Time  Nick         Message
00:37 Waylon       hiya all!
00:38 Waylon       koha 3.0presomething. How easy is it going to be to speed things up slightly, using memcached?
00:38 Waylon       And what about nginx? how do I use that with koha
00:38 Waylon       ?
00:39 eythian      3.6 supports memcached fairly easily I think. I doubt 3.0 does.
00:39 Waylon       during a siege using 15 concurrent, im getting returns of up to 16 seconds... thats.. bad.
00:39 eythian      I don't think you'd get much gain from nginx (though you might.) It'd still be a CGI app though.
00:40 druthb       o/
00:40 eythian      Waylon: but a slow query watcher on your database, maybe you've got something that's taking too long there, or IO troubles.
00:40 eythian      *put
00:40 eythian      perhaps a missing index somewhere.
00:40 Waylon       so, if i want speed, im going to have to upgrade koha?
00:40 eythian      it'd be the easiest way, yes.
00:41 Waylon       Longest transaction:	       17.45
00:41 eythian      You could always backport the memcached etc. patches, but it'd be a fair bit of work.
00:41 Waylon       Response time:		        6.38 secs
00:41 eythian      My money is on it spending that time in the database.
00:41 Waylon       right
00:42 eythian      I think the option you should turn on is slow-query-log, or something similr
00:42 Waylon       right
00:42 eythian      it'll make mysql dump out anything that takes longer than /n/ seconds
00:42 eythian      (or ms perhaps)
00:42 eythian      which helps figuring out where problems are.
00:43 rangi        thats 3.0 pre .. 3.0.0 was releases 10 aug 2008 ... so its even older than that so trying to backport to that will be a world of pain
00:44 Waylon       yeah. hell.. i tried doing a git review on this install.. i couldn't find a version match. its a mismash of pre and manually applied patches that I can't figure out, and the arabic mods.
00:45 Waylon       so... yeah.. backporting memcached patches... not going to happen.
00:47 eythian      I'd recommend upgrading, but in the short term check out your database.
00:48 Waylon       and getting nginx to replace apache, not going to be easy either?
00:48 eythian      Probably easier, but apache isn't adding 16 seconds of delay to your requests.
00:50 eythian      (unless something else is, and your apache maxchilds is set low)
01:04 Waylon       hmm.. slow queries are inserts into sessions, and a borrower_attribute increment due to login (tracking times logged in for that user, only done while logging in) both of them are query time 2. but once the first query is done by siege.. there isn't any more slow queries.
01:04 Waylon       yet the siege time still increaes.
01:06 Waylon       this server runs on a virtual machine, how much memory should i allow for each apache child in calculations?
01:06 Waylon       xen vm that is.
01:06 Waylon       not a first gen virtualbox.
01:20 Waylon       okay.. my apache is prefork type.
01:20 Waylon       StartServers          5
01:20 Waylon       MinSpareServers       5
01:20 Waylon       MaxSpareServers      10
01:20 Waylon       MaxClients          150
01:20 Waylon       MaxRequestsPerChild   0
01:42 druthb       how big is count(*) from sessions?
01:43 druthb       Waylon:  ^^   I've found that cleaning up the clutter there helps.
01:45 jcamins_away Yikes! Pre-3.0?
01:46 Waylon       3.0.0.pre2 i think it started out as.
01:46 Waylon       yea
01:46 jcamins_away Better you than me!
01:46 jcamins_away Waylon: I have no useful advice to offer, but I'd agree with druthb that your sessions table should be cleaned.
01:47 jcamins_away And with eythian that you should figure out a way to upgrade.
01:47 Waylon       eventually ill get it upgraded, and be developing with you guys.. until then... this job is mostly sysadmin and coding additions.
01:48 Waylon       1035 sessions.. yeah....
01:49 druthb       well, that's not the issue--the times I've seen it being a problem, it had tens of thousands of entries.
01:53 Waylon       its only 15 concurrent users for the siege.. i don't get why the response time increases.
01:55 Waylon       ah.. adding a branch limit on the zebra query, adds a second.
01:57 Waylon       but, a search that results in a precise match redirect, with only basic kw search... took 11 seconds to redirect.
01:58 Waylon       maybe i need to tune zebra?
02:02 eythian      do you know that zebra is the issue?
02:02 eythian      if not, you can't begin to tune it
02:11 Waylon       mysqld is at the top of the top list..
02:11 jcamins_away Waylon: I would be shocked if it were Zebra.
02:11 * Waylon     nods.
02:11 jcamins_away Waylon: in fact, if it were Zebra, I'd be so shocked that I might fall over.
02:11 jcamins_away ;)
02:12 Waylon       it may be archiac and mysterious, difficult to work with, but it was coded well?
02:12 jcamins_away Waylon: I don't know about "coded well," but it's pretty fast.
02:14 eythian      when it works, it works really quite well. When it breaks it gets mysterious.
02:14 * Waylon     nods.
02:14 eythian      I think this is due in no small part to me not really understanding how it works inside.
02:14 Waylon       cryptic messages like error: 114
02:15 eythian      yeah
02:15 Waylon       looks like mysql might need optimizing.
