Open Source Integrated Library System

Evergreen on IRC

#evergreen Logs for Wednesday, March 2nd, 2011

< Tuesday, March 1st, 2011Raw Log FileThursday, March 3rd, 2011 >
#TimeNickMessage
#00:33:11dbs has quit IRC
#00:42:57pmplett has quit IRC
#00:44:08pmplett has joined #evergreen
#01:36:10pmplett has quit IRC
#02:12:45moodaepo has quit IRC
#03:00:35jennam has quit IRC
#03:11:04jennam has joined #evergreen
#03:11:04jennam has joined #evergreen
#05:33:27phasefx has quit IRC
#05:34:50phasefx has joined #evergreen
#07:49:22rickd_ has joined #evergreen
#07:57:22rickd_ has quit IRC
#08:03:10kmlussier has joined #evergreen
#08:09:20sfortin has joined #evergreen
#08:17:19rickd_ has joined #evergreen
#08:31:50_bott_1 has quit IRC
#08:35:12_bott_ has joined #evergreen
#08:36:19dbs has joined #evergreen
#08:36:43dbsthanks for the tag fix, gmcharlt
#08:45:25dbs"the letters are in the right order, but they are not joined together" - knowing close to nothing about Arabic, is that a result of the font used to display the text?
#08:46:27tsbereI read that as "some of the letters are supposed to be combined for whatever reason, such as ae occasionally ends up being one 'letter'"
#08:46:48tsbereBut I don't know enough about arabic to be certain
#08:46:58dbstsbere: yes, that's what I'm saying
#08:47:00dbs peers at http://en.wikipedia.org/wiki/Arabic_alphabet#Ligatures
#08:47:35tsbereThing is, when I have seen ae combined (after being a and e) I have seen a unicode "joiner" between them.......seems a bit excessive to need that for a large portion of text.
#08:47:43dbswhen we used NFD, the display was much less pleasant for chars like é than NFC was
#08:52:50tsbereToo early for me to be trying to parse this, I think, but it looks like in Arabic things are supposed to look different depending on where they are in the word or sentence?
#08:53:27tsbereSo not joining words properly would result in it not rendering even close to correctly
#08:56:01tsbere would need examples of wrong compared to correct rendering just to consider starting working on a solution
#08:56:04dbstsbere: I think we lack sufficient information to have an intelligent conversation
#08:56:23dbswho wants to ask Christopher for said examples?
#08:57:07dbss/er//
#08:57:38tsbere lost the email in his tabs
#08:57:43tsberewas it sent to general or dev?
#08:58:00dbsgeneral
#08:58:04dbs needs to reboot
#08:58:06dbs has quit IRC
#08:58:58phasefx has quit IRC
#08:59:36phasefx has joined #evergreen
#09:01:36dbs has joined #evergreen
#09:01:36dbs has joined #evergreen
#09:04:15phasefx has quit IRC
#09:05:01Meliss has joined #evergreen
#09:05:07phasefx has joined #evergreen
#09:08:07tsbereI did some quick tests in firefox with some example words, and indeed the rendering changes (in firefox) depending on letter position. Insert whitespace in the middle and both sides of the whitespace change.
#09:08:23tsbereNot sure what triggers that yet, though
#09:08:44tsbereappears to be a standard utf-8 page
#09:08:59Callender has quit IRC
#09:13:01yboston has joined #evergreen
#09:16:50gdunbar has joined #evergreen
#09:18:16AaronZ-PLSQuick DB question: The view "payment_view" joins the money.payment table with "pg_class" to get the payment type (ie: forgive, cash, credit, etc). Where can I find this "pg_class"? I would like to do something similar as part of a different query, but having the payment type would be helpful
#09:20:43tsbereAaronZ-PLS: Tis a postgres thing, I believe
#09:20:51tsbereFrom a random version doc page: "The catalog pg_class catalogs tables and most everything else that has columns or is otherwise similar to a table."
