Evergreen ILS Website

IRC log for #evergreen, 2014-07-25

| Channels | #evergreen index | Today | | Search | Google Search | Plain-Text | summary | Join Webchat

All times shown according to the server's local time.

Time Nick Message
00:25 jeff_ joined #evergreen
01:14 jeff___ joined #evergreen
01:52 jeff_ joined #evergreen
02:14 jeff_ joined #evergreen
03:10 jeff_ joined #evergreen
05:45 dreuther_ joined #evergreen
07:41 jboyer-isl joined #evergreen
08:02 mrpeters joined #evergreen
08:07 * csharp thinks we should have a channel bot renaming contest
08:08 csharp "pinesol_green" has never caught my fancy
08:22 akilsdonk joined #evergreen
08:22 jboyer-isl Eversol_Pinehurst_Esq
08:23 jboyer-isl Very distinguished.
08:36 jl- left #evergreen
08:40 mmorgan joined #evergreen
08:47 Shae joined #evergreen
08:49 tspindler joined #evergreen
08:55 bshum Heh, happy sys admin day!
09:00 jboyer-isl I got you a gift!
09:00 jboyer-isl It’s a server crash!
09:01 * mmorgan guesses sysadmins should be appreciated at least once a year ;-)
09:04 Dyrcona joined #evergreen
09:05 kmlussier joined #evergreen
09:05 * Dyrcona now has to remember how to use yaz-client to check something.
09:07 kmlussier Good morning!
09:07 kmlussier tspindler: Chrome Vox was never a problem with autosuggest.
09:08 tspindler kmlussier: ok, is there something free we can test it with then  we have it on ou a dev server
09:15 tspindler @coffeee
09:15 pinesol_green tspindler: Zoia knows how to make fusilli.
09:15 kmlussier Heh
09:16 kmlussier Shouldn't that say pinesol_green instead of zoia?
09:16 kmlussier @coffee tspindler
09:16 * pinesol_green brews and pours a cup of "Kona Sweet" 100% Kona, and sends it sliding down the bar to tspindler
09:16 mmorgan Can't make coffee when you haven't had coffee yet ...
09:16 kmlussier It was the extra "e"
09:17 tspindler kmlussier: well I gusse mmorgan is correct
09:17 tspindler can't type either
09:17 kmlussier LOL
09:18 kmlussier I'm okay with keeping pinesol_green's the name as is. But I would like it on record that she's female.
09:30 krvmga joined #evergreen
09:32 krvmga i didn't see anything on launchpad so i thought i'd ask if anyone had received reports of Snow Leopard Safari 5.1.10 users not staying logged in the opac after they checked the Stay logged in checkbox?
09:33 krvmga i was not able to duplicate the problem on Windows 7 Safari 5.1.7.
09:33 collum joined #evergreen
09:35 Dyrcona I swear someone filed a bug about Z39.50 and UTF8 before, but I can't find it on Launchpad.
09:39 RoganH joined #evergreen
09:39 Dyrcona Nor do I find what looks like a relevant email in my bugs folder, and it goes back to November of 2012.
09:42 krvmga Dyrcona: i hate when stuff like that happens
09:42 Dyrcona Well, I could be confused. I imagine it would have been filed by paxed or Eva (from the Czech Republic).
09:42 Dyrcona Eva has filed a lot of character set related bugs.
09:44 Dyrcona But, nothing about Z39.50... I swear there was a bug with a subject like "...search national library..." but I'm hallucinating.
09:45 mrpeters left #evergreen
09:45 Dyrcona Fix ALL THE BUGS!!!
09:57 jeff </endmeeting>
09:57 * jeff grins
10:00 Dyrcona o.O
10:20 RBecker joined #evergreen
10:31 akilsdonk joined #evergreen
10:47 csharp @praise the world's sys admins
10:47 * pinesol_green the world's sys admins is one of the few who deserves to be praised
10:47 csharp heh
10:48 kmlussier sys_admins++
10:48 kmlussier y'all make my life easier. :)
10:53 krvmga re TPAC: one of the things that has always puzzled me is why the div code containing homesearch.tt2 is in home.tt2. it seems to work just as well if you put the code in homsearch.tt2 instead and just let home.tt2 be a container.
10:55 bshum krvmga: That's been an ongoing restructuring that's happened as people play with the tt2 code.  I know I did some of that when we were poking at some of the elements for mobile catalog.  If we can move (or remove) divs, etc. to be in their proper places, I'm +1 to that.
