IRC log for #koha, 2008-07-01

All times shown according to UTC.

Time S Nick Message
12:00 nengard slef - nope
12:00 nengard i'm not doing any development .. or know how to access the code etc
12:03 owen Hi everyone
12:03 owen welcome back, nengard
12:04 nengard thanks owen!
12:28 slef nengard: can you write the description or the help page?  then I'll take it today... else, please file a bug :(
12:29 nengard slef, i got kicked off - are we still talking about my sys pref issue or something else?
12:29 slef syspref... I was out to lunch :)
12:32 nengard slef - i added our notes to the documentation - you said you patched it - so what bug do I file?[…]Preferences--OPAC
12:35 slef nengard: blank help on http://intranet:8080/cgi-bin/k[…]bscriptionHistory perhaps?
12:37 nengard oh! I haven't been updating that help file ... in fact i have no idea how to do that (except on a local install).  I think we were trying to get the manual done and then were going to copy and paste the bits into the help ...
12:37 nengard this is something that should be addressed
12:37 slef nengard: oh! news to me :)
12:38 nengard slef the help on RC1 hasn't been changed at all
12:38 nengard at least not by me
13:27 nengard what is all this talk about sms - is this a feature in 3.0?
13:36 slef nengard: yes.  It's a feature that added in the last 2 weeks, requiring all 3.0 users to install more perl modules.
13:36 nengard slef - okay - i'll have to ask the guys what i need to add to the manual for this
13:36 slef nengard: I'm a little bit unhappy about major feature changes during beta/rc, in case you haven't noticed.
13:36 nengard :)
13:37 slef nengard: (it broke our two 3.0 test sites)
13:37 slef hdl, paul, are you awake?
13:37 paul seeing it's 3PM, fortunately, i am ;-)
13:37 paul hi slef& nengard
13:38 slef paul: any chance on fixing lists archives?  any help needed?
13:38 paul (mmm... I thought hdl solved the problem)
13:38 paul i'll ask him (he's away this week)
13:38 paul (and the next 2 weeks as well)
13:38 slef paul: would be really good to have working archives quicker than three weeks...
13:39 paul slef : do you know how I can run perl opac/ with a parameter (like q=something)
13:39 paul slef: of course !
13:39 paul that's why i'll ask hdl about the status of the ml
13:39 slef paul: run from where?
13:39 paul command line
13:39 paul (to do some profiling with -d:DProf)
13:40 slef ooh, not tried this for years
13:40 slef try export QUERY_STRING="q=something"
13:40 slef might need to set other CGI variables
13:41 paul it seems you're right :
13:41 paul Options must be name=>value pairs (odd number supplied) at /home/paul/ line 230
13:42 slef paul: have you set KOHA_CONF?
13:42 paul yes
13:42 paul export KOHA_CONF=/etc/koha-ipt3.xml
13:43 paul (which is correct & readable by my user)
13:43 slef should have a full list of CGI environment variables
13:43 slef
13:44 slef paul: other than that, check chris's website and search IRC logs - I'm sure this must be known already by someone else ;-)
13:45 paul time curl http... works like a charm. i've seen chris website. I wanted to investigate a little bit deeper
14:27 ricardo Hi all  :)
14:28 ricardo paul: are you there?
14:28 paul yep
14:29 ricardo Great!  :)
14:29 ricardo I don't remember if I have asked this question to you already:
14:30 ricardo Shouldn't the size of UNIMARC Leader field be 24 instead of 25 characters (in length)?
14:30 paul I think yes.
14:30 paul I always have problems with strings starting by position 0, or 1
14:31 ricardo Right... I asked this in IRC before (because I noticed that you did a commit changing a line from 24 to 25). HDL was here (in IRC) and told me that he thinks you changed it to 25 because of a problem with Zebra (?)
14:32 paul a problem with zebra ??? I don't remember that
14:33 ricardo I'll try to paste the IRC log (the relevant section) to here:
14:34 ricardo Creating tinyurl ...