#09:28:42AaronZ-PLStsbere:thanks
#09:29:56dbsyay, green across the top of http://testing.evergreen-ils.org/buildbot/waterfall
#09:36:48jenny has joined #evergreen
#09:37:35Callender has joined #evergreen
#09:46:50eeeviltsbere: re the rendering of some unicode character points seems to depend on normalization form of utf-8 (obv, when so encoded) ... for some scripts NFD works better and for others NFC is better ... and (at least in IE7 and before) that "betterness" seems to be reversed
#09:49:10dbsIf Christoph could provide us with some sample records, and some screenshots of the text contained within how it should look and how it currently looks, that would probably be helpful
#09:49:57dbseeevil: yeah, I mentioned the NFC vs. NFD thing to tsbere as well; NFC in Firefox was way the hell better for our purposes
#09:50:12dbsbut samples++ + tests++
#09:52:37eeevildbs: so you did. you++
#09:54:37r123 has joined #evergreen
#09:54:55r123 has left #evergreen
#10:11:24moodaepo has joined #evergreen
#10:31:35moodaepo has quit IRC
#10:34:01moodaepo has joined #evergreen
#10:43:59yboston has quit IRC
#10:49:41jenny has quit IRC
#10:49:59bshum has joined #evergreen
#10:50:46yboston has joined #evergreen
#10:59:53jenny has joined #evergreen
#11:28:08phasefxI've exposed a branch on evergreen-equinox.git called masslnc-vol-item-ui. The goal is to provide an option for unifying the volume/copy creator and item attribute editor, and to allow editing of parts and affixes
#11:29:18tsbereI may have to check that out
#11:29:51tsbereIf only due to being employed by a masslnc member
#11:29:58phasefxwould definitely appreciate the eyeballs
#11:30:15eeevilwell, the parts and affix UI parts will probably be in the bib_parts and cn_affix branches ...
#11:30:26eeevilfor the time being
#11:30:48collum has joined #evergreen
#11:31:08tsbereAnd now it may have become more work to look at than I have time for, and thinking about it I am not sure I could do looking at it justice.
#11:31:21tsbere can barely create items to begin with :(
#11:39:54phasebb has joined #evergreen
#11:58:14csharpI have a question for ldirector users...
#11:59:06csharpPINES is using ldirector to route to each of the brick heads, but looks to be using IPtables to route to our 2 SIP servers
#11:59:33csharpis there any reason why the SIP routing would be done with IPtables and not ldirector?
#12:00:02bshumcsharp: I planned to try ldirector routing with SIP (we only have one server atm)
#12:00:35bshumSo it's just a single entry pointing to our SIP server currently.
#12:00:43bshumNot sure what'll happen if we add additional ones
#12:01:32csharpbshum: thanks for the feedback
#12:02:12bshumcsharp: mrpeters-isl doesn't seem to be on, but he might have additional information. I know EI uses a bunch of SIP servers, and they also run ldirector
#12:03:39pmplett has joined #evergreen
#12:05:21jlamos has joined #evergreen
#12:05:28jenny has quit IRC
#12:06:12jlamos has left #evergreen
#12:07:17bshumcsharp: Don't know if you can trust it, but I found a doc where someone is making examples using SIP and ldirector: http://ur1.ca/3dfvb
#12:07:23bshumSo maybe it's fine?
#12:07:34csharpbshum: thanks!
#12:07:54bshumI'll be testing it soon enough.
#12:08:04bshumSo I'll let you know if we encounter any quirks on our end.
#12:09:30jlamos has joined #evergreen
#12:10:14chrissharp123 has joined #evergreen
#12:11:12jennam has quit IRC
#12:11:58bshumOr vice versa if you get there first ;)
#12:13:13csharpbshum: we're just looking to mirror the current setup, and asking questions along the way about whether X was set up for a particular reason or just because that's how someone chose to do it one day ;-)
#12:27:09phasefxtsbere: so masslnc-cataloging-enhancements is the real branch we'll work with going forward. It pulls in cn_affix, bib_parts, and masslnc-vol-item-ui
#12:28:26tsbere does a git fetch for his remote to get the new branch/updates
#12:28:56jenny has joined #evergreen
#12:30:49tsbereI must be getting used to git. I just asked git to tell me what actually changed...and was looking at the answer before I realized I had asked git that.