10:56 * bshum hasn't looked at that one in particular though
10:56 bshum Submit a patch, maybe that'll be the future ;)
10:57 mrpeters joined #evergreen
10:58 krvmga my home.tt2 is only this now:
10:58 krvmga [%  PROCESS "opac/parts/header.tt2";
10:58 krvmga WRAPPER "opac/parts/base.tt2";
10:58 krvmga INCLUDE "opac/parts/topnav.tt2";
10:58 krvmga ctx.page_title = l("Home") %]
10:58 krvmga [% INCLUDE "opac/parts/searchbar.tt2" %]
10:58 krvmga [% INCLUDE "opac/parts/homesearch.tt2" %]
10:58 krvmga [% END %]
10:59 krvmga i'll submit a patch for home.tt2 and homesearch.tt2 then
11:01 jboyer-isl I’d even vote to just put the contents of homesearch.tt2 into home.tt2 and just have the 1 file. homesearch.tt2 is not referenced anywhere else. (that a grep -r homesearch.t2 * can see.)
11:01 jboyer-isl It also has nothing to do with searching, per se. (because searchbar.tt2...)
11:05 Wyuli joined #evergreen
11:06 dMiller__ joined #evergreen
11:11 krvmga jboyer-isl: i could make an argument for that.  i think my bias is toward a lego brick approach: a plain base that you can blocks on.
11:11 krvmga put your blocks on
11:12 jboyer-isl I suppose so. I guess I just like to see at least one “block” in each file. :)
11:13 jboyer-isl (meaning actual content, now I don’t know if the blocks are the html or the INCLUDEs in this metaphor. :) )
11:13 krvmga jboyer-isl: :)
11:14 krvmga periodically, people come to me with "why don't you put buttons for some canned searches in the white space?" or "why don't you have a carousel that highlights what our libraries are doing?" and other things.
11:16 krvmga homesearch.tt2, after all, is just a name for a file defining the white space. it could just as well be our-glourious-library-carousel.tt2
11:17 * krvmga has spelling influenced by quentin tarantino obviously
11:17 kmlussier I like what jeff's library does with the whitespace http://catalog.tadl.org/eg/opac/home
11:17 bmills joined #evergreen
11:18 krvmga kmlussier: yes, a good example of what i meant.
11:24 jeff two other approaches that we have been experimenting with can be seen here: https://dev.tadl.org/responsive-web/ and here: http://dev.kalkaskalibrary.org/books/
11:27 jeff both of those are very much experiments and/or work-in-progress, hopefully as hinted at by the "dev" hostnames, robots.txt, and content like: "* just testing some things" :-)
11:28 krvmga jeff: i particularly like the responsive-web one.
11:29 krvmga jeff: do you have many academic libraries in your consortium?
11:31 jeff krvmga: no academic libraries in our catalog. we're also not a "consortium", per se. we are a district library, plus we're contracting with a neighboring county library to provide ILS/website/etc services (going live in Aug).
11:35 krvmga jeff: what are you using to build the responsive-web one? or are you doing it from scratch?
11:38 Wiiyuli joined #evergreen
11:42 krvmga i'll be back later.
11:45 jeff @later tell krvmga the responsive-web one is pretty much from scratch (with some jQuery). if we decide to continue down that path we might incorporate angularjs. I'm also interested in eliminating/reducing the role/need of the Rails app that currently works as a JSON<->TPAC screen scraping gateway.
11:45 pinesol_green jeff: The operation succeeded.
11:46 jeff Since they're experiments, there are a lot of things in there where we've done it, but learned enough to know that we would want to do it differently next time. :-)
12:18 Dyrcona @marc 505
12:18 pinesol_green Dyrcona: The titles of separate works or parts of an item or the table of contents. The field may also contain statements of responsibility and volume numbers or other sequential designations. (Repeatable) [a,g,r,t,u,6,8]
12:25 Dyrcona Perl code like this always looks "wrong" to me, even when it is in my own code: substr($_, 26, 1) = 'i';
12:44 mtcarlson joined #evergreen
12:48 collum joined #evergreen
13:06 krvmga joined #evergreen
13:21 krvmga jeff: back to my question from before: for the responsive-web example, what are you using to build it? or are you doing it from scratch?
13:26 jeff krvmga: hopefully you got my @later from pinesol_green just now. let me know if you need me to repeat.