14:34 ricardo paul:
14:36 ricardo paul: That's the result of a search in the IRC logs for the word "leader" (it may miss some sentences in-between that did NOT contain the word "leader")
14:40 paul ricardo : http://i15.bureau.paulpoulain.[…]uing/
14:40 paul (login : abel / abel)
14:41 paul is it correct now : Koha complaints only if the leader is not 24 or empty (if empty, he will be filled by calculated values only)
14:42 ricardo paul: logging in...
14:42 ricardo paul: Right  :)  Do you want me to submit a bug in then?
14:43 paul no, if it's OK now, i'll just commit the patched version
14:45 ricardo paul: Well, at least the "blur" seems to be working correctly (stays silent if the length is 24 characters or empty, but complains if it's, for instance, 25 characters)
14:45 ricardo paul: I did not try to create a COMPLETE record in the URL that you gave me, though
14:49 ricardo paul: Do you want me to create and save a complete record in or is that unnecessary?
14:49 paul you can do it, but i' think it's unnecessary. but feel free to do it anyway
14:52 ricardo paul: OK. Feel free to delete a book record with an author of Ricardo Marques (if I manage to create it, that is, eheh...)
14:54 ricardo paul: Interesting... Don't you use the subfield 200f ("First Statement of Responsibility") in your framework?
14:54 paul ???
14:54 paul yes I do, but it's not mandatory
14:56 ricardo paul: OK. It does NOT appear in your "Default" framework but it does appear in "Monographies"
14:57 paul monographies usually always have an author, where things like website does not
14:57 paul + it's a test/developper DB, it's not intended to be 100% consistent ;-)
15:02 ricardo paul: Right. Agreed (it's a test installation).  :)
15:05 ricardo paul: On the topic of Mailing List archives, I see by the IRC logs that slef already talked to you about this (and that you replied that hdl was on holidays but that you would try to reach him...)
15:05 ricardo I have a related question: I'm noticing that are still receiving new posts. Is this intentional?
15:05 paul I don't think so
15:08 ricardo paul: Right. Maybe a solution could be to "bounce" e-mails that are sent to, e.g., asking the sender to send it to ?
15:08 paul (on phone with hdl ;-) )
15:09 ricardo paul: Great!  :)
15:12 acmoore speaking of mailing list archives, markmail has started archiving the koha lists. see they will also apparently be loading the archives of these lists soon, too, so you will be able to search for older postings.
15:16 ricardo acmoore: Good to know that!  :)  Do you know whcih Koha Mailing lists will be archived by MarkMail?
15:16 ricardo s/whcih/which
15:17 acmoore ricardo, at least bugs, devel, and patches. I pointed them at all of the lists, and they said they started archiving them, but I don't see any of the "users" list posts being archived.
15:18 slef acmoore: boo! more non-free software on! (half-joking)
15:19 acmoore slef, what are you talking about?
15:19 slef acmoore:
15:19 acmoore slef, that's not on
15:19 slef acmoore: presumably you want a link from to
15:19 slef s/a link/links
15:19 acmoore slef, that's fine with me if you want.
15:20 slef On another subject, I just found the cause of some of my import_borrowers problems - C4::Members::AddMember doesn't auto-generate cardnumber.  Is there any reason why not?
15:20 ricardo acmoore: Right... I now see the names of those 3 lists in:
15:20 slef acmoore: do you know what's happening with plone on
15:21 acmoore ricardo, maybe I'll drop them a line and see if they'll add the users list, too.
15:21 acmoore slef, no, what's up with it?
15:22 ricardo acmoore:
15:22 acmoore slef, I don't know why it doesn't make cardnumbers. Is there a standard way to generate them?
15:22 acmoore slef, How are they currently generated?
15:23 slef acmoore: for import_borrower, the import file has to contain them at present
15:24 ricardo acmoore: Yeah, you may ask them to import the Koha list (at I'm partial no Nabble as an archiver, but I don't see anything wrong in having the lists archived by several archivers
15:24 ricardo s/no Nabble/to Nabble
15:24 slef acmoore: We were informed that will move to Plone[…]nr1199901668.html but I don't know the migration plan, so it chills work on the Kea site.
15:25 slef ricardo: do you have a full set of nabble koha list URLs handy?