02:16 eythian      there also have been speed improvements in newer Koha versions...just sayin' :)
02:16 Waylon       yeah.
02:19 Waylon2      ouch
02:20 jcamins_away Hmmm... I think Waylon is having some connection troubles.
02:21 Waylon4      ...
02:21 Waylon4      damn nz net..
02:21 jcamins_away Waylon4: what on earth are you doing?
02:22 jcamins_away (as an aside, I think you missed Waylon3)
02:26 Waylon23     router needed a complete reset.... not sure why.
02:26 jcamins_away 23? Wow, we missed 5-24!
02:26 Waylon23     hah. nah, i just added a random number.
02:28 Waylon       and my reindex cron job just started on the test server.. and i didn't load siege on a screen. damn.
02:28 Waylon       can't see the performance changes.
02:30 Waylon       anyway, afk, back later
03:20 AmitG        heya bag, rangi
03:20 bag          hey there AmitG
05:07 cait         good morning #koha
05:07 eythian      hello cait
05:07 cait         hi eythian :)
05:11 huginn`      New commit(s) needsignoff: [Bug 5473] 952 fields should be filled in by Acquisitions <http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=5473>
05:15 cait         eythian: what are you working on that late?
05:15 eythian      it's not late, it's still sunny outside.
05:15 eythian      I'm making the returns process be not dependent on barcodes.
05:15 eythian      i.e, you can do a return using an item number.
05:16 eythian      however, a lot of the code around there is Old Code that treats barcode like a primary key, so I'm having to refactor a whole lot.
05:17 eythian      the next step for this would be allowing searching for titles on the circ screen to issue/return things.
05:17 eythian      (for a new client we have who doesn't tend to use barcodes so much)
05:21 cait         ah
05:22 cait         well good luck :)
05:22 eythian      yeah, ta :)
05:43 AmitG        heya Robin, cait
05:44 cait         morning AmitG
05:44 cait         time to go - bbiab
05:47 eythian      hi
05:58 Waylon       hi
05:58 wahanui      hey, Waylon
05:59 Waylon       .... that was abit quick for a human.
05:59 eythian      wahanui is always on the ball.
05:59 Waylon       A bot?
05:59 wahanui      a bot are longstrings
05:59 Waylon       a bot.
06:00 Waylon       !commands
06:00 Waylon       no?
06:00 wahanui      somebody said no was failing unit tests
06:01 Waylon       koha?
06:01 wahanui      it has been said that koha is a free software ils see http://koha-community.org for more info
06:01 Waylon       liblime?
06:01 wahanui      liblime is no brandcuffs
06:01 Waylon       ptfs?
06:01 wahanui      ptfs is Piss-Taking Flaming Swines!
06:01 eythian      ‽
06:01 Waylon       omg..
06:02 eythian      Waylon: hipsters is http://www.harrisburgbicycleclub.org/images/stories/wheelclub1.jpg
06:02 eythian      err
06:02 eythian      wahanui: hipsters is http://www.harrisburgbicycleclub.org/images/stories/wheelclub1.jpg
06:02 wahanui      i already had it that way, eythian.
06:02 eythian      heh good.
06:03 Waylon       waylon?
06:03 wahanui      i heard waylon was having some connection troubles.
06:04 Waylon       wahanui, Waylon is the 33rd contributor to Koha. Currently works for Mandumah.com.
06:05 wahanui      ...but waylon is having some connection troubles....
06:05 Waylon       hrm.. damn.
06:05 mib_41oce5   Hello
06:05 eythian      no wahanui, Waylon is the 33rd contributor to Koha. Currently works for Mandumah.com.
06:05 wahanui      okay, eythian.
06:06 Waylon       hello mib_41oce5 !
06:06 mib_41oce5   I am new to Koha
06:06 mib_41oce5   one basic question: is Koha and LibLime koha different?
06:06 eythian      mib_41oce5: yep
06:07 Waylon       welcome! Yes, sadly, Koha and Liblime broke company some time ago... Liblime has a fork off of old code... and has extensively customised it.
06:07 mib_41oce5   ok, damn, two days wasted
06:07 mib_41oce5   I installed debain in my Mac using Virtual Box
06:08 Waylon       Hmm? at least you haven't invested money into a liblime solution.
06:08 mib_41oce5   and simply searched google for Koha and ended up installing liblime koha
06:08 Waylon       ooo.
06:08 eythian      sounds unpleasant
06:08 Waylon       how did that happen?
06:08 mib_41oce5   but now when I researched more
06:08 Waylon       so.. the first result that comes up is liblime koha?
06:08 mib_41oce5   yes
06:09 mib_41oce5   i ended up installing from version 4.08 from liblime but it's not working
06:09 Waylon       yeah... we can't help you with liblime.
06:10 Waylon       at least i can't.
06:10 mib_41oce5   have to uninstall and install 3.4 from koha
06:10 mib_41oce5   yes, I understand
06:10 eythian      http://wiki.koha-community.org/wiki/Koha_3.6_on_Debian_Squeeze <-- mib_41oce5, perhaps start here
06:10 eythian      3.6 is the current version that you want.
