base table but from the addition of the new column 2010-08-09T10:13:12 * berick thinks out loud 2010-08-09T10:13:19 perhaps the NOT NULL constraint on the auditor history also needs to be deferred? 2010-08-09T10:13:46 I'm just saying - we've added lots of new columns in the past with NOT NULL DEFAULT values, and the corresponding auditor table doesn't have a default 2010-08-09T10:14:39 you got me there 2010-08-09T10:14:59 0151 defines fiscal_calendar on auditor.actor_org_unit_history without a NOT NULL clause 2010-08-09T10:15:17 but when it is created from scratch, it gets the NOT NULL 2010-08-09T10:15:22 explains that one 2010-08-09T10:15:43 *** Dmagick-home has joined #evergreen 2010-08-09T10:15:50 dbs: the auditor tables don't need a default, and don't get NOT NULL constraints 2010-08-09T10:15:58 Wouldn't be surprised if it explains them all. So the schemas for an upgraded DB will not equal a from-scratch DB. 2010-08-09T10:16:26 because, if the base table starts with a nullable column, and then adds a NOT NULL constraint, how do we represnt that? 2010-08-09T10:17:00 they'll be equal 2010-08-09T10:17:49 (excepting changes to the auditor-table-only part of the definition) 2010-08-09T10:17:51 * dbs closes eyes and allows the schemas to diverge 2010-08-09T10:18:57 *** collum has joined #evergreen 2010-08-09T10:19:03 the auditor tables are not exact schema copies, just column order (but not ordinal position) and data type. constraints are never copied over, nor are DEFAULTs 2010-08-09T10:20:25 now, I'm not saying that we've always added columns to the auditor tables when the base table changes, but that very well should break UPDATEs to the base table and we'd notice. 2010-08-09T10:20:59 So, I have a bit of a large patch for automatic update related items. Anyone want to take a look for major issues I missed before I mail it off to the mailing list? 2010-08-09T10:21:15 as for fiscal_calendar, there was some controversy about that not being an OU setting ... I think it is now, and that column should be gone from both base and auditor 2010-08-09T10:21:40 IIRC 2010-08-09T10:22:09 tsbere: offering a paste? 2010-08-09T10:22:49 No, a zip. The patch file itself is over 45k. 2010-08-09T10:23:16 tsbere: well, what do you mean by "patch for automatic update related items"? 2010-08-09T10:23:26 like, batch update of asset.copy? 2010-08-09T10:23:34 staff client 2010-08-09T10:23:56 Yea, sorry, thought I typed staff client into my test, that was the notes file though. <_< 2010-08-09T10:24:00 er, text 2010-08-09T10:24:13 *** r123 has joined #evergreen 2010-08-09T10:24:39 * tsbere should really install an irc client at home again, using one over RDP can be problematic when lag hits the screen updates 2010-08-09T10:24:48 oh, well, I'm not qualified to evaluate that :) ... ignore me 2010-08-09T10:25:46 tsbere: I say go ahead and post it. I'll be able to give it a spin later today 2010-08-09T10:25:54 I don't THINK I broke anything else, but I messed with configure.ac and some makefiles, so I think it would be best if it got a good look over before being merged in with trunk. 2010-08-09T10:26:26 http://www.mvlc.org/updates.zip < That should be it. Has a notes file and a dco already. 2010-08-09T10:28:41 phasefx: You were complaining about the build id generation before, right? I should point out my patch changes that part of the build process. 2010-08-09T10:28:41 Earlier this morning I posted a bug fix for the chunking code (osrf_application.c). Anyone who is working with trunk, or with a release that includes the chunking code, should apply that fix. It affects non-atomic system methods; I don't know if it affects any non-system methods. See OSRF changeset 1991. 2010-08-09T10:32:48 back to the upgrade problem.. we just need to remove the NOT NULL from the first auditor change? 