15:25 slef ricardo: or a page listing all koha lists at nabble
15:26 ricardo slef: A full? Not really... But a search On nabble for Koha returns good results (in the last column). I can't paste things to here (I'm using irssi in an AJAX Console, and fixing variables of Firefox are not working for "fixing" clipboard security  :(
15:27 slef ricardo: Firefox 3?  I've got a fruity clipboard in that too :-(
15:27 acmoore slef, that sounds good. I don't know much about plone.
15:27 slef ricardo: is it
15:27 ricardo slef: Actually, still Firefox  (I'll switch to 3 when it reaches .0.0.14 too!  ;-)
15:27 ricardo slef: yes
15:27 paul ricardo & slef : there is a problem on koha ml archives that hdl don't understand. The links are wrongly builded every night, but correct every months. So, if you look may archives, it's OK. tomorrow, june archives will be OK as well.
15:27 ricardo slef: never mind... Nabble lists the koha lists here:
15:28 paul hdl will continue investigating those days, but he don't understand what's happening
15:28 ricardo slef:
15:28 ricardo paul: OK. And regarding bouncing mails sent to and/or redirecting them to ?
15:29 paul oups... I forgot to speak of that. I'll send a mail to him immediatly.
15:29 ricardo paul: OK
15:33 ricardo Am I still here? (I got disconnected from the console, had to login again, but the display is mostly black - apart from this message that I am typing now...)
15:34 slef ricardo: yes
15:34 ricardo slef: Great. Thanks for the feedback!  :)
15:38 slef OK, I've updated the - bug chris if it doesn't go live soon enough ;-)
15:39 ricardo slef: Can you update web content in Could you add a link to the download to Koha 3 RC1 in the Download page?
15:39 slef added markmail, updated gmane and nabble links to show all lists they have
15:40 slef ricardo: will do
15:40 ricardo slef: Great. Thanks!  :)
15:40 acmoore thanks, slef. How do you update that page?
15:41 slef acmoore: a clunky javascript web interface from katipo
15:42 hdl_laptop hello
15:42 ricardo Hi hdl! :) Sorry for interrupting your vacations!
15:42 slef ricardo: there seems to be a link there already, text "Download"
15:43 slef hdl_laptop: can anyone help fix the archives?
15:43 slef paul: any good reason for AddMember not to generate cardnumber if empty?
15:45 ricardo slef: Yes... but that page does NOT have a link to download RC1 yet (just some Beta)
15:45 acmoore slef, oh. Is that interface accessible by others?
15:47 paul ricardo: hdl is not in vacations. He's teaching Koha to some librarians (will be in vacation next week)
15:47 ricardo paul: Ah, sorry... I misunderstood then
15:48 slef ricardo: link is to[…]-stableRC1.tar.gz
15:48 slef acmoore: yes.  kados has access.  chris and russel are QA on it.  Not sure who else.
15:49 slef presumably rach and si have some access
15:49 ricardo slef: Right! :) And that link is in the home page but it's NOT in the download page
15:49 ricardo slef:
15:49 slef ricardo: link is from
15:50 slef ricardo: under "Koha Version 3.0 - June 23, 2008" and above "Older versions of Koha:"
15:50 ricardo slef: OK. Seems that I got a cached version of the page. Sorry about that
15:50 slef ricardo: go befriend your poxy-cache operator
15:51 ricardo slef: Maybe I should.. Oh well, F5 solved it
15:52 slef ricardo: I find a sharp stick works best.
15:53 ricardo slef: LOL!
15:54 ricardo hdl: Still here?
15:54 slef anyone know why AddMember should not generate cardnumber?  I'm also seeing a scary race condition in the web interface to add members... C4::Members::fixup_cardnumber says "Warning: The caller is responsible for locking the members table in write
15:54 slef mode, to avoid database corruption." but members/ calls fixup_cardnumber without locking anything.
15:56 hdl_laptop ricardo: yes
15:58 ricardo hdl: Would it be possible to bounce messages sent to koha list addresses and send messages to senders asking them to send then to the corresponding list in ?
15:59 hdl_laptop ricardo: yes.
15:59 hdl_laptop Could be but is there a reason why you want to do it ?
15:59 hdl_laptop Normally any message on should be resent on
15:59 slef hdl_laptop: lists still getting mail according to ricardo
16:00 slef ricardo: are mails being resent?
