IRC log for #koha, 2008-07-04

← Previous day | Today | Next day → | Search | Index

All times shown according to UTC.

Time Nick Message
13:29 ryan hdl: around ?
13:49 paul ryan: nope, hdl is teaching Koha. He should appear in a few hours, with hdl_laptop nickname
14:37 acmoore at YAPC this year, everyone was talking about how great the NY Times Profiler is http://open.blogs.nytimes.com/[…]es-perl-profiler/
14:37 perhps it would be a good took for those of us who are investigating performance improvements.
14:41 paul acmoore: you should drop a mail to koha-devel with this link...
14:41 gmcharlt greetings paul et #koha
14:41 paul hello gmcharlt (& acmoore)
14:47 acmoore OK, paul. I did.
15:04 atz acmoore: nice link
15:17 gmcharlt hdl: about?
15:18 paul gmcharlt: hdl is away during the day for the week (teaching Koha at a library)
15:18 should appear under nick hdl_laptop in a moment.
15:18 gmcharlt paul: then is he actively working on bug 2297?
15:18 paul he investigated it yesterday, and think that :
15:18 - he understood were the problem was
15:19 - was unhappy to see that someone broke a feature that used to work, probably by ignoring some specificities of the code (his fault : lack of POD & comments)
15:20 so he was sure enough that he could fix the problem quickly.
15:20 gmcharlt the thing is, I have a patch ready to go soon for it as well
15:20 paul send it & copy it to him.
15:20 gmcharlt the problem is the ModBiblio API
15:20 paul he will confirm or not.
15:21 is the "Columbia hostages freed " affair make as much noise in US as in France ?
15:21 gmcharlt paul: it made the front page of the Chicago Tribune
15:21 paul because here, you can't open radio, TV, internet or newspaper without hearing of it !!!
15:22 gmcharlt paul: but regarding 2297, I think I'm going to submit a patch to deal with all uses of ModBiblio and how it handles embedded items, since the issue is broader than just the subscription add
16:04 Kyle Hey all. We just got a vmware appliance up and running for Koha 3 RC 1 on Debian Etch.
16:04 paul hi Kyle
16:04 great !
16:05 gmcharlt Kyle++
16:05 Kyle Hi paul. I'd like to put the image on koha.org, does anyone know how I would go about that?
16:05 paul mmm... I would say "ask chris" & maybe "wait next week to upload the VMware of 3.0 official !
16:06 Kyle If 3.0 is out next week, I think waiting is a good idea.
16:13 On a side not,e, which do you all think would be better, to upload an image that is ready to be setup ( that is, when you go to port 8080, it takes you though the install wizard ) or to upload an image that is already set up with the example data.
16:14 or I could do one of each.
16:14 gmcharlt Kyle: one of each
16:14 Kyle choice is good ; )
16:14 gmcharlt Kyle: or perhaps better, the former along with a nicely detailed wiki entry on customizing the image
16:15 paul Kyle: the installer is supposed to solve this question.
16:15 as, if you have an empty DB, it will ask you what you want to setup
16:15 so, just put your sample DB in installer scope, and that's enough ;-)
16:16 gmcharlt paul: but if you want to just play around with Koha and have something that works with no setup whatsoever, that's also nice
16:16 Kyle yes, I really like that option of example data, afaik it doesn't create and sample items in the db, am I correct?
16:17 paul no, you're wrong. you can have a full sample DB, that overwrite all other parameters the user could choose
16:17 during install
16:17 kados: around ?
16:17 Kyle paul: you mean there is an option to put in example items?
16:17 paul it's SQL, so you can put whatever you want !
16:18 even a 100k production ready DB with 2 years of issuing history
16:18 gmcharlt Kyle: yes, the scripts in installer/data/mysql/<lang>
16:18 paul (just that it result in a large large .tar.gz...)
16:18 Kyle I mean, it creates example item types and patrons types, and example patrons, but I didn't see an option for dummy items during the web install.
16:19 paul because LibLime did not add a file for that.