06:10 * Waylon     nods. "For maximum support from this channel, yes.
06:10 Waylon       3.6...
06:10 mib_41oce5   Thanks for the link
06:10 mib_41oce5   oh 3.6
06:10 Waylon       man, im way behind on mandumah.
06:11 Waylon       prehistorically behind.
06:11 eythian      that's what we keep telling you ;)
06:11 mib_41oce5   ok great thanks
06:11 mib_41oce5   will have a got at it today
06:12 kf           hi #koha
06:12 Waylon       yeah. gotta recode all this stuff into a module that i can patch in and out without too much trouble though. And teach the designer how to use koha's translation system, and css customization to move things around.
06:14 Waylon       so yeah.. alot of work to upgrade.
06:14 eythian      yeah. The best way to deal with that really is just to keep on top of it all.
06:16 * Waylon     nods. "And then you got the os security updates, vunerablility reports, optimization, drupal security updates... properly on top of, its a fulltime job?"
06:16 * Waylon     is the sole admin of mandumah.com... and is feeling slightly bewildered by how much there is to do.
06:17 eythian      clearly you just need to automate things ;)
06:18 Waylon       how do you automate without the risk of breaking something?
06:18 eythian      you have a test server you roll out to first.
06:20 * Waylon     nods, "I have one of those. my 2 bare metal xen servers, have 1 live server and 1 non-public server, one will be used for dev,. the other for all updates.
06:20 * eythian    needs to develop some Puppet modules for Koha
06:20 Waylon       hmm? Puppet?
06:21 eythian      puppet makes automating server configs easier.
06:23 Waylon       ah.
06:25 Waylon       I think ill do rsync from the update server to the 2 live servers, of koha code and the various configs that need to be changed. (apache, bind), and the actual data archive.
06:28 eythian      http://www.kallisti.net.nz/~robin/kohacon11_upstreaming_handout.pdf <-- also read my KohaCon slides if you need more incentive ;)
06:28 Waylon       yeah.. incentives are good.
06:29 Waylon       wha? a arabic crowd?
06:29 Waylon       where was this?
06:30 eythian      I was doing a workshop thing in Malaysia, that's where that photo is from.
06:30 Waylon       ah wow cool.
06:31 Waylon       yeah.. i suppose, properly coded, my database within a database idea might be useful.
06:31 eythian      every problem can be solved just by adding another layer of abstraction, after all...
06:32 Waylon       that is, mandumah is seperated into three segments, currently. but all in the same mysql/zebra ... just abusing the branch table and marc tags.
06:33 Waylon       cause.. we don't have branches.
06:33 eythian      wouldn't it be easier just to pretend they're branches?
06:34 Waylon       yeah. thats what i do. for each biblio... those segments that have that biblio in them, have a item in the biblio.
06:34 Waylon       ... that changes in 3.6 though, right?
06:34 Waylon       zebra doesn't include items anymore?
06:34 eythian      zebra does, just the database doesn't
06:35 eythian      basically, 952 doesn't exist and is generated when needed.
06:35 eythian      (which includes sending to zebra)
06:35 Waylon       ah... hrm... so.. how would my system be ported then?
06:35 eythian      no reason you couldn't do it the same way.
06:36 eythian      you just might query differently inside Koha.
06:37 eythian      although my recommendation would be to rebuild it, and make it an option inside Koha so that you can make it public and not worry about it in the future :)
06:37 eythian      wholly chao, my code works.
06:37 Waylon       yeah. ill probably make an actual table with actual custom marc tags.
06:37 Waylon       i think...
06:38 Waylon       eh.. im not sure.
06:38 eythian      I'm going to leave it there and go away and hopefully it'll still be working tomorrow.
06:39 eythian      well, there is the extended marc thing still there.
06:39 Waylon       eythian?
06:39 wahanui      eythian is probably in NZ. ;) or a good influence
06:39 Waylon       what, thats all thats said? hrm.
06:39 Waylon       robin?
06:39 wahanui      robin is probably on annual leave
06:39 eythian      but, it does seem like you're just making more pain for yourself.
06:39 Waylon       Yeah. well.. its what the boss wants.
06:40 eythian      sure, but you could do it in a more portable fashion and ultimately not worry about it.
06:40 Waylon       hmm, any advice?
06:41 eythian      Not really. I'd write up what you need it to do and start a mailing list discussion. I don't really have enough information to be constructive.
06:41 * Waylon     nods.
06:42 eythian      later, y'all
06:42 Waylon       And the authentication by ip , iprange and virtualhost, may be useful to others too.
06:42 eythian      definitely.
06:42 eythian      we're always adding new mad authentication schemes
06:42 Waylon       and referer url
06:43 Waylon       so.. currently, virtualhost + referer url  or ip address + virtualhost = pass.
06:43 Waylon       on top of standard username/password.
