IRC log for #koha, 2009-04-08

All times shown according to UTC.

Time S Nick Message
15:22 pianohacker hi
15:27 owen Hi pianohacker
15:27 pianohacker Hello owen
15:28 pianohacker Eesh. I don't think holds have any other kind of problem
15:35 owen Seems like Koha isn't managing the holds priorities correctly.
15:35 owen ...but I'm not sure how to diagnose it.
15:37 kf own: item level hold vs. biblevel hold? wanted to test that out -never found time :(
15:37 owen biblio level hold
15:38 owen On the holds page I'm seeing things like, the top entry is marked as "waiting," and the next entries are "2" and "3." What happened to "1" ?
15:43 liz_nekls_ owen: we actually saw something simillar in our catalog
15:43 liz_nekls_ but we think it was staff error
15:43 liz_nekls_ accidentally placing item level holds instead of bib level holds
15:43 owen That's what I thought at first too
15:43 owen But I'm seeing it too often to believe that now
15:44 kf are your users allowed item level holds?
15:44 owen Not in the OPAC
15:44 liz_nekls_ same here
15:44 liz_nekls_ only staff can place item level holds
15:45 owen liz_nekls_: Do you let your staff change hold priority levels?
15:46 liz_nekls_ we do, but don't encourage it
15:46 liz_nekls_ we discovered that there are kind of infinite queues
15:47 liz_nekls_ what I mean is that if an item gets accidentally placed in item level, every item level that comes up behind that is in it's own queue
15:47 liz_nekls_ koha doesn't show this very well
15:47 liz_nekls_ it's logical, but the display ordering is weird at best
15:48 liz_nekls_ so say, you have a bib level hold for barcode 1
15:48 liz_nekls_ then the next hold is for barcode 2, item level
15:48 liz_nekls_ the third hold is for barcode 2
15:48 liz_nekls_ the fourth hold is itemlevel, no barcode assigned
15:48 liz_nekls_ er, i mean bib level
15:49 liz_nekls_ it's going to display that 1,2,3,4 in priority
15:49 liz_nekls_ when actually the person in priority 4 will probably get the book before the person listed in priority 3
15:49 liz_nekls_ because the item will be available sooner
15:50 liz_nekls_ it's confusing, we actually had a heated discussion one day hashing the display out.
15:50 owen Amazingly, I see what you mean :)
15:51 liz_nekls_ is this what you're seeing owen
15:51 liz_nekls_ ?
15:51 owen The confusing thing for me is that at some point koha "allocates" a particular barcode for a hold, and then changes the hold list display accordingly.
15:52 owen Maybe it's once the copy has been marked waiting, Koha says the hold is "only for 33000023994", making it look like someone placed an item-level hold
15:52 liz_nekls_ it may have something to do with this item vs biblevel hold deal idk
15:52 liz_nekls_ yes, that's what I think
15:52 liz_nekls_ er, inasmuch as I understand it
15:53 owen But we still see holds lists where, for instance, the priority numbers start at 2, there are 2 entries marked 27, two entries marked 30
15:53 kf liz: great analysis - i really have to test this later, we want to use item level holds for serials, where you need a special issue or volume on hold
15:54 liz_nekls_ owen: that's really weird, I don't think we've ever seen duplicate priority numbers
15:54 liz_nekls_ lemme ask though
15:54 liz_nekls_ 1s
15:55 owen Check your hold list for the Twilight DVD ;)
15:55 liz_nekls_ LOL
15:55 liz_nekls_ the title we actually hashed this out over was "beverly hills chihuahua"
15:55 liz_nekls_ LOL
15:56 liz_nekls_ i'm like... who would want that?!
15:56 liz_nekls_ j/k
15:56 liz_nekls_ yea, as it turns out we have seen that owen, we assumed it was a bug
15:56 liz_nekls_ ^.^
15:56 owen That's a good example for us to: The holds page in the staff client says 4 copies are waiting, an the priorities for the rest are 4, 5, 6, 7, & 7.
15:57 owen Two of the "waiting" entries say "Item being transfered to " with no branch, and list "No barcode."
15:58 owen If it's "waiting," it must have a barcode, since a single copy has been allocated.
15:58 liz_nekls_ yea, you would think that
15:59 atz there is a data problem once the item is "paired", there isn't a way to tell what type of hold it was.  it just looks like an item hold.
15:59 liz_nekls_ AH
15:59 atz so when staff manually reprioritize things, the schmo who was in the front waiting for a given item, will still be waiting *for that item*
16:00 atz which presumably just went out to somebody else...
16:01 atz i don't know about the transfer stuff...