16:19 but that's a choice, not a limit
16:19 you can just "mysqldump yourdb"
16:19 then add it to Koha, and that's OK, it can be "imported" by the installer !
16:19 mmm... another limit maybe the Apache timeout.
16:19 gmcharlt Kyle: uh, warning: doing that naively will violate patron privacy
16:20 paul so don't setup a 100k DB with 2 years history...
16:20 gmcharlt i.e., be selective in what you dump
16:20 paul gmcharlt: right ;-)
16:20 (I did a "anonymise.pl" script, that anonymise everything in a patron
16:20 '(mails, addresses, birthdate, phone, ...)
16:20 Kyle gmcharlt: yes, I understand that.
16:21 paul replacing it by a random sample or {rand}{street|av|bd}{something|somethingelse} {random zipcode} {random city}
16:21 Kyle paul: is that script in the Koha 3 repo?
16:21 paul nope.a
16:22 as it's frenchy addresses
16:22 but I can send it to you if you want
16:22 Kyle Ah, I see.
16:22 paul cities : {paris|Marseille|Lyon|Toulouse|...}
16:22 Kyle You don't need to. The example patrons already in there should work just fine.
16:23 For sample items, wouldn't something have to be done with zebra, as well as with mysql?
16:24 gmcharlt yes
16:24 installer would have to do something like kick off bulkmarcimport followed by a rebuild_zebra
16:26 Kyle Yes, that is why I think it might be nice to have a vmware image that is not yet setup, and one that is set up with all the example data, and some exmple items. That way if someone wants to try it out right away they can choose the already set up image, as it would have libraries, patrons, and items.
16:26 paul Kyle: ++
16:27 Kyle paul: thanks : )
16:27 atz a vmware image will get a lot more ppl who can't handle the linux admin parts of the install
16:28 they can run it on their windows workstation or whatever
16:28 Kyle agreed.
16:29 paul the best solution would be to have X versions : Zebra / NoZebra , UNIMARC / MARC21, en only / i18n, ...
16:30 atz so instead of the current "I installed 2.2.9 perfectly, but I got this error..." all the questions on [Koha] will be "So how do I edit the HTML?"
16:30 gmcharlt paul: hence why I suggested ading documentation about customizing the image
16:30 or we could just ask Kyle to do it all ;)
16:30 paul here in france, I think some libraries would be interested to have a vmware for "small library, that want simple cataloguing, with hide_marc=ON)
16:30 Kyle heh : )
16:30 paul having Kyle explaining is better, from far :D
16:31 Kyle If we can get a list of combinations hammered out like paul listed earlier ( Zebra/NoZebra, etc... ) I will see what I can do.
16:32 I think we should limit it to the big ones, so people wanting to try it out don't get overwhelmed by the options.
16:32 gmcharlt well, the default image w/o sample data could handle most of the options
16:32 atz Kyle: i wouldn't bother w/ a no zebra image.
16:32 gmcharlt e.g., have it have the Zebra processes be running, but user can choose NoZebra during installer if they wish
16:32 atz it is pretty easy to turn NoZebra on
16:33 right
16:33 gmcharlt isntall all of the langauges, etc.
16:33 then if you wanted a fr-FR NoZebra image with sample data, you could customize the base image
16:35 Kyle gmcharlt: do you mean customize the base image and upload it as another downloadable option, or have the user download the image and then customize it? I imagine having it pre-customized would be better. We can always have a "If you are not sure which version to download, pick this one" option.
16:35 gmcharlt the former
16:35 the default image would serve two purposes
16:36 1. version that allows a user to try out any combination of options
16:36 Kyle Yes, I see. That sounds good to me.
16:36 gmcharlt 2. basis for a furhter-customized image that someone does and uploads
16:38 Kyle yes. Essentially, we create version 1, then make copies of it and run the setup with whichever options we want, throw some items into the database, and uplaod it.
16:49 paul gmcharlt: hdl_laptop is here now ;-)
16:50 gmcharlt hi henri
16:53 hdl_laptop hi gmcharlt
16:53 paul told me you had a patch for subscriptions ?