07:05 magnuse      kia ora #koha
07:09 rangi        hi magnuse
07:09 kf           hi magnuse
07:13 bag          eveing
07:13 bag          heya magnuse rangi and kf
07:14 kf           hi bag
07:15 magnuse      hiya rangi kf bag et al
07:24 indradg      @wunder kolkata
07:24 huginn`      indradg: The current temperature in Kolkata, India is 21.0�C (12:50 PM IST on December 19, 2011). Conditions: Mist. Humidity: 64%. Dew Point: 14.0�C. Pressure: 29.95 in 1014 hPa (Falling).
07:25 magnuse      @wunder boo
07:25 huginn`      magnuse: The current temperature in Bodo, Norway is 1.0�C (7:50 AM CET on December 19, 2011). Conditions: Mostly Cloudy. Humidity: 70%. Dew Point: -4.0�C. Windchill: -4.0�C. Pressure: 29.42 in 996 hPa (Steady).
07:27 alex_a       morning koha
07:29 magnuse      bonjour alex_a
07:33 kf           hi alex_a
07:33 alex_a       bonjour magnuse
07:33 alex_a       and kf
07:35 julian_m     hi #koha
07:38 kf           hi julian_m
07:39 julian_m     hi kf
07:44 sophie_m     hello #koha
07:57 Oak          kia ora #koha
07:57 Oak          magnuse
07:57 magnuse      Oak
07:57 Oak          :)
07:58 kf           hi Oak
07:58 Oak          Guten Morgen kf :)
08:11 asaurat      hi!
08:12 Oak          hello asaurat
08:13 kf           hi asaurat
08:20 gaetan_B     hello
08:20 wahanui      hey, gaetan_B
08:21 kf           hi gaetan_B
08:21 kf           Gentle reminder to everyone: plz sing off on patches even if it's not GBSD!
08:22 magnuse      kf++
08:23 AmitG        heya gaetan_B
08:28 kf           and I am serious!
08:30 Oak          kf++
08:30 kf           no karma, sign offs plz! :)
08:36 Waylon       can Location be a zebra search parameter?
08:36 Waylon       like branch: is?
08:42 reiveune     hello
08:44 kf           Waylon: I think so
08:44 kf           Waylon: in current releases there is a free text search option for location
08:45 kf           and hi reiveune
08:45 * Waylon     has been abusing the branch system to create divisions within the same database into different sections of the database.
08:45 Waylon       so a user that had access to A branch and B branch could only search A or B or both branches.
08:45 Waylon       not C branch.
08:58 huginn`      New commit(s) needsignoff: [Bug 6296] Allow authentication to Koha via PKI / x.509 certificates <http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=6296>
09:15 francharb    hi
09:15 wahanui      salut, francharb
09:30 mbalmer      so, if instead of writing all of SQL directly in the code, some of it is generated using a Perl module, that Perl module can produce optimal, database dependant SQL, e.g.:
09:30 mbalmer      $sql = "select * from table where " . datediff(now, then);
09:31 AmitG        heya francharb
09:31 wahanui      francharb falls asleep easily at sea.  Don't believe me?  Just ask Brooke.
09:31 mbalmer      so datediff produces the right SQL code for mysql or pgsql, depending on which db is being used.
09:31 mbalmer      this technique could be used for low-level stuff, but also for high level operations:
09:32 mbalmer      $sql = sort_all_books;
09:40 mbalmer      I suggest two different modules for this:  SQL for the lower-level SQL formatting, with functions like SQL::today, and DB for higher level procedures, which may boild down to many lines of perl code in the Mysql case while simply calling a stored procedure in the PostgreSQL case.
11:46 huginn`      New commit(s) needsignoff: [Bug 5607] Make fields from issues table available in overdues <http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=5607>
12:04 AmitG        heya NateC
12:04 mtj          mbalmer:  you could call a stored procedure in mysql too?
12:05 mtj          http://dev.mysql.com/doc/refman/5.0/en/create-procedure.html ?
12:06 mtj          http://net.tutsplus.com/tutorials/an-introduction-to-stored-procedures/
12:08 mbalmer      mtj, that was just an "example".
12:09 mbalmer      i.e. to show the flexibility of such a solution
12:11 mtj          someone recently corrected me when i said 'mysql doesn't have stored procedures'
12:11 mtj          ...so i learned something
12:12 mtj          (looks like mysql got them in version 5)
12:15 NateC        Hi AmitG!
12:16 NateC        and a good day to everyone else!
12:24 * kf         waves
12:25 * mtj        waves to nate n cait
12:26 kf           heya mason :)
12:27 mtj          heya, long time, no chat :)
12:28 * magnuse    waves
12:33 kf           mtj: I blame the time zones :)
12:36 mtj          kf:  yeah, i agree :) and its getting close to the end of the year, too
12:37 kf           it's been crazy here
12:37 kf           timezones can be blamed for a lot of things
12:38 kf           and they don't complain about it
12:48 tcohen       hi #koha
12:56 sophie_m     kf : I have sent a patch for bug 5607, if you want to sign-off
12:56 huginn       04Bug http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=5607 enhancement, PATCH-Sent, ---, sophie.meynieux, ASSIGNED , Make fields from issues table available in overdues
12:59 kf           sophie_m: I already moved it on my todo list :)
12:59 kf           thx!