16:01 ricardo slef: I think at elast koha-translate is being "splitted" but I may be wrong. Let me check this again
16:04 slef ricardo: thanks.  koha-translate would be the worst to split, because some of its members aren't on any other koha development list.
16:07 ricardo slef: Well, the archives of koha-translate at only list 2 messages in June (and Nabble lists *lots* more sent to koha-trasnlate AT nongnu).
16:07 slef ricardo: the archives at are broken, remember?
16:07 slef hdl_laptop, ricardo: I think we need to fix that bug before we can confirm this one :-/
16:07 ricardo slef: Right... But is the problem only related to "No subject" messages?
16:07 ricardo slef: Agreed
16:08 slef ricardo: Not as I understand it.
16:08 slef ok, I've got to go play with propane for 30mins or so... back after
16:08 ricardo slef: And you are probably right
16:08 ricardo slef: Careful. Playing with propane may end with a bang (James Bond style! ;-)
16:10 hdl_laptop slef: as I understand, mailman has problems making archives of current month.
16:11 hdl_laptop the month has to be completed before Mailman makes a good archive of it.
16:12 hdl_laptop i may be wrong, some parameters I could have overlooked.
16:13 ricardo hdl: Assuming that's correct, at least the messagest that have been sent to koha-translate AT seem to not have been archived at :
16:13 ricardo Compare :
16:13 acmoore hdl_laptop, are you saying that this is a bug in mailman? That sounds a little odd, since we didn't seem to have this problem in previous months. And, other mailman installations like[…]-July/thread.html don't seem to have this problem.
16:15 ricardo hdl:[…]8-May/thread.html
16:16 ricardo hdl:[…]008-05/index.html
16:17 ricardo (not being able to copy + paste URLs to this AJAX console is really bad  :(
16:36 hdl_laptop ricardo: yes.
16:37 ricardo hdl_laptop: meaning...?
16:38 hdl_laptop I have seen the problem.
16:38 hdl_laptop I will try and synch those lists.
16:38 ricardo hdl_laptop: Ah, OK. Thanks for the feedback!  :)
16:38 hdl_laptop And get some more information on how to get archive work at best.
16:39 ricardo hdl_laptop: Great!  :)
16:52 ricardo Got to go. Take care people!
17:31 hdl_laptop gotta grab some food
18:06 ryan slef: around ?
18:18 owen If I have a change to the template that would result in a translation change, do I have to wait to submit it until we have a post-3.0 branch set up?
18:22 slef ryan: back now
18:22 slef that propane tank was entirely too much fun!
18:22 slef (yes, that was sarcasm)
18:56 slef oops
18:56 slef[…]w_bug.cgi?id=2288
19:02 acmoore slef, have you been able to make a test case that shows the race condition, or is that just from reading the cod?
19:02 acmoore s/cod/code/
19:03 slef acmoore: reading the cod, but it seems pretty obvious - if something else calls AddMember before whatever just called fixup_cardnumber does, it'll have the same "unique" number
19:04 slef acmoore: process 1 fixup_cardnumber, process 2 fixup_cardnumber, process 2 AddMember, process 1 AddMember -> FAIL
19:04 slef acmoore: improbable but possible
19:04 slef now, I'm really off to cook with gas
19:04 slef any messages won't get an answer for an hour or more
19:04 acmoore slef, yeah, it looks like it can happen to me, too. I was just wondering if it had actually bit you.
19:05 slef acmoore: not yet.  Just think if someone uses my upgraded import_borrower while someone else is doing a memberentry, it's a lot more likely to happen
19:06 slef so I'm making things worse, really, really ;-)
19:06 slef oh whoops
19:06 slef I wanted to invite comments on[…]w_bug.cgi?id=2287
19:06 slef but I pasted entirely the wrong browser window
19:06 slef very sorry everyone
19:07 slef these URLs are too damn similar - we need something like wordpress's permalinks with the title in the URL
20:47 slef well, the bbq works
20:47 chris :)
20:48 slef chris: update to notified
20:48 chris yep triggered it this morning before i caught the bus
20:48 slef chris: does it mail you those?  It seems to say it only mails russ
20:49 chris it emails russ, but i can spot changes easy
20:50 slef cool... I'll not prattle about it any more
20:50 slef Are you working on other fun stuff today?