#12:38:29phasebb has quit IRC
#13:06:59tsberewell, looking at things, I would only have one suggestion so far
#13:07:14tsbereAnd it is really isn't that important
#13:07:15eeevilcsharp / bshum: the primary reason for not using ldirector for SIP is (when PINES' EG cluster was built) ldirector didn't have a way to add arbitrary tests for health -- it knew about http (and thus ping.txt) natively, and could check a local file, but had nothing that would actually test the SIP pipe. So, because it didn't do more than iptables, and iptables is faster, we went with that
#13:09:51tsberephasefx: Do you want my not really all that important thoughts?
#13:09:59tsbereor rather thought
#13:10:37phasefxtsbere: I love thoughts
#13:10:58bradleeevil: I also remember an issue where the SIP connections wouldn't die correctly and it would eventually kill the SIP servers
#13:11:02tsbereWell, in regards to the SQL queries in particular for call number prefix/suffix stuff
#13:11:11bradlberick: remember that? --^ (using ldirector for sip load balancing)
#13:11:19tsbereCurrently they look like cnp.label || CASE WHEN cnp.label <> '' THEN ' ' ELSE '' END || cn.label || CASE WHEN cns.label <> '' THEN ' ' ELSE '' END || cns.label
#13:11:49tsbereI think that they are slightly more readable, and possibly slightly faster, if they look like this: CASE WHEN cnp.label <> '' THEN cnp.label || ' ' ELSE '' END || cn.label || CASE WHEN cns.label <> '' THEN ' ' || cns.label ELSE '' END
#13:12:01eeevilbradl: ahhh... yes indeed. ld kept closed connections half open, and the sip server wouldn't notice that
#13:12:03tsbereAnd that is the extent of my thoughts so far.
#13:12:21bradlwe never solved why it did that... we just used iptables..
#13:12:36bradl10% wrong, 90% done, to quote Rogan ;)
#13:12:42eeevilheh
#13:12:59eeeviltsbere: that's a fine point, and thanks for the eyeballs
#13:15:29eeeviltsbere: actually, I'm going to use use a slightly different test, int >, which should be slightly faster still than a text <>
#13:15:37berickbradl: yes, that matches my recollection. the "up test" chewed up backends
#13:16:43bradlyeah, that's right
#13:17:42dbscollective_memory++
#13:18:08gmcharltcollective_memory_recorded_by_pinesol++
#13:18:54phasefxpinesol_on_old_community_server-- :)
#13:19:15eeevilso, the take-away is, "it'll work until it brings the whole thing to a screeching halt" ;)
#13:21:07tsbereeeevil: Actually, I take back my "only that" comment. I didn't notice the default for prefix, at least, because the insert into is asset.call_number_preffix instead in 040.schema.asset.sql - and for that matter the insert for suffix is using sufffix too.
#13:23:59dbs hopes to add a "run eg_db_config.pl" test to the CI server for Evergreen to catch schema errors
#13:24:14dbsor, you know, if anyone else wants to take a stab at editing buildbot.cfg...
#13:26:42tsberedbs: Potential issue is that on a dry run on a true fresh DB there are a number of error lines that can be safely ignored
#13:27:00bradlwe are the hive brain
#13:28:32eeeviltsbere: fixxing the exxxxtra fs
#13:29:59jenny has quit IRC
#13:30:25dbstsbere: yes, you can filter known non-problems
#13:32:54csharpphasefx: pinesol_on_lupin?
#13:34:13phasefxcsharp: need to put everything on lupin. maybe next weekend I can work on that
#13:34:30csharpphasefx: lemme know how I can help
#13:34:40phasefxwill do
#13:36:43dbsbecause everybody was dying for this, I put the current rules of governance draft into AsciiDoc format in http://svn.open-ils.org/trac/ILS-Contrib/browser/governance/governance.txt
#13:36:54gmcharltdbs++
#14:04:30bshumHmm, on the holds pull list page, there's a "Print Page" button on the upper right hand corner that doesn't seem to ever work.
#14:04:38bshumIt just claims "We are unable to Print or Print Preview this page."