12:59 sophie_m     kf :-)
13:00 kf           glad rangi did the last sign off for the other overdue bug
13:01 kf           hard to find time :(
13:01 kf           but this one should not be hard to test :)
13:02 sophie_m     no, very easy
14:11 magnuse      \o/
14:34 kf           hm
14:34 kf           is there a reason we have no timestamp in borrowers?
14:34 kf           and would it be nice to have one?
14:34 jcamins_away kf: no, and yes, respectively.
14:36 kf           hm ok
14:50 mbalmer      I did a few tests with (semi)generated SQL vs. static SQL as Koha uses now.  It seems to be a route that workable any only minimal invasive to the existing codebase.
14:58 kf           hi maximep :)
14:59 maximep      gooooooooood morning
15:01 mbalmer      maybe we expand the userbase in .ch in january a bit ;)
15:01 kf           ah, a library interested in koha?
15:02 mbalmer      yes, maybe.  at least we get the chance to present it.
15:07 wizzyrea     mornin
15:08 kf           hi wizzyrea :)
15:09 kf           mbalmer: I gave a presenttion about koha at inetbib in zürich hm last year
15:09 kf           got a lot of interest
15:10 mbalmer      this will be more discussion style.  this library is my customer since many years,  and they are a bit "unhappy" with their current product.
15:48 Koha         Hi
15:48 wizzyrea     hi there
15:48 Koha         Amit@
15:50 Koha         bg thr
15:52 Koha         @wounder delhi
15:52 huginn       Koha: I've exhausted my database of quotes
15:53 wizzyrea     well that was random
15:55 Amitg        Hi
15:55 wizzyrea     welcome back
15:55 wizzyrea     ;)
15:56 Amitg        Hi irc on my mobile
15:57 Amitg        bg around?
15:57 hdl          hi Amitg
15:57 bg           yup
15:57 Amitg        Hi hdl
15:57 Amitg        Irc on my mobile
15:59 Amitg        Hi indra
16:12 sekjal       does anyone have an opinion on the usage of $dbh->last_insert_id()?  it's used very rarely in Koha, and I'm wondering if there is a particular reason
16:12 jcamins_away sekjal: it's not thread-safe.
16:13 jcamins_away sekjal: unless it's implemented in a different way in Perl/DBI than it is in libmysql.
16:13 sekjal       I could see there being a race condition... if another INSERT happens before the function is called, you'd get the wrong ID
16:13 jcamins_away Right.
16:14 sekjal       wondering how likely that actually is
16:14 hdl          last_insert_id can be and should be retricted to table.
16:14 jcamins_away hdl: still not thread-safe.
16:14 hdl          jcamins_away: yep.
16:14 jcamins_away sekjal: in my experience using C (granted, that was pretty high-performance), very.
16:14 hdl          But no other way to get the latest id in table.
16:15 jcamins_away hdl: yes, and it's a problem, but that's probably why Koha doesn't use it.
16:15 sekjal       suppose I could do a query looking for a row that matches all the field values I just put in...
16:15 mbalmer      with postgres you can use a serial, maybe mysql has sth similar.
16:15 hdl          otherwise, we should do a select on the table... which imho is quite... awkward too.
16:15 sekjal       that seems so inelegant, though
16:16 sekjal       what I'm looking to do:  alter EnqueueLetter to return the message_id of the newly added message, instead of just "true"
16:17 mbalmer      check if mysql does have the equivalent of PostgreSQL serials, then it is easy.
16:17 jcamins_away mbalmer: if Koha used transactions, it'd be easy.
16:17 mbalmer      not using transactions is a crime anyways.  but that is a different story.
16:17 jcamins_away Unfortunately, I'm pretty sure Koha uses autocommit=1.
16:18 mbalmer      the whole SQL underpinnings seem a bit, well, questionable.  but of course that leaves room for improvement.
16:18 mbalmer      ;)
16:18 mbalmer      and that means fun.
16:19 hdl          mbalmer: so many ways to make improvements :D
16:19 mbalmer      yeah ;)
16:20 mbalmer      I just hope I get a few customers that use it, so I can justify putting some effort in to the code.
16:24 sekjal       hmmm... I can't assume that the information INSERT'ed in EnqueueLetter is unique... it would be safe to do so with the timestamp, but that's using MySQL's NOW()
16:24 huginn       New commit(s) needsignoff: [Bug 6323] Attach/move items -- error handling cleanup <http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=6323>
16:25 sekjal       but if I order by the timestamp desc, and limit to 1, that should reduce the odds fetching the wrong info to virtually nil
16:27 Amitg        Hi sekjal
16:27 sekjal       hi, Amitg
16:27 mbalmer      sekjal, no hacks, do it right ;)
16:28 mbalmer      MySQL _has_ the SERIAL data type.  so use it.
16:29 sekjal       mbalmer:  are you proposing I change the Koha data structure for my project?
16:29 mbalmer      1) $id = next serial value
16:29 mbalmer      2) insert into … (id, …) values ($id, ...)
16:29 mbalmer      3) return $id
16:30 sekjal       mbalmer: so the 'next serial value' command reserves the ID?