13:26 jeff krvmga: something i didn't include in the @later, so you probably didn't see:
13:26 jeff Since they're experiments, there are a lot of things in there where we've done it, but learned enough to know that we would want to do it differently next time. :-)
13:27 krvmga jeff: thanks, i missed that.
13:27 krvmga jeff++
13:27 jeff krvmga: you saw the bit pinesol_green relayed with mention of jQuery/angular/Rails/etc?
13:29 krvmga jeff: yes, i saw it now.
13:31 jeff great!
13:52 mllewellyn joined #evergreen
14:02 gmcharlt eeevil: bug 1348731
14:02 pinesol_green Launchpad bug 1348731 in Evergreen "add optional nonce to permit distinguishing multiple simultaneous auth using the same username" (affected: 1, heat: 6) [Medium,New] https://launchpad.net/bugs/1348731
14:03 eeevil gmcharlt++
14:08 jeffdavis re: the mention of carousels earlier, I found this to be useful reading: http://cynng.wordpress.com/2014/01​/30/death-to-the-website-carousel/ (the post title comes on stronger than the post itself does)
14:19 jeff jeffdavis: at this point, i'm not sure which to call a "slider" and which to call a "carousel"
14:21 jeff jeffdavis: we've been using "carousel" to refer to "multiple book jackets with titles beneath that when clicked lead to the catalog record page for that title, optionally scrolling/touch-drag-whatever to see more items"
14:33 kmlussier Ha ha. I was on vacation when bug 1339190 was submitted and didn't see that bug title.
14:33 pinesol_green Launchpad bug 1339190 in SIPServer "SIPServer is heavy, not my brother" (affected: 1, heat: 6) [Undecided,New] https://launchpad.net/bugs/1339190
14:33 kmlussier eeevil++
14:34 krvmga jeffdavis: just read death-to-the-website-carousel , interesting read
14:34 krvmga jeffdavis++
14:35 eeevil kmlussier: :) ... I have a followup to that that's in testing ... I'll pull the pullrequest for the ... nonce (that'll be funnier later)
14:36 jeffdavis krvmga: glad you found it worthwhile!
14:37 jeffdavis it seems that TADL's carousels/sliders/whatever avoid a lot of the accessibility issues that article raises, which is cool
14:37 jeff jeffdavis: my intent wasn't to defend them, just to comment on how i've used slider/carousel to mean different things over time. :-)
14:38 jeff and our main site still maintains a "big image that changes automatically" element, which is what's being critiqued as bad :-)
14:40 jeffdavis jeff: until recently my preferred term was "you know, that thingie that lets you scroll through cover images for a dozen or so different books" ;)
14:41 jeffdavis <- feral web developer
14:41 * jeff laughs
15:24 mrpeters left #evergreen
15:39 kmlussier @love spell check
15:39 pinesol_green kmlussier: The operation succeeded.  kmlussier loves spell check.
15:40 bshum Ugh
15:40 bshum mceraso++ # found an error in the postgres install for the alpha build
15:40 bshum Looks like a duplicate key issue of some kind....
15:41 bshum Breaks fresh installs :(
15:41 kmlussier mceraso++
15:42 bshum That's weird
15:42 bshum INSERT INTO config.upgrade_log (version, applied_to) VALUES ('2.7.0-alpha1', :eg_version);
15:42 bshum Ended up in the 002.schema.config.sql file twice
15:42 bshum Sigh
15:44 bshum Not sure what happened there
15:49 Dyrcona bshum was it once around line 94 and again around line 920?
15:49 bshum It was around 95/96
15:49 bshum NOt sure about again
15:49 bshum I can check further
15:50 bshum Nothing I can see further down in the script
15:50 Dyrcona The second one is in a function definition and it uses variables.
15:50 Dyrcona I was just thinking if the make release script messes with that file it might have duped it some how.
15:51 Dyrcona I also wanted to check it didn't get unnecessarily doubled in master.
15:51 bshum There isn't an entry in master for the version like that.
15:52 bshum Or at least not one I could see
15:53 bshum That is so weird...
15:53 bshum The branch I used to make the tarball has the upgrade script in it
15:53 bshum But not the actual tarball itself
15:54 bshum So that too is apparently missing from the alpha tarball
15:56 Dyrcona right. I thought maybe the make release script would scan the file for the existing insert and modify it.
15:56 Dyrcona I don't know if I've ever used make release, and if I did it was quite some time ago.