16:53 gmcharlt hdl_laptop: bug 2297
16:54 change will be to have ModBiblio live up to its name and modify only the bib portion, not change the items
16:54 I'm writing a test case for it now
16:55 hdl_laptop Is it a problem of ModBiblio ???
16:56 This would be surprising to me.
16:56 gmcharlt a problem of its design, anyway
16:56 it goes like this:
16:56 1. retrieve MARC bib - it has items embedded in 952/955/whatever tags
16:56 2. munge the bib (assume that you don't care about the items)
16:56 3. call ModBiblio()
16:57 what ModBiblio does is add all of the item tags from the original version of the bib
16:57 to the MARC::Record that you passed in
16:57 then save the bib to DB
16:57 however, if you got that MARC::Record from GetBiblio
16:57 it *already* has the item fields
16:57 so ModBiblio will add them back
16:58 most clients of ModBiblio remember to remove the item tags first before calling ModBiblio
16:58 but not all
16:58 if, on other hand, ModBiblio() ignores any item tags from the incoming bib, problem goes away - APIs like AddItem() shoudl be used to handle the items
17:01 hdl_laptop I trust you intestiguated more than me.
17:01 I investiguated yesterday...
17:02 And was surprised to see that PrepareItems would use AddItem but not link serial created with the itemnumber.
17:02 but it was really late.
17:02 So I can have overlooked some parameters.
17:03 gmcharlt hdl_laptop: that part is probably a separate issue from the creation of duplciate item tags
17:08 hdl_laptop ok.
17:08 gmcharlt: are you sending your patch to the list or are you pushing it quite soon ?
17:08 gmcharlt hdl_laptop: I'll be sending patch to list soon
17:21 hdl_laptop thx gmcharlt
17:21 bbs
17:40 rhcl Anybody online here from LibLime? Maybe they are getting an early start for the holiday?
17:40 atz yeah, several of us are here
17:40 what's up?
17:41 acmoore here!
17:41 atz in fact, rhcl, from here you almost look like one of'm (that sometimes goes by "rch")
17:42 acmoore yeah, I caught myself in a double-take over that, too.
17:44 nengard i'm here too - wish i was getting an early holiday start!!
17:44 even though i just got off vaca
17:44 heh
17:44 atz raining here...  :\
17:44 nengard atz they say that's coming our way
17:44 gmcharlt perfect day in Chicago
17:45 acmoore we got a few inches of rain last night, designed just perfectly to try to prevent fires from the fireworks this weekend.
17:48 rhcl I just received an inquiry from a major software vendor of ILS "accessories". I'd kinda like to ask a brief question via email or pm of someone who can better speak for the Koha community than me.
17:48 gmcharlt rhc1: can it be asked publically on koha or koha-devel?
17:49 atz koha has verifiable SIP support now, if that's the question.  :)
17:50 acmoore gmcharlt, did you intend to add a "use CGI" in that patch you just sent me?
17:50 rhcl Maybe it's not so critical, but I would prefer a direct email with someone.
17:50 gmcharlt acmoore: yes, I did - required if you're running a single test class
17:51 rhcl Oh, no, it's not about SIP.
17:51 acmoore why? I don't see where it's used or anything.
17:51 gmcharlt acmoore: startup_30_login
17:52 acmoore gmcharlt, hmm. why was it working before?
17:52 (I'm not saying you're doing it wrong. I'm just saying I don't see it yet)
17:52 gmcharlt when running full test, CGI.pm must have been slurped in somewhere use
17:53 acmoore weird. well, whatever. It's obviously used, so it should be "use"d
17:53 OK. I'll apply this and try it out in a few minutes. good idea!
18:14 gmcharlt, I'm taking a look at your patch to the test suite. Have a second?
18:14 gmcharlt acmoore: ok
18:15 acmoore I can't quite get it to work, I guess. I get a lot of errors about redefined subs and "cannot test anonymous subs - you probably loaded a Test::Class too late" and then the whole test suite still runs.
18:15 gmcharlt acmoore: the TEST_FILES= is required
18:16 acmoore on the command line?