16:31 mbalmer      a serial is a 'BIGINT UNSIGNED NOT NULL AUTO_INCREMENT UNIQUE'
16:31 maximep      i don't get at all the difference with last_insert_id and getting the next id
16:31 mbalmer      in postgresql you use 'nextval(serial_name)', there exists a similar idiom in MySQL
16:32 mbalmer      maximep, getting the id from a serial upfront is safe, thread safe as well.
16:32 sekjal       mbalmer:  does nextval(serial_name) reserve that new id?  so no other process could slip in and claim it in the interim?
16:32 mbalmer      beacuse each values is guaranteed to be produced only once
16:33 mbalmer      sekjal, exactly, that is what serials are for.
16:33 maximep      i dont understand that serials stuff... the mysql doc says it's just a good old auto_increment unique
16:33 hdl          sekjal: I would not take  it  for granted it works same as postgresql. That Would need some test...
16:33 mbalmer      it's for sure worth the time to read the docs about them
16:33 mbalmer      hdl, ack.
16:34 mbalmer      mysql is such a POS that I would also test it before using it ;)
16:34 sekjal       mbalmer:  making that kind of change to the Koha datastructure is outside my specification
16:34 sekjal       I'll have to find an alternate route
16:34 mbalmer      like non working, not safe code?
16:34 mbalmer      ;)
16:35 maximep      i dont understand how last_insert_id is not threadsafe. It returns the last insert from that particular connection, doesn't it ?
16:35 mbalmer      what if, in the meantime, someone else inserted sth?
16:36 maximep      someone else will be on a new connection to the db
16:36 hdl          maximep: should be... But if you are using persistent connection, then you may have some strange behaviour.
16:38 maximep      hmmm, pretty sure cgi scripts will always create a new connection
16:38 mbalmer      well, if you use MySQL you probably don't care to much about your data anyways, so the implementation details don't matter either ;P
16:38 sekjal       http://www.gossamer-threads.com/forum/General_C8/Perl_Programming_F14/MySQL%27s_last_insert_id_and_mod_perl_.._How_does_it_work%3F%3F_P171325/
16:39 maximep      threadsafe from the same sth
16:40 sekjal       ok, so by that logic, modifying EnqueueLetter should be okay.  dbh is created, sth is set and executed, then subroutine ends
16:40 mbalmer      using stuff like last insert ids only creates non-portable code.  creating a new unique id upfront, and using it in the next insert can be made portable.
16:41 maximep      i just dont understand how you can create a new id upfront
16:41 maximep      never seen anyone do that in myusql
16:41 mbalmer      by using a monotonically increasing number.
16:42 sekjal       mbalmer:  if you file a sub-bug to your "Koha uses MySQL" omnibus, detailing the methodology required for this, we can work on implementing it for Koha 4.0
16:43 sekjal       until then, I'm going to stick with last_insert_id(), as it's safe in this specific instance, and does have precedent elsewhere in the codebase
16:43 mbalmer      no two callers can get same number if you use the right mechanism.
16:43 mbalmer      yeah, repeating errors makes it correct use ;)
16:44 mbalmer      but go on, really, it is easy enough to fix later.
16:44 sekjal       mbalmer:  I'll happy to fix it in both my instance, and the legacy ones, at such time as there is a basis for it
16:45 mbalmer      as I said before, personally I think that code should be correct.  But then maybe that is because I come from a different corner ;)
16:45 mbalmer      (development wise, that is)
16:46 reiveune     bye
16:46 maximep      you come from a postgresql snob corner :p
16:47 wizzyrea     ^^
16:47 maximep      i wish I had a project to learn to use all it's powerful features
16:47 maximep      because i know it's way more powerful than mysql. someday :/
16:47 mbalmer      maximep, only partly true (the snobcorner), it's actually worse ;)
16:48 mbalmer      with the introduction of a shim layer, I think that both mysql and pgsql can be used.
17:01 kf           ok, time to leave - bye all :)
17:06 rhcl         time to hide
17:09 hdl          see ya
17:17 gaetan_B     bye
17:41 cait         hi again #koha
17:41 ketchup      hihi
17:41 wizzyrea     woot cait is back
17:42 cait         missed me? :)
17:42 ketchup      missed the cookies
17:42 * cait       sighs
17:42 cait         ok
17:42 * cait       hands over the cookies
17:42 ketchup      or rather, yes, we missed you cait
17:43 cait         too late!
17:43 cait         ;)
17:43 * wizzyrea   missed cait
17:44 wizzyrea     and her shining, smiling face.
17:44 cait         shinnig?
17:44 wizzyrea     http://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&ved=0CCYQtwIwAA&url=http%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DiCQ0vDAbF7s&ei=_nfvTrDRGMKItgezloWECw&usg=AFQjCNFanIdji5CxhT-Gi6teLsg-mRmYEw
17:46 * wizzyrea   gets to work on everybody's digital xmas gifts.
17:46 wizzyrea     also: nice unibrow.
17:47 cait         can't watch the video - blocked from here
17:47 wizzyrea     NO!