2010-08-09T10:32:50 tsbere: just skimmed the notes for now; looks very nice. will test later 2010-08-09T10:32:56 +ALTER TABLE auditor.asset_call_number_history ADD COLUMN label_class BIGINT; 2010-08-09T10:33:03 berick: correct, I'm going to commit that 2010-08-09T10:33:10 oh, great, thanks dbs 2010-08-09T10:33:12 had somebody drop by to talk to me IRL 2010-08-09T10:33:43 *rolls eyes at RL* 2010-08-09T10:35:36 And then the question will be - wait, shouldn't the auditor schema changes happen first? 2010-08-09T10:35:56 * berick was wondering that as well 2010-08-09T10:37:31 and that's what the subsequent commit is for 2010-08-09T10:37:44 * dbs is powering up his trivial commit tally 2010-08-09T10:38:57 thanks for the assistance and patience, y'all 2010-08-09T10:40:03 tsbere: btw, see https://bugs.launchpad.net/evergreen/+bug/615376 2010-08-09T10:40:32 adding_patches_to_launchpad_bugs++ 2010-08-09T10:41:02 gmcharlt++ # for using launchpad 2010-08-09T10:41:32 phasefx: That....is a good point. 2010-08-09T10:42:16 On the other hand, at one point I was RUNNING the staff client in one folder while installing it to another. Hmmm........ 2010-08-09T10:42:32 depends on a few things. 2010-08-09T10:42:36 tsbere: dbs has a good one too; when you are ready to share this with a wider audience, you should make a wishlist bug in launchpad describing what you want to do, and then attaching the patch. Then you can point folks to the launchpad entry on the mailing list 2010-08-09T10:42:39 i would expect that to work, since it was different paths. 2010-08-09T10:42:47 jeff: exactly 2010-08-09T10:43:14 use case was somebody overwriting an existing installation 2010-08-09T10:43:19 berick pasted "db error for dbs 3" at http://paste.lisp.org/display/113338 2010-08-09T10:43:26 i'm liking what i've been able to do with nsis, though i wish it had some better support for editing configs 2010-08-09T10:43:29 dbs: you've created a monster ;) 2010-08-09T10:43:47 dbs: fwiw, i don't see the source of the problem 2010-08-09T10:43:57 Quick test on my installer, at least it comes to a halt and has a retry on the main program. <_< 2010-08-09T10:44:05 Assuming it can't overwrite 2010-08-09T10:44:29 independent of tsbere's work, I can imagine libraries including download links to the staff cient in the portal pages, and causing problems 2010-08-09T10:44:29 berick: augh, it's the index moving around 2010-08-09T10:45:44 and another commit. ka-ching 2010-08-09T10:46:19 Who is going to figure out which of the 364+ upgrade scripts actually need to be run for a 1.6.x upgrade, anyway? 2010-08-09T10:46:52 * tsbere is focusing on the staff client side of things, partially due to looking at the upgrade scripts and deciding he didn't want to touch that this month 2010-08-09T10:47:16 * gmcharlt knows who is going to help - a library who has a dedicated test database and server ;) 2010-08-09T10:47:55 Our current method of database update is "wipe and re-load the db", I believe. Yay being only a test setup for now. 2010-08-09T10:48:23 dbs: isn't that what rel_1_6[_*]/.../Pg/upgrade/* is for? 2010-08-09T10:48:37 * phasefx does both for development, but hasn't updated this morning yet 2010-08-09T10:48:39 heh. Unfortunately our dedicated test database and server isn't functional at the moment - eh, r123 ? 2010-08-09T10:49:14 but I'm not doing schema diffs 2010-08-09T10:49:35 berick: uhh... to get from 1.6.x to 2.0? 2010-08-09T10:50:38 mck9: thanks for the OpenSRF heads-up 2010-08-09T10:51:20 * tsbere wonders if it is an abuse of the term "wishlist" if you write the wishlist based on what you already coded 2010-08-09T10:51:23 dbs: oh, i read 1.6.x. upgrade as an upgrade to (say, a prev 1.6.x version), not from 2010-08-09T10:51:32 *** bshum has quit IRC 2010-08-09T10:51:52 *** bshum has joined #evergreen 2010-08-09T10:51:52 I guess we can cross that bridge when we get to it 2010-08-09T10:51:57 and.. that sentence made no sense. please to ignore 2010-08-09T10:52:20 berick: somehow it made sense to me 2010-08-09T10:52:25 judge me by my speak, do you 2010-08-09T10:53:45 I figured we'd figure that (1.6-2.0) out after alpha, since that will surely precipitate changes. I fear it less than 1.2-1.4, fwiw 2010-08-09T10:54:50 dbs: db upgrade successful 2010-08-09T10:55:02 gracias 2010-08-09T10:56:17 phew! I'm deeply sorry for the half-hour * 2 or 3 devs burned on that 2010-08-09T10:59:19 does IE8+Chrome Frame failing to render search results in Windows XP (but not Vista) ring any bells? 