18:16 gmcharlt when you're using test-single - otherwise, defaults to database_dependent.pl, and has effect of trying to call runtests twice
18:16 acmoore: yes
18:17 acmoore I can't quite get it. What should the command line look like?
18:20 gmcharlt make test-single TEST_FILES=lib/KohaTest/Biblio.pm
18:21 acmoore aaahh.
18:22 atz i would have been fooled by that into doing:  TEST_FILES=lib/KohaTest/Biblio.pm  make test-single
18:22 like w/ other make vars
18:22 acmoore atz, I was, too.
18:23 gmcharlt, so, now it runs tests at least, but by running a few of them, I think I have my database well botched, so none of them really pass anymore.
18:23 gmcharlt the one is setting an env var, the second is setting a make variable
18:24 acmoore I'll run the whole suite again and see if I can restore the database enough to be usable again.
18:25 gmcharlt acmoore: calculated risk - one could put the whole DB clearing step into the INIT, but that would mean slowing the tests
18:25 long-term, IMO, better to design the test classes so that they can assume that the DB tables exist, but not make any particular assumptions about their contents, and to clean up after themselves
18:25 acmoore gmcharlt, have you sucessfully used this in a test-code-test-code-test cycle?
18:25 gmcharlt yes
18:26 acmoore OK, then it's probably worth me signing off on.
18:26 hdl_laptop hi again
18:26 acmoore I'd like to fix the docs, though. The commit message is misleading about the usage, I think.
18:26 gmcharlt I'll fix it
18:26 acmoore gmcharlt, can I edit the change before I sign it somehow?
18:26 gmcharlt I see now that I missed the TEST_FILES=bit
18:26 acmoore or, I guess, what's the procedure here?
18:27 gmcharlt acmoore: better for me to resubmit
18:27 acmoore OK. Whenever you get to it.
18:28 gmcharlt semt
18:29 sent, even
18:29 acmoore gmcharlt, and then I have patches for the overdue notices script. I'm sure I've missed some stuff and could use a second pair of eyes if you get some time. Can you give me 20 minites or so?
18:30 gmcharlt acmoore: sure, in about 10 minutes
18:33 acmoore I'll mail them. You don't have to apply them, but I'm sure you'll see something like missing docs or something.
18:34 gmcharlt, your patch signed and sent.
18:34 gmcharlt thx
18:48 acmoore gmcharlt, sent you 5 patches for 2274. If you get a moment to at least eyeball them, that would be great.
18:52 is Fr?d?ric some kind of super-linguist? He deserves a cookie or something.
19:14 frederic hello
19:14 Is there something wrong with koha-devel and koha-translate lists?
19:15 My messages appear in koha-devel hours after I post them
19:15 And don't appear at all in koha-translate
19:16 acmoore I saw a recent post from you on the translate list.
19:16 Here are some recent ones from you: http://koha.markmail.org/search/?q=frederic
19:16 RTL language stuff?
19:17 frederic acmoore: thanks. I haven't seen for koha-translate. Something wrong on my side
19:17 But it take forever to appear
19:17 Yes RTL
19:18 acmoore oh, it looks like markmail may not be archiving koha-translate.
19:19 hdl_laptop frederic: I had to ackowledge your message.
19:19 and had no internet connction.
19:20 frederic hdl_laptop: thks. And why do you have to?
19:21 hdl_laptop you were not member on that email before posting.
19:22 it seems.
19:22 frederic koha-translate or koha-devel?
19:23 atz OO barcodes modules sent.  nothing important yet, but pretty cool to look at (much better model for extension)
19:24 hdl_laptop koha-devel
19:24 koha-translate, no message pending approval
19:25 have you sent messages to that list recently ?
19:27 I have one from you on that list in my box sent today.
19:27 Koha Arabic Version
19:30 acmoore atz++
19:32 frederic atz: are you the one who know how to do OO programing in Perl?
19:32 atz sure, it's not that hard
19:33 frederic hdl_laptop: I sent 2 messages this morning (our morning) and no one appear immediatly, neither koha-devel nor koha-translate. And then I saw thme this evening.