#14:04:44bshumAnd might be a bit misleading
#14:04:53bshumThere is a print button in the lower left hand corner
#14:04:56bshumThat does work
#14:10:23dbs has quit IRC
#14:11:43phasefxbshum: we need to hide that button
#14:12:08bshumphasefx: Is that all one bar across the top? I see there's a reload, etc on the left side too
#14:12:11phasefxit's geared toward printing HTML pages, which the pull list no longer is in that context
#14:12:23bshumAha
#14:16:32kmlussier has quit IRC
#14:36:57bshumAlso, it seems that we cannot clone any of our existing A/T event definitions. It's bumping up the sequence but not actually doing anything.
#14:37:15bshumIn our log, the error reads: "null value in column "hook" violates not-null constraint"
#14:40:06kmlussier has joined #evergreen
#14:50:27dbs has joined #evergreen
#14:50:27dbs has joined #evergreen
#14:51:01dbskmlussier: http://masslnc.cwmars.org/ instead of http://masslnc.org ?
#14:51:58kmlussierIt redirects there, but I usually use masslnc.org. Is it not redirecting?
#14:54:11dbshmm. Didn't work earlier for me, but works now
#14:54:32dbsnever mind :)
#14:54:50tsbereI believe redirection errors are my department on that one. Unless someone at cwmars can make it just know what masslnc.org is so I can point an A record thataway.
#14:55:44pmplett has quit IRC
#14:55:54bshumSo, it seems that attempting to clone an A/T event definition loses the hook, reactor, validator field entries from the original.
#14:56:26tsberebshum: That sounds significantly less like a "clone" than it should be
#14:56:33moodaepotsbere++
#14:56:48bshumtsbere: Yeah I'm trying to see if maybe it's because we're attempting to clone our old 1.6 A/T definitions vs. newer 2.0 ones
#14:56:55bshumAnd if there's a significant difference between them
#14:57:16tsbereI would expect a clone to make as close to a dupe as possible in that context, regardless of rule age.
#14:57:34bshumRight, it used to work in 1.6
#14:57:41bshumBut hasn't been working in 2.0 for us
#14:58:42kmlussierdbs: Glad you made it in. Was starting to worry I would need to set out yet another correction on a URL. :-)
#14:58:57bshumOh weird, it works for some but not others.
#14:59:15bshumGoing to try to narrow down the differences
#15:01:32chrissharp123 has quit IRC
#15:01:44Andy__ has joined #evergreen
#15:02:08tsbereeeevil: In the readability improvements for the sql in biblio.pm, I think line 465 was intended to be removed (the old label build compared to the new that was put above it)
#15:03:03dbs has quit IRC
#15:04:51dbs has joined #evergreen
#15:05:24eeeviltsbere: that sounds like something I'd do... sec
#15:07:36eeevilthansks
#15:08:11jenny has joined #evergreen
#15:11:05pmplett has joined #evergreen
#15:39:32collum has quit IRC
#15:41:00AaronZ-PLSSpeaking on A/T event definitions. Is it possible for an A/T event to clear fines?
#15:42:52phasefxAaronZ-PLS: I imagine you can create a Reactor for that; but there's not a stock one already doing it that I see
#15:44:17pmplett has quit IRC
#15:44:25bshumAlright, so it's intermittent, some A/T we can clone, and others we can't. Nothing coming up different in the event_definition other than having different org units. And spontaneously, it seems that every so often, a definition we could clone at one point cannot be cloned a few moments later.
#15:45:02bshumPoking at the logs for clues, but so far, coming up empty
#15:45:08AaronZ-PLSWe have several libraries who give fine free status for a year to people that contribute $100+ to a capital fund, but they are not fine free at other libraries and some of them are abousing that status. I would like to create an event that voids the fines for those people if the checkout library is the library which has granted the fine free status
#15:45:23justin_mo has joined #evergreen
#15:46:30AaronZ-PLSIs there a better way to do this?
#15:47:15tsbereHow do you know they are fine-free status to begin with?
#15:48:03AaronZ-PLScurrently there is a stat cat. We are looking to add a new patron group "donor" to replace that stat cat
#15:48:41tsbereHow do you plan to differentiate where they are a donor?