2010-08-09T11:00:00 gmcharlt: I hear nothing 2010-08-09T11:00:35 No bells here, but this cat haz no experience using Chrome Frame 2010-08-09T11:05:36 *** cerpy has quit IRC 2010-08-09T11:32:05 Suggestion: add launchpad to some more visible place on open-ils.org. As of yet, I have not found a link to it beyond a specific bug or two in the dokuwiki. I may have missed it, granted, but if I haven't spotted it yet it may not be visible enough. ;) 2010-08-09T11:32:57 I think that's a good idea. Launchpad was an experiment, but I think it's getting a good bit of use now 2010-08-09T11:36:30 oh, phasefx, I think I forgot to mention the additional eg.conf config for apache to make the updates "site" work. Don't forget that if you are testing the automatic updates themselves. 2010-08-09T11:36:52 could change the Contributing document if using launchpad like that seems like a good idea (I think so, easier to lose track of submitted patches in email, but it is more overhead) 2010-08-09T11:37:08 tsbere: roger roger 2010-08-09T11:39:01 tsbere: and just to be clear, this form of auto-update only works if your users are running as members of the Local Administrators group? Or is it just a function of "users have write access to the install location"? 2010-08-09T11:39:31 tsbere: Does this update xulrunner, or just the evergreen application? 2010-08-09T11:40:07 jeff: Combination thereof. On Vista/7 it will also work if they can be prompted by UAC. Also, if you make generic-updates only it will just update the evergreen app. The win-updates and linux-updates variants will update xulrunner too, provided that the app directory was structured properly. 2010-08-09T11:40:24 got it. thanks! 2010-08-09T11:41:12 jeff: General idea, if you have firefox automatically updating, the evergreen client works the same way. Extension mode is different (I did auto updating for that too) as it can be installed system wide or per profile, and permissions vary for each. 2010-08-09T11:41:24 * dbs slowly teaching open-ils.cat.biblio.record.marc_cn.retrieve to fetch call numbers based on asset.call_number_class 2010-08-09T11:42:02 tsbere: is the UAC interaction documented somewhere on mozilla's side? 2010-08-09T11:42:04 dbs: that method may also be squeezing certain subfields too close together (no intervening whitespace) 2010-08-09T11:42:15 *** Dmagick-home has quit IRC 2010-08-09T11:42:44 jeff: updater.exe, I think, handles that elevation. There is a nsiupdatesomething javascript in the xulrunner package that does the "can I attempt to update" checking. 2010-08-09T11:43:48 *** Dmagick-home has joined #evergreen 2010-08-09T11:43:58 jeff: Can't find anything beyond my testing that it prompted me, but I never dug into the source code of the binaries, just the javascripts 2010-08-09T11:44:48 yeah, i just started to have all kinds of questions about the autoupdate + UAC logic, and thought it might be documented somewhere else by someone else so i didn't have to pester you, since they're not specific to your patch 2010-08-09T11:45:42 nsUpdateService.js around line 170 in 1.9.2.8 is where it checks elevate possibility, I think. 2010-08-09T11:45:50 That should be in xulrunner/components 2010-08-09T11:48:02 * tsbere should consider a version of the windows installer nsis that, when a flag is set, marks the entire install directory as writable by all or something 2010-08-09T11:48:46 i have similar here. 2010-08-09T11:49:03 we store our prefs in the app dir, not the user profile dir. 2010-08-09T11:49:20 Yea, that would require some permission changes 2010-08-09T11:49:38 so our install process handles the acls and prefs changes. 