20:51 chris bugfixing mostly i think
20:52 chris oh and some enhancements .. mostly bugs though
20:55 slef oh well
20:55 slef someone's got to squash them bugs
20:55 slef it can be satisfying when you get a real cockroach squashed
20:55 slef but the moths need to go too
20:55 chris :)
20:57 slef hdl_laptop: types of bugs
20:57 hdl_laptop ok.
20:57 hdl_laptop Thanks
20:58 hdl_laptop slef : on which subscription bugs are you working on ?
20:59 slef hdl_laptop: 1mo
21:00 slef hdl_laptop: 1536 is most immediate puzzle to me
21:03 hdl_laptop d should not occur.
21:03 hdl_laptop I had many serials without any items and the appeared.
21:04 hdl_laptop c : should we use HTML::Scrubber to parse input ?
21:08 slef hdl_laptop: searches for itype=SER... how does itype exist if there are no items yet?
21:10 hdl_laptop itype was for itemtype.
21:10 slef hdl_laptop: I thought itemtype was itemtype?
21:10 hdl_laptop But changed once again quite soon.
21:11 hdl_laptop (Joshua changed it just before RC1)
21:12 slef argh!
21:12 atz slef: itemtypes are at the item level, or biblio level or even biblioitems level
21:13 atz the items table field is "itype", but inside Koha everything should try to use the API so that it doesn't have to know or care what level sets the itemtype
21:14 slef atz: so how do you think searching for a new subscription biblio should work?
21:15 slef hdl_laptop: c. not sure... maybe should just check whether it is a valid item type?
21:15 atz not really my area...
21:15 atz i'd have to research for a bit
21:15 hdl_laptop slef:we should just add an itemtype selection box.
21:15 hdl_laptop And not hardcode SER or PER selection.
21:16 slef hdl_laptop: it is a selection box, but that's not to say that $evil_librarian won
21:16 slef 't edit the URL
21:16 slef I know $evil_librarian isn't a major concern for us yet, but still better to avoid if possible.
21:17 hdl_laptop who is $evil_librarian ?
21:18 slef someone trying to attack us through the intranet
21:26 hdl_laptop g nite.
23:35 Odd_Bloke Aha, have finally added a book to my library. \o/
23:37 Odd_Bloke The OPAC interface is really nice, BTW.
23:39 Odd_Bloke But is there anything I can do to make it easier for people on the admin side to add records?  ATM the 'Add MARC Record' screen is pretty terrifying. :p
23:41 masonj i think currently , the add-marc-record screen is the only way
23:41 masonj except for the bulk-edit process
23:41 Odd_Bloke Yeah, I'm looking at single-book entry.
23:42 Odd_Bloke What would creating a new framework entail?
23:42 masonj yep, i think thats it
23:43 masonj you can create a new framework, and just include the marc tags that are important to you
23:44 Odd_Bloke Hmm, I don't know enough to work out which those are. :p
23:45 Odd_Bloke How can I diagnose where a z39.50 search source is going wrong?
23:45 masonj look in your koha-error.log for starters
23:48 Odd_Bloke is the relevant error.
23:49 masonj ta, ill take a little peek...
23:50 Odd_Bloke Thanks. :)
23:55 Odd_Bloke OK, I may have got the wrong record type.
23:55 Odd_Bloke Let me try a couple and see if it helps.
23:55 slef Odd_Bloke: 3.0 or 2.2?
23:56 Odd_Bloke slef: 3.0.
23:56 slef Odd_Bloke: If you can patch,[…]play-a-little.txt makes the z39.50 search error a bit more useful / less crashy
23:57 slef patch path/to/cataloguing/ <0009-Improve-Z39.50-result-che​cking-and-display-a-little.txt
23:57 slef might work
23:59 Odd_Bloke Aha, "238: Record not available in requested syntax: Available syntax: SUTRS: Bib-1"
23:59 slef Odd_Bloke: British Library?
23:59 Odd_Bloke Yup.