16:03 bshum Yeah
16:03 bshum I guess it's a quirk from running make_release more than once
16:03 bshum It keeps adding another line
16:03 bshum That must be what happened to me
16:03 bshum I'm resetting my branch and making a new build
16:03 bshum 2.7.0-alpha2!  I suppose.
16:04 dbwells I'd probably stick with alpha1, but it doesn't matter too much.
16:05 bshum Yeah I could just reload it again
16:05 dbwells Assuming we're just redoing the build, and not adding more stuff in.
16:05 bshum Nothing new has been committed yet, so it's basically the same thing
16:05 bshum Just with the proper scripts this time
16:07 bshum Can just push a new alpha1 up to the web server once this dance is done.
16:13 mmorgan just reading about clustering in postgres
16:13 mmorgan anyone using this on (just as a random example) permission.grp_tree to sort permission groups in the patron registration/edit form?
16:14 mmorgan did it on our training server and it worked like a charm
16:17 artunit joined #evergreen
16:18 phasefx mmorgan: hrmm..how are things sorting without that?  I would expect ascibetically at each level in the hiearchy
16:18 phasefx hierarchy, even
16:19 mmorgan seem to sort in the same order they are in the database, which is affected when rows are edited
16:20 mmorgan for reference, lp 816615
16:20 pinesol_green Launchpad bug 816615 in Evergreen "patron profile permission group sort order" (affected: 4, heat: 30) [Wishlist,Triaged] https://launchpad.net/bugs/816615
16:21 gmcharlt mmorgan: the cluster command is really meant just to improve query perfomance if you know that the primary mode of access for rows in table is  in order by a particular index
16:22 gmcharlt but it really shouldn't be counted on to preserve sort order for use by an application that's not using an 'order by' clause in an underlying SQL query
16:22 gmcharlt IOW, a better and permanent fix for 816615 would be to update the code to request the permission groups sorted by name
16:23 * phasefx is with gmcharlt.  Treat order as undefined if there is no order by
16:23 gmcharlt (or to do a sort after the staff interface retrieves the list)
16:24 mmorgan gmcharlt: I get that certainly, and would be thrilled with a permanent fix, but in a pinch, is there a downside to clustering that table? It certainly gives us a better list than we have now.
16:24 tspindler left #evergreen
16:25 mmorgan we don't edit that table often, but whenever we do, the order changes, and to our users it sure appears random now.
16:25 gmcharlt mmorgan: no particular downside, though since cluster takes an exclusive access lock on the table, the cluster command should be run (or re-run, if you add or edt pgt) after hours
16:29 mmorgan I would plan on running it only after an edit, which is rare, and thanks for the caution about the exclusive access lock.
16:30 eeevil mmorgan: the reason they're not sorted right now, btw, is that because it's a tree, name-sort adjacency doesn't imply proper ordering ... the display code should be sorting them at each "level" of the tree
16:32 eeevil if the code isn't doing that, but a sorted list has the side-effect of proper sorting in tree mode, that's fine ... but there's not guarantee that every JS implementation, say, will display that same side effect
16:32 eeevil so, once we move to the browser client, some browsers may not have that side effect you want ... just as a warning
16:33 eeevil (or they all may, or the code may be preserving enough order that it's more than a side effect, but I haven't looked to be certain)
16:34 eeevil point is... with SQL, an order-by is the only sure way to get a specific ordering (or: what gmcharlt said :) )
16:36 mtcarlson joined #evergreen
16:37 mmorgan ... and I get that, certainly, but it doesn't do it now, and if there's something we can do locally to make it better for our users without causing any additional side effects, it works for me in the short term.
16:40 phasefx mmorgan: I think we're just slightly horrified but there's really no short-term harm :)  I used to be a huge fan of quick & dirty
16:40 mmorgan I'm not going to do it on Friday afternoon before going home for the weekend, though ;-)
16:40 phasefx mmorgan++
16:42 eeevil mmorgan: to be certain, I'm not saying don't cluster ... just want to make sure the caveats are clear.
16:43 eeevil (and I'm being verbose for the logs as much as anything)
16:43 mmorgan phasefx: why is it slightly horrifying? Maybe I'm missing something, but this seems minor. Not mucking around in the software, or much in the database either, just clustering one small table when we make an edit.
16:44 mmorgan yes, I certainly agree that all the caveats should be stated clearly.