2010-08-09T11:50:16 You take a look at my patch yet? May be more compatible with your install process than the original. 2010-08-09T11:50:50 yes, but we don't have need for the mozilla-style automatic updating. 2010-08-09T11:51:34 You mentioned that before. I ensured that part wasn't part of the "make a client" stuff beyond the config options being there if an automatic update host was set. 2010-08-09T11:52:09 yep! i appreciate that your patch doesn't force the use of automatic updates. 2010-08-09T11:52:18 (doesn't seem to, haven't looked it over completely) 2010-08-09T11:52:55 it defaults to not including automatic update settings at all in the client, you have to take extra steps to make that config happen. Although it does default the updates folder into existence. 2010-08-09T11:58:52 * tsbere should really have double-checked his notes file, he has confused - and _ in a lot of the makefile targets in the notes file 2010-08-09T12:13:38 *** jamesrf has joined #evergreen 2010-08-09T12:46:40 *** natschil has joined #evergreen 2010-08-09T12:52:22 *** natschil has quit IRC 2010-08-09T12:58:38 *** youdonotexist has joined #evergreen 2010-08-09T12:59:10 *** youdonotexist has joined #evergreen 2010-08-09T13:19:43 *** r123 has quit IRC 2010-08-09T13:32:24 *** r123 has joined #evergreen 2010-08-09T13:34:11 grrr. "open-ils.cstore: Attempt to reference non-existent column "id" on asset.call_number_class (acnc)" - I'm sure the IDL is okay, the table exists in the database, and I've restarted the services. but no cstore love 2010-08-09T13:36:16 dbs: fm_IDL.xml is the culprit 2010-08-09T13:36:22 no for id 2010-08-09T13:37:01 huh - I thought the oils_persist:primary attribute took care of that 2010-08-09T13:37:12 nossir 2010-08-09T13:37:16 Bah, guess not. 2010-08-09T13:37:17 dbs-- 2010-08-09T13:39:26 That works a lot better. thanks berick 2010-08-09T13:39:34 word 2010-08-09T13:39:58 Sad how you can stare at something for ages and try a few dozen different things without seeing the obvious 2010-08-09T13:40:25 That's funny "russian roulette", not funny "haha" 2010-08-09T13:40:25 hrm. is it just me, or is dojo getting uglier? 2010-08-09T13:40:34 http://download.dojotoolkit.org/release-0.4.4/dojo-0.4.4-src/tests/widget/test_FilteringTable.html 2010-08-09T13:40:41 vs http://www.dojotoolkit.org/reference-guide/dojox/grid/EnhancedGrid.html#nested-sorting 2010-08-09T13:42:18 *** r123 has quit IRC 2010-08-09T13:42:19 jeff: claro theme vs. whatever was in 0.4.4? 2010-08-09T13:42:57 yeah, not just the look, but also the way the nested sort works. 2010-08-09T13:44:57 the UI is... powerfully-complex, i suppose. 2010-08-09T13:45:06 well - does example #1 support nested sorting? 2010-08-09T13:46:30 the 0.4.4 example does, yes. 2010-08-09T13:46:39 albeit in a simpler, and less powerful way. ;-) 2010-08-09T13:47:21 Yeah, it's not working for me 2010-08-09T13:47:50 If I sort by "Date added", then try to nest the sort with "Date Modified", Data Modified just kills Date Added. 2010-08-09T13:48:03 (in the 0.4.4 example) 2010-08-09T13:49:04 on the first, the nesting is fixed at two levels. the most recent thing you sort is your primary sort. 2010-08-09T13:49:22 and it seems to work for me, using date added and date modified columns to sort, in chrome. 2010-08-09T13:51:12 Hmm. Maybe I selected something else first time around and now even a hard reload won't clear that in Firefox 2010-08-09T13:51:30 browsers-- 2010-08-09T13:51:33 uis-- 2010-08-09T13:52:04 not working for me in Chrome either 2010-08-09T13:52:13 ah well. just venting after a "oh, i want a client-sorted table here. dojo can help me there!" moment :) 2010-08-09T13:52:42 Maybe I can't figure out how to sort by multiple fields in 0.4.4, which would be a good positive argument for the newer approach (for me) 2010-08-09T13:53:45 There should be five rows with a date added of "4/23/2002" if I sort by date added, and then those five rows should still be at the top but shuffled around when subsequently sorted by date modified 2010-08-09T13:54:19 *** r123 has joined #evergreen 2010-08-09T13:54:30 dbs: yeah, it's the reverse. 