16:02 liz_nekls_ is there a patch in the works for that atz?
16:02 liz_nekls_ or another tough nut to crack
16:02 atz the original holds design did not include ability for staff to reprioritize everything
16:02 atz i don't know who is working on that anymore
16:02 atz galen had looked at it months ago, iirc
16:03 liz_nekls_ yea nekls had some holds work done that might have affected that (local holds type stuff)
16:03 liz_nekls_ but maybe not, idk
16:04 atz yeah, that's an extra level of complexity
16:05 atz the simplest fix might just be to warn the staff member that the hold is being "changed" to be title-level when they go to reprioritize a "paired" hold
16:06 atz because that is most likely to get the right results most of the time... if they really did want the item though, then it causes problems.  
16:06 liz_nekls_ yea, or also to make the ability to reprioritize a syspref
16:06 liz_nekls_ and turn that crap off
16:06 liz_nekls_ >.>
16:06 owen We used to run a nightly script to reprioritize holds based on their original hold date, so that if the staff changed someone's priority it would get reset
16:06 atz yeah, i'm always like "why did you have to mess w/ that?"
16:07 liz_nekls_ owen: smart
16:07 liz_nekls_ sneaky lol
16:07 kf ill keep your solution in mind owen :)
16:07 atz the date isn't the factor though, since it would have to be a title level hold to get any new results
16:07 owen Yeah, but it was frustrating when you really *did* want to move someone on the list, for a good reason (like they got taken off by mistake)
16:08 atz ( for a paired item )
16:08 liz_nekls_ yea, I can feel that. Maybe it would be better to be able to modify the date? it'll be abused no matter what
16:10 liz_nekls_ or a permission to rearrange holds
16:10 liz_nekls_ kind of like the manager functions at the grocery store
16:10 liz_nekls_ you need a key and a code to rearrange
16:11 liz_nekls_ and put your initials on it and a reason
16:11 liz_nekls_ and show that on the list
16:11 liz_nekls_ that would deter for most of our libraries >.>
16:11 owen Someone ought to file a bug report for that ;)[…]w_bug.cgi?id=2970
16:12 owen FWIW.
16:12 liz_nekls_ lollerskates
16:12 liz_nekls_ this would be a definite argument against standard circ accounts in favor of staff member specific accounts
16:13 liz_nekls_ (we use generic staff accounts with tiered privileges, so all circ staff logs in as lib-circ, catalogers as lib-tech, directors as lib-director)
16:15 owen liz_nekls_: did you guys ever settle on a method for sending printed overdues?
16:18 liz_nekls_ don't think so, we had the first meeting about it last month
16:24 pianohacker liz_nekls_: Can I tell you about the solution I made for my library for a closely related problem?
16:25 pianohacker (Sending invoices for fines)
16:27 pianohacker Whenever a book is marked lost, a message is added to the message queue (of type 'print', rather than 'email' or 'sms')
16:28 pianohacker This print queue can then be dumped as a PDF or managed (deleting messages, marking them as sent, etc.)
16:31 pianohacker liz_nekls_: Does that sound like something your library might be interested in
16:31 atz pianohacker: good idea
16:34 kf bye #
16:35 kf bye #koha ... its really time to leave work ..
17:10 liz_nekls_ pianohacker: sorry I missed your messages, it certainly sounds like a good solution.
17:13 liz_nekls_ pianohacker: are you going to kohacon
17:16 pianohacker liz_nekls_: Yup
17:19 pianohacker There's always the 'attending' section of the Kohacon 2009 page on facebook
17:57 liz_nekls_ pianohacker:  genius (gah I get behind)
19:32 pianohacker Do any devs have thoughts on C4::Service?
22:42 chris heh
22:42 chris oh hey atz .. does clay look after the bugzilla?
22:43 atz the system/application anyway
22:43 chris its not long er
22:43 chris hmm too fast on the return key there
22:43 chris its no longer mailing the bugs mailing list
22:44 chris that was set as the qa for each bug .. looks like that option is gone
22:45 atz probably happened at the last upgrade... a few settings were reverted/overwritten
22:45 chris hean
22:45 chris gah i cant type
22:45 chris yeah even
22:46 chris i guess if we set it to default cc for all the components, that will get it back
22:47 atz chris: is there actually a "bugs mailing list" ?
22:47 chris yep
22:47 chris has been for years
22:47 atz addy?