16:44 eeevil mmorgan: I think phasefx means "horrified that the front end doesn't sort the names" :)
16:44 phasefx mmorgan: because it's a kludge, relying on undefined implementation specific behavior
16:44 paxed Dyrcona: i doubt it was me, i don't think we had any finnish z39.50 servers in use at that time.
16:44 mmorgan eeevil: Ah. Yes, we had that in week one after going live :)
16:45 phasefx unless clustering (which I'm not really familiar with) is supposed to define order.. and then it's still not ideal because the UI should be doing it out of the box
16:45 Dyrcona paxed: I hallucinated the ticket that I was looking for. It doesn't exist.
16:45 paxed Dyrcona: ah, ok.
16:45 * bshum waves at paxed
16:45 kmlussier Good morning paxed!
16:46 kmlussier Oh, wait, it's not morning there yet, is it?
16:46 eeevil phasefx: it physically orders the contents of the table based on an index, so that similar rows land on the same or nearby pages, and there's less seeking in the heap file
16:46 paxed kmlussier: late evening
16:46 * paxed is at a 'con right now.
16:47 phasefx eeevil: but the SQL standard itself doesn't specify what order rows are returned in without an order by, right?
16:47 eeevil phasefx: correct
16:47 phasefx it's just an implementation quirk of postgres that could change
16:47 eeevil cluster is about access speed
16:47 eeevil and it's pg specific
16:47 mmorgan phasefx: but it does seem to be a fairly benign kludge, compared to some others...
16:47 phasefx mmorgan: I agree :)
16:47 eeevil mmorgan: indeed
16:48 phasefx thus the "slgihtly" :)
16:48 phasefx hrmm.  I'll blame encrypted synergy for that typo :D
16:48 eeevil phasefx: but, yeah, in SQL, a relation is a "bag" (or a "set" if it has a pkey), not a list, which implies order
16:49 phasefx eeevil: makes sense to me
16:49 gmcharlt it would be interesting... for certain definitions of that word... to add a switch to Pg to randomize the order of rows returned by a query that doesn't specify an explicit order
16:49 Dyrcona sybase has the notion of clustered indexes, which sounds similar in purpose, but possibly different in implementation.
16:50 eeevil gmcharlt: I think I saw that proposed, actually, to remove side-channel info leakage
16:50 phasefx gmcharlt: we can go ahead and use cluster on the QA server to put everything in backwards orders
16:50 mmorgan In our experience on Evergreen thus far, the permission group list in the edit screen has been badly sorted, and pretty consistently so from day one. If clustering can sort it better even for a short time, it's a win for us :)
16:50 eeevil back when all the selinux extensions were first being designed
16:50 phasefx gmcharlt: well, s/QA/demo/ or manual test server
16:51 phasefx not quite the same, but partway there for finding bad assumptions in the code
16:52 eeevil Dyrcona: that's, IIRC, index organized tables, where the heap tuples live on the leaf pages of the index ... yes, similare purpose, but not possible in pg in the broadest sense
16:52 eeevil Dyrcona: however, for stable dataset and a covering index, you can get an index-only scan
16:52 Dyrcona yep.
17:06 pinesol_green Incoming from qatests: Test Success - http://testing.evergreen-ils.org/~live/test.html <http://testing.evergreen-ils.org/~live/test.html>
17:11 mmorgan good weekend all! don't have nightmares about clustering...
17:11 mmorgan left #evergreen
17:11 kmlussier mmorgan: Have a nice weekend!
17:11 kmlussier Ah, too late!
17:17 * jeff stops pulling at puppet's strings and heads home
17:32 * kmlussier really shouldn't write e-mails on a Friday afternoon.
17:35 bshum kmlussier++
17:35 bshum cookies++
17:35 kmlussier I'm wondering if I can get custom-made cookie boxes with the MassLNC logo.
17:35 * kmlussier checks
17:50 gmcharlt kmlussier++
17:50 * gmcharlt makes note to self: kmlussier & jcamins cookie bake-off whenever there's a joint EG/Koha conference
17:51 kmlussier I'm looking forward to it!
17:51 kmlussier I still haven't figure out the whole packing cookies thing, but I figured I could consult with jcamins when the time comes.
19:05 dMiller_ joined #evergreen
19:22 rangi` joined #evergreen
19:34 jboyer-isl joined #evergreen
19:53 kmlussier left #evergreen

| Channels | #evergreen index | Today | | Search | Google Search | Plain-Text | summary | Join Webchat