2010-08-09T13:55:29 i guess i'll see if the new style grows on me. 2010-08-09T13:56:16 Ugh. Okay, I guess I can see how that works, but the new style seems clearer 2010-08-09T13:57:06 the other thing i'm trying to get used to is that the first one makes the table sortable. the second one populates a grid with the contents of your (presumably hidden) table. 2010-08-09T13:57:42 so i have to deal with things like height/width and/or iframe-y scrollbars. 2010-08-09T13:57:52 all important things to starting learning to deal with, though. 2010-08-09T14:24:33 phasefx: fetchOrgSettingDefault() - should that be visible to server-side XUL? I'm a-hoping so. 2010-08-09T14:29:19 dbs: if it's in opac utils, I think so 2010-08-09T14:29:28 org_utils 2010-08-09T14:30:08 if you have a staff client open, put cursor focus in a remote xul interface, hit control+shift+f7. then type fetchOrgSettingDefault in the textbox at the top and click Exec 2010-08-09T14:34:44 cool trick! 2010-08-09T14:36:12 seems to be there. thanks, one more hurdle cleared 2010-08-09T14:36:52 if we could get Firebug embedded in there, I'd be using it 2010-08-09T14:37:17 * phasefx broke things horribly last time he tried getting chromebug to work. gave up for a while 2010-08-09T14:38:50 strict barcode currently checks codabar checksum, with no org-unit setting for otherwise, right? 2010-08-09T14:38:54 i suppose i could look... 2010-08-09T14:38:55 * jeff does o 2010-08-09T14:38:56 so 2010-08-09T14:39:50 mck9++ # rocking the Perl 2010-08-09T14:49:23 * Dyrcona has found the secret sauce to load bibs in a hurry. 2010-08-09T14:49:51 ha. haha. When the debug dialogue says "Request is not defined", it means the Request object is not defined - not the specific request you're trying to make. Ha. Ha. 2010-08-09T14:50:30 Dyrcona: strip all tags but 100 and 245? 2010-08-09T14:51:06 phasefx: Batch SQL statements in Java. They're even faster than PgSQL's COPY command. 2010-08-09T14:51:23 mmmm 2010-08-09T14:51:57 COPY and \COPY ? 2010-08-09T14:52:26 dunno \COPY. 2010-08-09T14:53:30 damnit phasefx, I just caught myself looking up how chromebug would be installed in a xul app. <_< 2010-08-09T14:53:43 All I can say is that 420,000 bibs have loaded in the last two hours. It took about five hours to load that many with COPY doing 7 simultaneous process. (I'm using 7 threads in Java.) 2010-08-09T14:54:35 tsbere: by all means, continue :D 2010-08-09T14:54:41 No triggers or indexes disabled either. With COPY, I disabled some FTI indexes. 2010-08-09T14:55:04 * Dyrcona runs off to reimplement Evergreen in Java. :P 2010-08-09T14:55:25 tsbere: you may have better luck than me, and it's been a while 2010-08-09T14:55:43 Dyrcona: I bet OLE would love us then :) 2010-08-09T14:55:46 * tsbere notes that chromebug is more of a firebug ADDON than a standalone 2010-08-09T14:56:23 * phasefx was actually trying it with firefox and firebug, and not xulrunner.. with the extension version of the client 2010-08-09T14:56:31 and made my firefox unstable 2010-08-09T14:56:47 That.....would be a lot easier to get working, I think. :P 2010-08-09T14:57:21 combine that with self-upgrading extensions, and I'd certainly try to work like that on a regular basis 2010-08-09T14:57:41 Well, hey, I got ya a self-upgrading extension codeblock. :P 2010-08-09T14:58:06 going the xulrunner route, you'd probably need to stub out some entry points to make things look a bit like firefox 2010-08-09T14:58:17 tsbere++ 2010-08-09T14:59:25 I've been musing having xulG.new_tab, etc. work with the native tabbrowser in Firefox, rather than have the extension use a wholly separate chrome window 2010-08-09T15:07:19 I can't find it, but was there once documentation that pointed to what all the various abbreviations stood for in an OPAC search URL? (such as rt=title, tp=title) 2010-08-09T15:08:00 bshum: some crappy