#15:49:02AaronZ-PLSHome library
#15:49:08csharpAaronZ-PLS: we recently had that request in PINES, but it was for a fines-free week - we did not do it in the software, but I'm interested in being able to
#15:49:51csharpin that case, it was an entire library system within our consortium, and I figure it would be done with circ rules in 1.6-2.0
#15:49:53bshumOne of our libraries apparently had a fines free week too. Lucky there's an "amnesty" checkin mode in 2.0 that voids all fines off incoming items.
#15:49:55tsbereWell, when fallthrough comes around (2.1) you could rig it up so that the "donor" patron group when aligned with the correct user home ou gets a "no fines" recurring or max fine rule
#15:50:06csharpbshum: nice!
#15:50:12eeevilAaronZ-PLS: you could use a new bool staff-only user setting type as an opt-in setting for an A/T
#15:50:13bshum2.0++!
#15:51:27AaronZ-PLSSo the rule would be: "IF <PatronGroup> = Donor AND CircLib = HomeLib" void the bill with the message "Bill voided for Donor Patron"
#15:52:08AaronZ-PLSeeevil: how would that work?
#15:56:06AaronZ-PLSWould that be a new stat cat?, or is there another way to add a user setting?
#15:56:26Meliss has quit IRC
#16:07:23eeevilAaronZ-PLS: you can add user settings
#16:07:42eeevil(sorry ... attacking code ... shoulda just stayed quiet ;) )
#16:08:36eeevilAaronZ-PLS: but, as bshum said, having those users in a subgroup of Patron would be more straight forward
#16:09:31eeevilwell, I see bshum didn't say that, but mentioned amnesty
#16:09:31bshumtsbere actually, but yes!
#16:09:46eeevilright, tsbere
#16:09:50eeevilbshum++
#16:09:55eeeviltsbere++
#16:10:28bshumIs the action_trigger_filters.json.example an important thing to have a real file equivalent for in an /openils/conf folder?
#16:12:08justin_moHello all. I'm the IT guy for a library consortium in Missouri and we are considering implementing an open-source option for those libraries that might be interested. When researching options, Evergreen and Koha are the obvious front runners. Again and again I've seen it said (mostly by people who ended up going with Koha) that Evergreen is "more suited for large libraries and consortia" but haven't ever really seen any data to
#16:12:59phasefxjustin_mo: IRC will truncate long lines.. we didn't see what comes after "ever really seen any data to"
#16:13:13justin_moOh, sorry.
#16:13:18justin_mo...back this up.
#16:13:25justin_moThat's all you missed.
#16:14:05justin_moLooking for something to support the assertion that 'Evergreen is better for consortia'
#16:15:41phasefxit was designed to support consortia, but small libraries use it as well
#16:16:23phasefxit may be better to see what feature-sets are more closely aligned to what you're looking for when comparing these applications
#16:16:50gmcharltjustin_mo: a lot depends on who you are organizing the option you are presenting to your libraries
#16:17:09dbs has quit IRC
#16:17:19gmcharltif you're building (or have aims of buidling) a state wide resource-sharing system and shared catalog, that's what Evergreen was designed for
#16:18:02gmcharltif the libraries would be adopting individual open source catalogs, you might consider giving them option of picking either Evergreen or Koha based on features they need
#16:18:22gmcharltfor example, if you need EDI support out of the box, Evergreen currently has that but Koha does not
#16:18:27justin_mogmcharlt: We actually have a state-wide resource sharing system, but it's almost completely academic libraries, and all are running Millennium
#16:18:37gmcharltconversely, if you need LDAP integration out of the box, Koha currently has that but Evergreen does not
#16:19:01gmcharltof course, both projects I think will ultimately converge on "checkbox" features like that
#16:19:57justin_mogmcharlt: w/r/t giving them the option, I don't think we have the bandwidth (personnel wise) to handle two different systems. Does Evergreen have qualities that make it inherently more scalable, cheaper to develop, more performant, etc?
#16:20:17justin_moBasically I'm talking about "non-checkbox" features
#16:22:15dbs has joined #evergreen
#16:24:35gmcharltwell, in the context of a shared Evergreen catalog, it's explicitly designed to be more scalable, particularly in terms of making it easy to add additional application servers as transaction growth warrants
#16:25:23gmcharltin particular, it lends itself well to setting up a bunch of relatively cheap application servers
#16:25:32justin_mogmcharlt: That's exactly the type of thing I'm curious about.