22:47 chris
22:47 chris last mail was march 19 .. so taht tallys with the upgrade
22:48 atz i've got clay on another channel
22:48 chris cool
22:48 chris
22:48 chris if that could be the default cc for everything that would rock
22:51 atz chris: clay added addy as a "globalwatcher"
22:51 chris fab
22:51 atz should do it
22:51 chris thanks
23:19 chris hey jo
04:07 bignose-work chris: my backlog doesn't contain it, so I'm asking you again: what address should I send a report about the wiki "here's your new account" message never arriving?
05:01 brendan bummer - no nightly greeting from Amit tonight :(
05:01 chris ah yeah he is at a conference
05:01 brendan well good evening everyone!
05:01 chris bignose-work: the koha-devel mailing list
05:02 brendan Yeah - I remember.
05:02 brendan I guess I've become a fan of the nightly greeting...  haven't I
05:02 chris heh
05:55 mason good evening brendan, chris
06:15 SelfishMan I'm curious what hardware you guys are using for Koha and what is the average page load time for search results?
07:39 chris its around 2 seconds for my test box SelfishMan for page render, then up to 5 to pull all the google images in
07:40 chris its my home server, running pootle and a bunch of other things too
07:40 chris but its a decent box
07:42 chris dual core at 2ghz, with 4gig of ram
07:43 chris SelfishMan: you are using mod_expires eh?
07:57 chris serving the yui stuff locally and pushing the expires headers out for that, the images and the css makes the whole site a lot faster
07:58 SelfishMan yeah, using mod_expires
07:59 chris and serving the yui stuff locally or fetching it from yahoo?
08:07 chris hi nicomo
08:12 SelfishMan hmm...never thought about the yui stuff
08:12 chris depends on what your connection to the world is like
08:13 SelfishMan bad
08:13 chris from over here in nz, making people fetch stuff from the us on every page doesnt work :)
08:13 chris yeah you can set it in sys prefs to use local files
08:13 SelfishMan looks like all the yui stuff is local
08:14 chris cool
08:15 chris expires are set on it eh?
08:17 SelfishMan yep.  Everything except the actual page contents is set to expire in one month so that part is fine
08:18 chris cool
08:18 SelfishMan The staff are complaining about it taking 6-10 seconds to load a page so I'm trying to get a point of reference
08:19 SelfishMan The hardware that koha is running on is about four years old but it seems to perform as well as every other koha opac I've tested
08:19 soul9 SelfishMan, do you clean the sessions table some times?
08:20 soul9 we have caught sessions getting pretty long...
08:20 soul9 3 million rows :-D
08:21 soul9 that may slow mysql down a bit...i'm not sure though, depends on the mysql setup too, i guess.
08:21 SelfishMan soul9: I was looking at that earlier but I haven't had a chance to check it yet
08:21 chris at work, we do sessions in memcached
08:21 soul9 SelfishMan, sessions either uses an sql table, or uses a dir. in /tmp
08:21 soul9 chris, ++ :-)
08:22 soul9 chris, and how long do you keep the sessions?
08:23 chris 10 hours
08:23 soul9 yeah, that sounds good
08:23 soul9 chris, any motivation to share the specifics of the setup? ;-p
08:24 chris :)
08:24 chris 5 webservers, running nginx, memcached, apache2 and mod_perl2
08:25 chris in a load balancer
08:25 soul9 oh
08:25 soul9 heheh
08:25 chris with one server running postgres in the abck
08:25 soul9 nginx <3
08:25 chris (its for
08:25 soul9 youah
08:25 soul9 ah i see
08:25 chris nginx talks directly to memcached
08:25 soul9 big traffic eh
08:26 chris yeah round 500 gigagbytes a day
08:26 soul9 huh damn
08:26 soul9 yeah, that's pretty impressive
08:27 chris so the application (perl and template toolkit) caches pages in memcache and nginx serves them
08:27 soul9 i see
08:27 chris and fails through to apache2 (and the application) if the page is expired
08:27 soul9 and you use koha with mod_perl?
08:28 soul9 i thought there were issues with mod_perl + koha
08:28 chris ive been testing with it, there is unsafe code
08:28 chris just not cleaning up
08:29 chris and circular references, which perl doesnt clean up
08:29 chris so it leaks
08:29 chris but a couple of days clear and we could tidy that up
08:29 SelfishMan Yeah I found that too
08:30 chris theres the easy way and the right way :)
08:30 chris easy way is write an apachehandler that explicitly destroys things
08:31 chris right way, is track down the circular refs, and break them by either doing weak refs, or not making htem in the first place :)
08:36 soul9 hmmm
08:37 soul9 i see
08:37 chris they easy easy way
08:38 chris is using Apache2::SizeLimit
08:38 chris PerlCleanupHandler Apache2::SizeLimit
08:38 eiro Template::Toolkit + memcached + fcgi + nginx <3
08:39 eiro + lighttpd (die, die apache :-) )
08:40 SelfishMan I'm curious about that config
08:41 chris which config?