docs were worked out to try and support libx - search for libx in the wiki 2010-08-09T15:08:18 dbs: Thanks, I'll search for that. 2010-08-09T15:08:57 otherwise, see /openils/var/web/common/js/config.js for some source that defines those terms a bit further 2010-08-09T15:09:15 e.g. var PARAM_OFFSET = "o"; /* search offset */ 2010-08-09T15:10:12 phasefx: okay, given that fetchOrgSettingDefault uses Request and config.js constants, I'm guessing that I should put together my own network request instead of pulling those other pieces into volume_copy_creator.js? 2010-08-09T15:11:38 RemoteRequest.js should already be bundled via util_overlay.xul. Might be okay to include config.js as well; let me skim it for likely collisions 2010-08-09T15:13:33 yeah, config.js looks safe 2010-08-09T15:13:34 Or I could just unpack the one constant that it would want to suck in 2010-08-09T15:15:44 xulrunner should cache the file the first time it's loaded, so it shouldn't add much to existing interfaces if they don't need it 2010-08-09T15:17:40 right - but fetchOrgSettingDefault() calls the Request object, instead of the RemoteRequest object, and you don't get the Request alias for RemoteRequest unless you also load opac_utils.js 2010-08-09T15:18:06 it's a bit of a mess to pull in for one network call 2010-08-09T15:20:26 dbs: there's also fieldmapper.aou.fetchOrgSettingBatch(orgId, [set1, set2, ...]); 2010-08-09T15:20:36 which is present if dojo is loaded 2010-08-09T15:20:47 berick: right, but I was trying to go with the existing coding style 2010-08-09T15:20:54 util_overlay.xul is pulling in /opac/common/js/utils.js CGI.js md5.js RemoteRequest.js {locale}/OrgTree.js and org_utils.js 2010-08-09T15:21:07 no opac_utils.js 2010-08-09T15:21:23 ah 2010-08-09T15:21:59 phasefx: would you mind if I start mixing in more Dojo-based calls? 2010-08-09T15:22:18 I say suck in as much as the opac code as needed. replace existing conventions and utility functions with dojo where needed. code_duplication-- 2010-08-09T15:23:22 dbs: no, I think that'd be great, since Dojo has momentum with most of the other Evergreen developers, and it'd give me examples in my "home ground" to crib from 2010-08-09T15:24:07 okay, cool. Thanks phasefx for making me feel like I'm slightly less crazy, and thanks berick for being the voice of reason :) 2010-08-09T15:24:17 dbs: I've started using fieldmapper.standardRequest for some interfaces instead of util.network myself 2010-08-09T15:24:29 So nice and clean :) 2010-08-09T15:24:58 only thing I miss is logging, but I'm sure there's some easy way to hook that in 2010-08-09T15:25:50 So, I think I have chromebug working WITH xulrunner. But can't fully test, as I can't log into our test system. 2010-08-09T15:26:04 berick and I re-implemented the same wheels a lot. Be nice to see dojo as a common ground 2010-08-09T15:26:40 tsbere: I have a trunk server running you can get to 2010-08-09T15:26:59 http://evergreen-ils.org/dokuwiki/doku.php?id=community_servers#community_development_servers 2010-08-09T15:27:07 last updated end of June, har :) 2010-08-09T15:27:59 * phasefx needs to spend some effort making it upgrade itself. Which will be nice with the self-upgrading staff clients 2010-08-09T15:30:26 I think that is out of date 2010-08-09T15:30:30 Or at least the client is 2010-08-09T15:33:39 *** Dyrcona has quit IRC 2010-08-09T15:33:40 some of the menu options for acq and/or booking may have changed since then, but for the most part a recent client should work as long as it's looking for /xul/current/server/, /xul/acq_20100630/server/, etc 2010-08-09T15:33:42 * tsbere pokes around main.xul with chromebug anyway 2010-08-09T15:34:36 * phasefx always builds client with "current" during development 2010-08-09T15:34:49 but would hate for such a thing to be seen in the wild 2010-08-09T15:35:20 I found a bad cert override file in my profile, was preventing the client from acting properly. Seems to work all of a sudden. 