#16:26:20gdunbar has quit IRC
#16:26:21gmcharltcheaper to develop? it's a wash. IMO, the barrier to entry to develop for Koha is lower, but Evergreen has a younger code base and has a bit less cruft
#16:26:38gmcharltKoha has a rather larger development community (and more commercial developers), though
#16:27:45justin_moFrom what I've seen the Koha devs are much more international.
#16:28:11phasefxincidentally, http://koha-community.org/ is the definitive site for that project, in case you've been elsewhere
#16:28:12justin_moAlso, despite the larger community of Koha, Evergreen seems to be adding new features more rapidly
#16:28:13gmcharltjust as a point of information, by the way, I'm a committer to Evergreen and the release manager of Koha 3.2.0, so if you're detecting that I'm firmly straddling the fence, you'd be correct :)
#16:28:23dbsHey, Evergreen has international! USA _and_ Canada!
#16:28:32justin_moAhaha
#16:28:34phasefxand elsewhere
#16:28:36gmcharltdbs: and Georgia the country!
#16:29:34dbsSorry, I was talking about code contributions. Definitely, I recall a trip to Armenia a few years ago
#16:29:51justin_mogmcharlt: tbh, someone straddling the fence is exactly what I need. We were put in the position to make a decision, so it really puts us on the same fence... except we have to eventually pick a side.
#16:30:03gmcharltas far as rate of adding features - also a wash; some things that are new to Evergreen 2.0 Koha has had for years; on the other hand, some things that Evergreen does Koha doesn't currently support
#16:30:43phasefxbut we all love open source over proprietary ;-)
#16:31:23justin_moproprietary is the ditch running on both sides of said fence I think
#16:31:26dbsKoha's next conference is in India, whereas Evergreen's is in Georgia (the state, not the country) :)
#16:32:21gmcharltjustin_mo: as far as talking to (relatively close-by) libraries are concerned, you can get a good Koha perspective from several consortia in Kansas
#16:32:29gmcharltand the Evergreen POV from Indiana
#16:33:58dbsgmcharlt and atz are probably the only people really qualified to talk about both koha and evergreen, so you lucked out justin_mo!
#16:34:16justin_moWe actually took a trip up to Michigan to visit MLC before the INCOLSA merge.
#16:34:26justin_modbs: I'm a lucky guy!
#16:34:45kmlussier has quit IRC
#16:36:29justin_moThanks for the start guys.
#16:42:29bshum has quit IRC
#17:00:00jenny has quit IRC
#17:06:36dbs has quit IRC
#17:34:51justin_mo has quit IRC
#17:37:50yboston has quit IRC
#17:47:31jenny has joined #evergreen
#17:59:14jenny has left #evergreen
#18:09:46dbwells_ has joined #evergreen
#18:13:33dbwells has quit IRC
#18:13:41dbwells_ is now known as dbwells
#18:23:14sfortin has quit IRC
#18:34:31rickd__ has joined #evergreen
#18:35:39rickd_ has quit IRC
#18:36:15atheos has quit IRC
#18:41:06atheos has joined #evergreen
#18:55:18r123 has joined #evergreen
#20:39:10rangi has quit IRC
#20:39:11mjgiarlo has quit IRC
#20:43:20rangi has joined #evergreen
#20:43:20mjgiarlo has joined #evergreen
#20:45:33rangi has quit IRC
#20:45:34mjgiarlo has quit IRC
#20:51:06rangi has joined #evergreen
#20:51:07mjgiarlo has joined #evergreen
#21:07:23phasefx has quit IRC
#21:09:52rickd__ has quit IRC
#21:19:17phasefx has joined #evergreen
#21:33:45dbs has joined #evergreen
#21:33:49dbs has joined #evergreen
#23:54:25pinesol has joined #evergreen
#23:54:25pinesol has joined #evergreen
#23:55:31phasefx2 has quit IRC
#23:56:09dave-esi has quit IRC
#23:56:10phasefx2 has joined #evergreen
#23:56:17dave-esi_ is now known as dave-esi
< Tuesday, March 1st, 2011Raw Log FileThursday, March 3rd, 2011 >