23:59 slef I think you have to pay to get MARC21 out of them :(
23:59 Odd_Bloke /o\
00:00 slef +1
00:00 slef still worked last I knew
00:00 slef also has some useful ones
00:01 slef I think there's a better list on somewhere
00:03 slef includes syntaxes
00:07 masonj Odd_Bloke: in your, set $DEBUG to 1 too, and watch your error.log
00:11 slef masonj: I think my patch showed the problem... BL not offering MARC21 to the public interface.
00:12 Odd_Bloke Yeah, though Oxford is now giving me
00:12 masonj all good then ;)
00:13 masonj yow, thats a new error
00:15 slef oooh you know lamby?
00:15 slef </OT>
00:16 Odd_Bloke <OT> slef: Yeah, we're both at the University of Warwick. </OT>
00:17 slef Odd_Bloke: dang, maybe ox have got more strict
00:19 Odd_Bloke Never mind, a combination of Durham and LoC seems to be working for most things.
00:20 slef[…]k%3A210%2FADVANCE suggests not.
00:20 slef hmm
00:20 Odd_Bloke That said, they're showing BL as working...
00:25 slef Not as usmarc (aka marc21) record syntax
00:25 slef BL works - just not as anything useful ;-)
00:31 Odd_Bloke So the library that I'm managing (with massive and quite enjoyable overkill) is small enough that displaying all of the items contained within on a few pages would be both relatively pain-free and useful (to save users stabbing in the dark with searches).  Is there a good way to do this?
00:31 Odd_Bloke (Disclaimer: I've probably missed parts of the OPAC entirely, I've been doing backend stuff thus far)
00:33 slef probably virtual shelves but I've not used them in a while - spent too much time on cataloguing and patrons recently
00:34 Odd_Bloke slef: That's Lists, if my reading of something I found on Google is correct?
00:49 Odd_Bloke So I'm playing around with the OPAC with a couple of users now, and I don't seem to be able to put a hold on a book from the user's account.  Is this expected behaviour?
00:50 Odd_Bloke Furthermore, books are still listed as 'available' even when I've used a staff account to put the hold on them.  Is there any way I can get them to automatically switch to a held state?
00:51 Odd_Bloke Apologies for all the questions, I'm just quite excited that it's working. :)
01:07 slef Odd_Bloke: not to me but it might depend on sysprefs; I can't remember; no worries but you might get better answers either on-list or when Europeans are awake or some Americans aren't at a big conference.
01:11 Odd_Bloke As the library I'm setting this up for doesn't really have a 'checking out' system (the books are on shelves in a public area, it's trust-based), I figure people taking and releasing holds will serve as well as formal checkouts/ins.
01:28 Odd_Bloke Aha, found the problem: "AllowOnShelfHolds" was OFF.
01:37 Odd_Bloke OK, something funky with holds is going on here, clicking 'Holds Queue' in 'Home > Circulation' 404s.
04:10 Odd_Bloke The OpacBrowser and OpacCloud features look like they could be really useful to me.
07:13 paul chris around ?
07:15 chris yep
07:15 paul hello.
07:15 chris hi
07:16 paul can you explain how memcached work ?
07:18 chris
07:18 chris explains it better than I can
07:19 paul can you send me the memcached code you already did ?
07:21 paul + your mod_perl vhosts (mc is away this week)
07:21 paul (at rmll :
07:22 chris[…]inline/logs/cache
07:22 chris you can grab it from there, the cache branch
07:23 chris basically for
07:23 chris i cache the entire page in memcached
07:24 chris then if the user isnt logged in, the script just gets it from the cache
07:24 chris instead of connecting to the db, and building up the template etc
07:24 chris for other pages, you can instead just cache parts of it. Caching things like the config file
07:25 chris and fetching that object instead of rereading and parsing the xml multiple times in each script call makes sense too
07:25 chris email on its way
07:29 atz yeah, the main targets for caching afaict are the config files and sysprefs
07:29 chris also things like branches
07:29 chris we fetch the list of libraries/branches in a lot of places
07:30 atz and sessions would be much faster in RAM
07:30 chris so you cache them, and just change the admin scripts to rebuild the cache object when a branch is added/changed/deleted
07:30 chris ditto itemtypes etc
07:31 paul is memcached a standard package on linux distro ? what about windows ?