08:42 SelfishMan mostly the nginx with memcached config specifically for koha
08:43 chris well with koha, you could serve the static content with nginx
08:43 chris and fall thru to apache for the dynamic
08:44 chris there arent that many pages you can cache in their entirity with koha
08:45 SelfishMan That's what I figured but my mind shut down hours ago
08:46 chris but there are a few
08:46 chris the main page of the opac for one (if a user isnt logged in)
08:46 chris and a bunch of the staff site
08:47 chris for stuff we cache fragments of html we have a template toolkit plugin that can get an include from a cache, or render it and cache it
08:48 chris then we cache at the object level
08:48 soul9 you have all the koha templates in template::tolkit?
08:48 soul9 toolkit*
08:48 chris no this is for stuff, not koha
08:49 chris for koha, we can cache at the object/data structure level .. and maybe some pages
09:02 soul9 ah
09:02 soul9 yeah,i guess makes sense
09:07 chris i got a big win caching the output of _build_langauges_arrayref
09:07 slef is
09:07 slef    
09:07 slef is aklibrarian1 on here?
09:08 chris not currently no
09:08 slef what nick, do you know?
09:08 chris i dont think ive ever seen her on here
09:08 slef and if you see them when I'm not about, let them know their comments has gettextisms and an eyetest that stops me commenting and is offensively worded.  Got an email for her?
09:09 chris is she in the planet?
09:10 slef yep, but no email on it
09:10 slef about page currently reads "This is an example of a WordPress page, you could edit this..."
09:11 slef I guess I could dig through email archives to find a contact email
09:11 chris  <-- that one?
09:11 slef that one
09:11 slef this is for stuff ->
09:12 slef sorry, just reading scrollback
09:12 chris yep
09:12 chris my day job :)
09:12 slef my day job appears to be "which web service provider has just started using javascript to break our customer authenticated proxy"
09:13 chris heh
09:13 chris yeah mine is, 'which js ad is going to break the site this week'
09:15 chris actually its mostly just always adding new features to the editors interface
09:16 chris but fight with adservers is a regular hassle too
09:16 slef thing is, most of these providers don't offer anything except username/password (not feasible for 000s of patrons and 3 staff) and IP restriction
09:17 slef no openID, usually no shibboleth or athens
09:17 slef definitely not LDAP
09:18 slef so is there a better alternative than an authenticated proxy?
09:18 slef is the sort of design I find myself trying to cope with daily just now
09:19 chris heh
09:19 chris thats an entry for dailywtf if ever i saw one
09:21 chris
09:21 chris i showed you guys the code quality metric eh?
09:22 chris
09:25 slef 1mo, just flaming another captcha user
09:25 slef well, dripping sarcasm over them anyway
09:25 chris heh
09:26 slef heh @
09:26 slef sarcasm @[…]35.htm#msg3887821
09:27 slef what app's stuff built on and is it any good?
09:28 slef tire mulch resting underneath the White House playground equipment poses a <a href="">health risk</a> to Sasha and Malia
09:28 slef boggle
09:30 slef[…]LED-Clothing-WTF/
09:33 chris slef: custom built
09:33 slef chris: no framework at all?  for performance or because there wasn't a suitable one?
09:33 chris for performance
09:35 chris we do some tricks with views in postgres
09:36 chris and different schemas
09:36 chris so the frontends talk to the live schema only whihc has views that never show unpublished material
09:37 chris means the exact same code can be used on the preview machine as on the live ones ... with just a different db user
09:37 chris which talks to different views
10:01 chris right bedtime for me
10:04 kf sleep well chris
10:10 Amit hi kf
10:12 soul9 nite chris
10:13 kf hi Amit
10:22 slef nigh chris
10:23 slef night even
11:14 fallor hello all
11:17 soul9 hi
11:25 fallor i'm in trouble with koha 2.2 -> 3.0 upgrade, you wouldn't happen to know about it soul9?
11:25 Crusoe hello
11:25 fallor hello crusoe
11:26 Crusoe I'm experimenting with koha now and have a problem tying to save a record
11:27 Crusoe browser says Wide character in null operation at /usr/share/perl5/MARC/Charset/ line 96.
11:27 Crusoe in a log i see also something like: Use of uninitialized value $encoding in string eq at /home/www/koha-1/lib/C4/ line 3128.
11:28 Crusoe the latter preceeds the former in log
11:28 Crusoe any ideas&
11:28 Crusoe ?