19:33 atz it's just very easy to do procedural stuff
19:34 the SIP abstraction layer (ILS) is also OO
19:34 frederic atz: for you, it's easy. What's very confusing in Context.pm is the mix OO/procedures
19:34 atz yeah, that happens in many places
19:34 frederic Yes, I've seen. Very clean design
19:34 atz Dates.pm has that too, because it has legacy functions from Date.pm
19:35 frederic I would be so easy to have for example a Syspref class which hide the implementation... No SQL UPDATE in the code...
19:35 Later
19:36 atz frederic: yeah, that would be a good move
19:37 frederic Does anybody know someone who had to deal with right-to-left language support in Koha?
19:37 I can see that there is an Hebrew translation for example
19:37 acmoore I think gmcharlt is intending to write a mutator to let us set sysprefs without SQL.
19:37 frederic a mutator! I have to google
19:38 atz the Physics ppl in Islamabad probably did rtl, i think
19:39 acmoore frederic, or a "setter", a method to let you write the value of an attribute.
19:39 frederic what's strange is that rtl support requires specific care in template and CSS. For me it's not here yet. So I'm surprised there hasn't more feedback on that
19:39 acmoore: http://en.wikipedia.org/wiki/Mutator_method
19:39 I found
19:56 hdl_laptop gmcharlt: I have seen that DelItem still required (dbh) to be passed along as parameter.
19:57 I thought we decided not to pass dbh as function parameter but to use C4::Context->dbh
19:57 just to point it.
19:57 gmcharlt hdl_laptop: evidently it got missed - I didn't add back the $dbh
20:33 hdl_laptop gmcharlt: Is there anyone at liblime who has jquery knowledge and is available interested  in what I have been doing on subscription-add.pl ?
20:33 gmcharlt hdl_laptop: I can take a look at it
20:34 ccatalfo would be better, though
20:35 ccatalfo hdl_laptop: i can take a look, too.  (I'm ok on jquery knowledge; not so good on koha knowledge)
20:35 hdl_laptop It is is a really poor state at the moment. I am full of  ideas. but has no time and little skills on jquery so that it takes me too much time.
20:39 ccatalfo hdl_laptop: how far along have you gotten?  How much is left to be implemented?
21:04 acmoore The recent email brought my attention to the 'select 1' clause in C4::Context::dbh. Is that just there to make sure that we're connected to a database? It seems like there might be a better way to do that.
21:04 Isn't there a DBI::ping method or something? or, can we not depend on auto_reconnect to work?
21:05 I guess I'm wondering what problem occurred that caused us to add that.
21:09 hdl_laptop ccatalfo_: couldnot get too far unfortunately.
21:10 I detail in a mail.
21:22 gmcharlt hdl_laptop: got the email; will comment tomorrow
21:28 hdl_laptop thx.
21:36 ccatalfo_ hdl_laptop: thx for the email; i'll check it out tomorrow, too
21:36 hdl_laptop gmcharlt: is C4::Context->dbh access slower than passing dbh ?
21:37 + I thought at first glance that dbh was not used in DelItem
21:37 So that tehere would have been no good not to remove it from _koha_delete_item.
21:49 gmcharlt hdl_laptop: marginally slower, perhaps - I just think it's one thing to not pass in $dbh in the public API, and a different in the private one
21:54 hdl_laptop I agree.
22:04 getting tired.
22:04 good night
22:04 (midnight here)
22:54 rach howdy
22:54 I was thinking to put the venuzuala announcement on the koha website
22:54 any objections?
22:54 masonj oooh, great idea rach
07:22 mc hello world
07:22 désolé pour la grasse mat.: je viens juste d'arriver
07:25 oopps ...  i'm on #koha ... not #koha-fr
07:25 so please don't carre about the french line
07:25 and excuse me
08:18 paul mmm... investigating cpan mirrors, I find http://www.cs.uu.nl/stats/mirmon/cpan.html#fr
08:19 that says there is something wrong with cpan.catalyst.net.nz ...
11:45 js who know where is the firefox plugin to search in koha database ?
11:45 hi (before ;))

← Previous day | Today | Next day → | Search | Index

koha1