07:31 chris windows shmindows :)
07:32 paul memcached is in contrib repository of my mandriva box
07:32 chris paul: i have a syspref usecache
07:32 paul yes, i've seen
07:32 chris and using require, import
07:32 chris you can make it totally optional
07:32 Odd_Bloke It's in Etch, too.
07:32 chris night joe :)
07:32 paul sweet dreams atz
07:33 chris so my plan would be that if you want to run with memcached you can, but you dont have to
07:35 chris paul, because memcached is distributed .. you can also hit the cache locally, and talk back to central database server (if you want to have caches in your branches for example)
07:36 paul ???
07:37 chris if you have an opac server in your branch for eg, that uses the central database, but a local cache
07:37 chris its also good if you use a loadbalancer, because not matter which server your request ends up going to, because memcached is distributed, you can use the cache locally
07:38 paul hey, that's a nice idea to improve perfs !
07:39 chris pie probably has some good ideas too
08:14 chris cool email paul
08:15 lloyd Hey.. Anyone know if google jackets have been added to the staff interface yet?
08:21 chris not me
08:53 pie chris: anything specific about Memcached or just chatting in general
08:53 pie ?
08:54 paul hi pie. I think it's about optimization. (memcached or not)
08:54 pie yeah, as chris said, make it optional
08:54 pie but with it on, it's fly like the wind :-)
08:55 pie we have a setup in work whereby we have a stack of 6 memcached servers spread across the webservers
08:55 chris pie: are you on the koha-devel mailing list?
08:56 pie no, I should though :-)
08:56 chris
08:57 pie heh, the quick links on the homepage have instructions for chat, but not for mailing
08:57 pie :-)
08:57 pie ta
08:59 pie am on it now
09:01 chris pie:[…]2-to18211341.html
09:15 pie that RFC looks fine to me
09:16 pie I presume those tables aren't too huge
09:16 chris nope they are all pretty tiny
09:16 pie awesome
09:16 paul mmm... the syspref table has almost 200 lines.
09:16 chris the framework one would be the biggest
09:16 paul which is small, but cost a lot of CPU as it's loaded everytime.
09:16 paul chris: ++
09:17 paul (& i'm not 100% sure it's useful to memcache it, as it's not oftently used)
09:17 paul at least, it's not the 1st reference table to memcache ;-)
09:17 pie I guess the logic is slightly unfinished since you'd have to put the "$result=..." line somewhere after the end of the initial if :-)
09:17 pie but I'm just picking, that's all
09:18 pie those tables are usually just used as look-up tables then?
09:18 paul new mail "branching git for 3.0 / 3.2" sent on koha-devel...
09:18 chris yep, 99% of the time
09:18 paul yep.
09:18 chris mostly set and forget
09:18 pie I guess it's the low-hanging fruit and over time we might be able to put it into other places too
09:19 chris yeah, i think that we could cache a borrowers info
09:19 chris for say 5 mins
09:19 chris (name, address etc)
09:19 chris at the start of circulation
09:19 chris so we dont have to keep looking that up each book we issue
09:20 pie that sounds like a great idea
09:20 pie how much of their 'info' can they change via the same interface?
09:20 pie if anything
09:21 chris only the issue detail, ie what they have out, will change
09:21 chris the rest is all changed in the members section
09:22 chris which could flush the cache entry
09:22 pie sounds like a plan :-)
09:23 chris paul: id go with the one repo with a new branch
09:23 paul me too
09:23 chris then we can do cherry pick
09:24 chris and switch between 3.0 and 3.2 easily
10:09 slef hi all
10:11 slef chris: you're asleep like 80% of the time, in my perception :)
10:12 pie I think he's gone to sleep!
11:28 slef Does atz have a website?
11:45 slef is Apache test page.  Someone tell ricardo when he reappears.
11:52 owen "irsiDynix Enterprise unveiled as new faceted search product...featuring industry-first fuzzy search technology"
11:52 owen I wonder how they define "fuzzy" ...
11:52 slef owen: probably like - it autocorrected brand names into dictionary words last I tried.
11:52 slef (very very annoying)

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