11:29 Crusoe tying = trying
11:29 fallor somekind of character encoding problem obviously, but no idea what's going on
11:30 fallor i've been fighting with utf-8 characters and koha 2.2, but i've never run into that problem
11:30 fallor sorry :(
11:32 soul9 Crusoe, my guess would be your notice doesn't contain the charset.
11:32 Crusoe what is "notice"?
11:33 paul_p s/notice/biblio/
11:33 paul_p soul9: notice is french ;-)
11:33 soul9 :-)
11:33 soul9 Crusoe, record
11:35 Crusoe oh!
11:35 soul9 the record should specify what charset it's encoded in
11:35 Crusoe in a leader?
11:37 Crusoe ok, it is 100$a in unimarc
11:37 soul9 yeah
11:37 soul9 sorry, i don't have the marc specs around ;-)
11:38 soul9 and the cost too much :-((
11:38 soul9 s,the,they
11:40 fallor can anyone tell me what this is about:
11:40 fallor <yazgfs> isn't a variable assignment, skipping it<!-- [scheme:]host[:port][/databaseName] --> isn't a variable assignment, skipping it<!-- scheme: tcp, ssl, unix, http, sru --> isn't a variable assignment, skipping it<!-- can run all servers on tcp, but the unix socket is faster --> isn't a variable assignment, skipping it<listen id="biblioserver" >unix:/var/run/koha/zebradb/bibliosocket</listen>
11:40 fallor isn't a variable assignment, skipping it<listen id="authorityserver" >unix:/var/run/koha/zebrad​b/authoritysocket</listen> isn't a variable assignment, skipping it<!-- public server runs on tcp --> isn't a variable assignment, skipping it<!-- <listen id="publicserver" >tcp:@:9999</listen> --> isn't a variable assignment, skipping it
11:40 Crusoe I have set both mandatory character sets  to utf-8 - but it does not help :-(
11:40 fallor whoa, that was long
11:40 fallor sorry about that
11:41 Crusoe ... and nice to see you talking sed, guys %-)))
11:41 Crusoe s/talking/speaking/ %-))
11:41 fallor writing maybe ;)
11:42 fallor anyway, that's what says to me :(
11:42 mason fallor, i think you have a typo in your koha.xml file
11:43 fallor well it's created by the koha installer
11:43 soul9 Crusoe, ;-)
11:43 fallor i haven't touched it, but it looks like the script doesn't understand xml at all
11:43 mason looks like xml::simple doesnt like your conf??
11:44 mason where do you see the error?
11:44 fallor is there some tool that i can check the validity of xml?
11:44 hdl_laptop fallor: it seems you are trying to use a koha-config file for 3.0 with a 2.2 designed script
11:44 mason in a log , or from starting zebrasrv?
11:44 fallor where...? err... on my console if i try to run the script
11:45 fallor koha3 works fine, i can log in and do all kinds of stuff there
11:45 fallor it's just the search that doesn't work
11:45 mason export KOHA_CONF=???
11:45 fallor obviously because the index has not been created
11:45 fallor KOHA_CONF is ok, it's exported already
11:46 mason your KOHA_CONF is pointing to an incorrect/old version of koka.conf?
11:46 fallor it's pointing to the correct version afaik
11:46 fallor ito: /etc/koha/koha-conf.xml
11:46 fallor which is the file created by the koha3 installer
11:47 Crusoe wow!!!
11:47 Crusoe I've got it!!!
11:47 soul9 :-)
11:47 Crusoe marcflavour variable was not defined
11:47 Crusoe I just use "add variable" button
11:48 Crusoe s/use/used/
11:48 fallor that was simple enough then :D
11:48 mason fallor: use 'xmllint' to check your conf
11:51 fallor is there some dtd that i should check it against?
11:51 mason sorry , i didnt get that far :/
11:53 fallor just "xmllint --valid" prints the file on screen with no error messages
11:53 soul9 fallor, nah, i think it just should be well-formed xml
11:53 fallor so, it seems to be ok then
11:55 mason so...    $ perl -d ./
11:56 mason hit 'n'
11:56 nahuel chris, ?
11:56 nahuel chris, around ?
11:56 mason step thru the script unitl it barfs, which should be somewhere in xml::simple, i think?
11:57 fallor the same message repeats
11:57 fallor and then there is: skipping itmain::(./$|=1; # flushes output
11:58 fallor then i get this weird looking prompt that sayt DB<1>
11:58 fallor hitting enter says: main::(./ $limit;# = "LIMIT 100";

| Channels | #koha index | Today | | Search | Google Search | Plain-Text | plain, newest first | summary