2010-08-09T15:38:25 * tsbere thinks that getting chromebug working was too easy 2010-08-09T15:38:36 * tsbere doesn't use it often enough to test it, though 2010-08-09T15:39:22 it was very experimental at the time I tried it 2010-08-09T15:39:42 The firebug parts aren't working, though. I probably shouldn't be surprised. 2010-08-09T15:39:59 * phasefx uses venkman sometimes, but firebug is nicer 2010-08-09T15:40:44 firebug can't find anything to latch onto. Chromebug still works without that, apparently. 2010-08-09T15:41:57 I can poke around the DOM, the HTML, the CSS, etc on the chrome end though. 2010-08-09T15:43:32 * phasefx figured chromebug would just be a way of getting firebug to work with chrome 2010-08-09T15:44:12 No, chromebug was designed for debugging outside of what firebug can see. Normally you put chromebug on firefox. <_< 2010-08-09T15:44:34 can it catch network activity like firebug? 2010-08-09T15:45:02 * jeff schedules a report two days in the past and watches it run twice 2010-08-09T15:45:07 (amused) 2010-08-09T15:46:44 If it can....I don't know how to see it doing so. The net tab is empty. 2010-08-09T15:47:53 I think the limitation is that it is a wrapper to move firebug up a few levels, effectively. And firebug doesn't work with xulrunner normally. :( 2010-08-09T15:48:05 jeff: ammused, but not surprised, right? 2010-08-09T15:48:14 nope :) 2010-08-09T15:48:35 which is a good feeling 2010-08-09T15:49:01 * miker_ scheduled an hourly recurring a year in the past as a demo just to show that yes, it really will do what you think 2010-08-09T15:49:33 kill the server? :P 2010-08-09T15:49:35 clicking on the output folder over and over to reload ... /that/ was amusing ;) 2010-08-09T15:49:44 yes :) 2010-08-09T15:49:48 tsbere: nah, just ran lots of reports in a row 2010-08-09T15:50:10 need to teach my output-grabber how to delete output once fetched 2010-08-09T15:51:01 jeff: if it's authenticated (the grabber), there are APIs (used to build the report output table) ... are you using those? 2010-08-09T15:51:31 I recall "parsing email" being typed in here before, which is why I ask 2010-08-09T15:54:33 i parse the email to get the output url, from that i know the output id and i can make the open-ils.reporter.schedule.delete call. 2010-08-09T15:55:39 i have a ses cookie from logging into the oilsproxy to download the output, so i'd just re-use that value. 2010-08-09T15:59:56 * tsbere gives up on chromebug in xulrunner, it apparently isn't useful enough there. 2010-08-09T16:01:13 miker_: did that answer your question? 2010-08-09T16:02:37 whenever miker asks a question i try to make sure i'm not missing the lesson that may be contained within. ;) 2010-08-09T16:04:41 *** shopkins has quit IRC 2010-08-09T16:06:29 tsbere: there was a guy on here before, emrikol, I believe, that was producing an extension for the staff client. Did you see that? nifty stuff :) 2010-08-09T16:06:55 heh. Didn't see that, but could be useful, right? 2010-08-09T16:06:55 used the Extension Manager in xulrunner 2010-08-09T16:07:52 he actually hit upon some of the same issues you did, with version numbers in application.ini, etc. 2010-08-09T16:08:30 I imagine he would have 2010-08-09T16:08:45 Some of the version functions in particular are apparently brutal <_< 2010-08-09T16:09:13 jeff: no missing lesson here ;) 2010-08-09T16:11:25 I think my code does a decent enough job with the version numbers, though. Not sure I like depending on the README file though. >_> 2010-08-09T16:13:16 miker_: yay! 2010-08-09T16:13:41 *** collum has left #evergreen 2010-08-09T16:18:41 btw phasefx, you taken a closer look at the patch yet for major issues I somehow missed? 2010-08-09T16:19:29 tsbere: I skimmed it all, but haven't tested anything. really need to do that after hours 2010-08-09T16:19:35 grabbing 0365 2010-08-09T16:20:29 wonder if we should be adding NSIS and related utils to the pre-requisite fetcher, since packaging a client is optional 2010-08-09T16:20:48 should be or should not be 2010-08-09T16:20:57 I don't think you are currently 2010-08-09T16:20:59 The only other related utils would be zip and unzip. Everything else is fetched if needed. 