17:47 wizzyrea     aw
17:47 cait         digital xmas gifts?
17:48 wizzyrea     i'm getting you all presents and wrapping them in small urls so you don't know what they are ;)
17:49 cait         oh
17:49 cait         that's wow
17:49 * cait       panics because she has nothing to give
17:49 wizzyrea     don't
17:51 wizzyrea     does this work cait: http://grooveshark.com/s/Shiny+Happy+People/365FuD?src=5
17:51 cait         grooveshark is normally ok :)
17:51 cait         but youtube - lots of problems
17:52 wizzyrea     yea :(
18:31 magnus_afk   cait, wizzyrea: Shiny Happy People works for me! ;-)
18:31 cait         because yo have no gema
18:32 rangi        morning
18:32 cait         morning rangi
18:32 rangi        @wunder wellington nz
18:33 huginn       rangi: The current temperature in Wellington, New Zealand is 12.0�C (7:00 AM NZDT on December 20, 2011). Conditions: Mostly Cloudy. Humidity: 67%. Dew Point: 6.0�C. Pressure: 30.15 in 1021 hPa (Rising).
18:35 rangi        bus time
18:37 rangi        I love this week on the bus, no schoolkids, and not many others
18:43 rangi        !jenkins status
18:43 jenkins_koha status of all projects:
18:43 jenkins_koha Koha_3.2.x: last build: 36 (15 j ago): UNSTABLE: http://jenkins.koha-community.org/job/Koha_3.2.x/36/
18:43 jenkins_koha Koha_3.4.x: last build: 68 (10 j ago): UNSTABLE: http://jenkins.koha-community.org/job/Koha_3.4.x/68/
18:43 jenkins_koha Koha_3.6.x: last build: 25 (5 j 0 h ago): SUCCESS: http://jenkins.koha-community.org/job/Koha_3.6.x/25/
18:43 jenkins_koha Koha_master: last build: 578 (3 j 1 h ago): SUCCESS: http://jenkins.koha-community.org/job/Koha_master/578/
18:43 cait         hm
18:43 cait         could be worse?
18:43 cait         what's the problem with 3.4?
18:44 rangi        just some patches need to be cherry picked
18:45 rangi        its test fixes, not broken functionality but be nice to stop them complaining
18:55 rangi        ibeardslee: did francois talk to you about the freedom lunch idea?
18:58 ibeardslee   no he didn't
19:00 rangi        ahh was an idea that one of the days, say these ppl are gonna be eating lunch someplace (level 6 kitchen maybe) you should join them to talk about software freedom
19:00 rangi        thats as far as we had thought
19:01 rangi        bus stop bbiab
19:01 ibeardslee   the academy?
19:06 rangi        back
19:07 rangi        yeah
19:07 rangi        try and catch the ones who are interested
19:10 ibeardslee   24 of them, but not after a substitute for the session that gets everyone
19:12 ibeardslee   'freedom' is an important part of the Academy
19:12 rangi        yup
19:12 rangi        this was to catch the ones that really got it
19:12 ibeardslee   ahh, and additional session
19:12 ibeardslee   sure no harm in that
19:27 cait         @wunder Konstanz
19:27 huginn       cait: The current temperature in Taegerwilen, Taegerwilen, Germany is -1.7�C (8:24 PM CET on December 19, 2011). Conditions: Mostly Cloudy. Humidity: 91%. Dew Point: -3.0�C. Windchill: -2.0�C. Pressure: 30.20 in 1022.6 hPa (Falling).
19:27 maximep      let's see what random city i will get
19:27 maximep      @wunder yqc
19:27 huginn       maximep: The current temperature in Kangirsuk, Quebec is -16.0�C (2:00 PM EST on December 19, 2011). Conditions: Low Drifting Snow. Humidity: 72%. Dew Point: -20.0�C. Windchill: -28.0�C. Pressure: 29.51 in 999 hPa (Rising).
19:28 cait         brrrrr!
19:28 maximep      @wunder ymx
19:28 huginn       maximep: The current temperature in Montreal Mirabel, Quebec is -2.0�C (2:00 PM EST on December 19, 2011). Conditions: Light Snow. Humidity: 86%. Dew Point: -4.0�C. Pressure: 29.67 in 1005 hPa (Falling).
19:28 maximep      closer to the truth
19:36 bag          @wunder sba
19:36 huginn       bag: The current temperature in Santa Barbara, California is 13.0�C (10:53 AM PST on December 19, 2011). Conditions: Overcast. Humidity: 75%. Dew Point: 9.0�C. Pressure: 30.07 in 1018 hPa (Falling).
19:38 wizzyrea     can anyone explain why this appears? http://screencast.com/t/KxbFdxkPQ6U
19:39 rangi        hmm not me, never seen that before
19:39 rangi        what does the .tt file say
19:39 wizzyrea     [Mon Dec 19 13:35:30 2011] [error] [client] [Mon Dec 19 13:35:30 2011] batchMod.pl: Problem = a value of 1 has been passed to param without key at /home/liz/kohaclone/C4/Templates.pm line 185., referer: http://koha-staff.