2010-08-09T16:21:50 thought I saw something else 2010-08-09T16:22:28 mar? mbsdiff? They come together in a zip. 2010-08-09T16:22:47 Yes, the linux binaries are distributed in a zip. >_> 2010-08-09T16:24:01 yeah, mar and mbsdiff. not familiar with those. bzip2, that might already be in the prereqs 2010-08-09T16:24:18 bzip2 might just be in the os by default. 2010-08-09T16:24:53 mar is for making the .mar files (mozilla archive). mbsdiff is a binary diff that the mozilla updater understands. 2010-08-09T16:25:28 *** jamesrf has quit IRC 2010-08-09T16:26:06 dbs: arg! :) ... I'll take a later one, then ;) 2010-08-09T16:26:12 (go ahead, not done) 2010-08-09T16:26:33 har 2010-08-09T16:26:43 did I miss your 0365 lock? 2010-08-09T16:26:53 *** bshum has quit IRC 2010-08-09T16:33:45 Let's see how this commit / upgrade goes. 2010-08-09T16:33:48 * dbs crosses fingers 2010-08-09T16:34:09 dbs: need a tester? 2010-08-09T16:34:40 * berick upgrades his dev box obcessively 2010-08-09T16:34:41 I'll take a tester :) 2010-08-09T16:36:39 * dbs builds clean again 2010-08-09T16:36:45 dbs: no explosions 2010-08-09T16:39:00 *** yboston has quit IRC 2010-08-09T16:39:31 grabbing 0366 2010-08-09T16:42:25 dbs++ # for remember the cdbi bits! 2010-08-09T16:42:44 oh yeah, those too :) 2010-08-09T16:42:45 CRAP 2010-08-09T16:43:06 clean build explosion. no upgrade munging required thankfully. 2010-08-09T16:44:21 resolved in 17142 2010-08-09T16:55:06 Hmm. Got some more damage to repair apparently: "ERROR: column cn.record does not exist ... FROM asset.call_number_class cn" - somehow acn and acnc are mixed up in open-ils.storage.biblio.record_entry.copy_count.staff.atomic 2010-08-09T16:55:16 Think that goes back to Friday's commit though. 2010-08-09T16:56:00 dbs: luckily, I think I just made that obsolete 2010-08-09T16:56:06 heh 2010-08-09T16:56:16 asset::call_number->table( 'asset.call_number_class' ) 2010-08-09T16:56:17 fixing it won't hurt, mind you, in case we have to roll back /my/ changes 2010-08-09T16:56:35 I will proceed as planned, and look forward to your changes :) 2010-08-09T16:56:38 errant %s//, eh? 2010-08-09T16:56:58 changes committed already ... no more calling that method (or the metarecord version) 2010-08-09T16:57:45 yeah 2010-08-09T16:57:53 huzzah! 2010-08-09T16:59:08 looks nice, miker_++ 2010-08-09T17:00:10 not super DRY, but simpler and faster than building dynamic queries (in or out of the db) 2010-08-09T17:09:50 I love it when a plan comes together (re: call number classification scheme normalization stuff). 2010-08-09T17:10:48 I suppose we could avoid the fetchOrgSettingDefault call on every "Add volumes" invocation if we just grabbed the new OU setting at client initialization time along with the others. Ah well, premature optimization and all that. 2010-08-09T17:19:03 dbs++ 2010-08-09T17:33:03 *** dbs has quit IRC 2010-08-09T17:39:51 *** jamesrf has joined #evergreen 2010-08-09T17:46:42 *** jenny has left #evergreen 2010-08-09T17:52:30 phasefx: Do you think I should expand on the windows installer and add an option for the developer settings being installed as well (default off)? 2010-08-09T17:59:53 tsbere: i'd say go for it, if only because i want to see more examples of your approach. ;-) 2010-08-09T18:00:35 * tsbere wonders if that is a good thing or a bad thing ;) 2010-08-09T18:00:47 hah! a good thing. 2010-08-09T18:00:49 tsbere++ 2010-08-09T18:00:53 * jeff -> off 2010-08-09T18:40:55 *** youdonotexist has quit IRC 2010-08-09T19:31:58 *** jamesrf has quit IRC 2010-08-09T19:44:01 *** jamesrf has joined #evergreen 2010-08-09T20:49:29 *** jamesrf has quit IRC 2010-08-09T21:04:02 *** dbs has joined #evergreen 2010-08-09T21:04:09 *** dbs has joined #evergreen 2010-08-09T21:16:34 *** rjackson-isl has quit IRC 2010-08-09T21:17:58 *** rjackson-isl has joined #evergreen