19:39 rangi        ie, how does that get set
19:39 wizzyrea     that's next ;)
19:42 cait         hm some permisson thing perhaps?
19:42 cait         indy branches?
19:42 wizzyrea     no, I don't think so
19:42 cait         user has no rights to edit catalog?
19:42 wizzyrea     is superlibrarian (but not the koha user), no indy branches
19:42 cait         ok, out of ideas :)
19:42 cait         I am looking at your new patch
19:43 cait         i was wondering
19:43 jcamins_away wizzyrea: comma missing in a template->param() block.
19:43 cait         why it always asked me to "try again"
19:43 wizzyrea     what do you want it to say
19:44 * wizzyrea   has changed the wording on advisement 3x now
19:44 wizzyrea     what's one more
19:44 wizzyrea     oo
19:44 wizzyrea     jcamins_away: I will look for that.
19:45 jcamins_away wizzyrea: or an equal when there should be an =>
19:49 cait         heh
19:50 cait         only thought it sounded like there was an error
19:50 cait         when it actually succeeded
19:53 wizzyrea     I can make it say something different when it succeeds
19:53 wizzyrea     if that's better
19:53 cait         not sure it's necessary
19:53 * wizzyrea   doesn't care, the workflow on that screen is awkward and needs to be fixed.
19:53 wizzyrea     which is what the patch does.
19:53 cait         was only what I was thinking looking at it
19:53 cait         yep
19:53 cait         it's a nice enhancement :)
19:54 wizzyrea     naw if you think it's bad wording we should change it. I'll fix it
19:55 mbalmer      re
19:55 cait         wizzyrea: sorry for cuasing you more work
19:57 wizzyrea     ah it's all good
20:14 jcamins_away Why's nengard recording new videos for 3.2?
20:16 rangi        shes a hipster?
20:17 bag          what are you talking about jcamins?
20:17 bag          sorry jcamins_away
20:17 jcamins      :)
20:17 rangi        http://www.web2learning.net/archives/4901
20:17 rangi        jsut came through in the feeds
20:17 jcamins      She just posted a new video on the Koha 3.2 OPAC, didn't she?
20:17 magnus_afk   perhaps it's the medication? ;-)
20:17 * magnus_afk hides
20:18 bag          maybe she linked to the wrong video
20:19 jcamins      Possible. I just read the title and was confused.
20:19 jcamins      Either that or it could be that the title is wrong.
20:25 mbalmer      gn8 koha world.
21:07 rangi        w00t w00t
21:07 rangi        liw: check this out http://www.kansalliskirjasto.fi/kirjastoala/uutiset/1323856166325.html
21:08 rhcl         cool, really cool
21:08 rangi        yeah it is
21:08 rhcl         what's it say?
21:08 wizzyrea     there's a "in english" link at the top
21:08 wizzyrea     an*
21:09 rangi        summary
21:09 rangi        National Library of Finland cancels Ex Libris Primo pilot and investigates Open Source for national digital library
21:09 jcamins      Awesome!
21:10 cait         finland++ :)
21:12 rhcl         the english link at the top took me to their index page, not the document of interest, but translate.google.com gave me the gist of it
21:13 sekjal       cait:  you were mentioning something about a timestamp on the borrowers table earlier today
21:13 cait         ype
21:13 cait         yes
21:13 cait         something like a last update date
21:13 sekjal       I was thinking, it would be handy to have timestamps for emails, addresses and phone numbers, for "date last verified"
21:14 cait         I am not sure if this isn't overcomplicating things
21:14 sekjal       I"m sure it is
21:14 cait         you can get the update information from action_logs I think
21:14 sekjal       the library I was training last week was very information in a verification method for emails, because they don't trust the data they have
21:14 cait         it's only a bit hard to get out and a timestamp would also show the date when a borrower was created or transferred - a bit faster than querying action_logs
21:17 cait         transferred to deletedborrowers
21:17 cait         which brought up that thought today :)
21:17 cait         I was wondering about the fastest way to find out when a borrower was deleted
21:17 wizzyrea     action logs?
21:17 wizzyrea     oh
21:17 wizzyrea     nm
21:17 wizzyrea     reading fail
21:17 * wizzyrea   goes back to her hole.
21:23 cait         hm
21:23 cait         bed time :) night #koha
22:21 liw          rangi, meh, the page now says the original press release has been replaced with a new one, which says they'll continue to try to use primo, while using some open source stuff too
22:21 rangi        ahhh
22:32 rangi        liw: the inside story is "That's correct.But it doesn't say they'll use Primo, but Primo Central, which is the metadata index only
22:32 rangi        "
22:34 rangi        which is just a big catalogue essentially
22:34 rangi        but they are gonna use open source to expose it
22:58 liw          rangi, yeah, that sounds about right, from the Finnish text -- it's still weird that they backed out from a press release
23:00 rangi        yeah
23:12 rhcl         ping melia
23:12 melia        hi
23:12 liw          rangi, I don't see a mention there of which open source library software they'll use -- koha? something else?
23:58 bag          NateC?
23:59 rhcl         bag: whatup?