IRC log for #koha, 2008-03-16

All times shown according to UTC.

Time S Nick Message
13:05 hdl hi gmcharlt
13:05 gmcharlt hi hdl
13:05 hdl are you around ?
13:05 kados hi guys
13:06 hdl hi kados
13:06 gmcharlt hi kados
13:07 kados hdl: plan to do a bugfix meeting today still?
13:07 hdl Reading my message about bug sqaushing sessions back today, I realised that time schedule was wrong.
13:07 hdl But session I would like to have it today. Yes.
13:08 kados OK
13:08 hdl What time would be better for you all ?
13:09 kados well, /me wants to roll the 3.0 beta this weekend
13:09 kados and that could take about 6-8 hours :-)
13:10 hdl could we have it begin at 19pm French time ?
13:10 hdl Or should we begin now.
13:10 kados hdl: what time is that in GMT?
13:10 hdl 17GMT
13:11 kados so Noon US time
13:11 hdl OK.
13:11 hdl So we should start now.
13:12 kados hdl: I was thinking we should work on bugs, but also, I'd like to migrate the mailing lists next week
13:12 hdl And stop it at noon.
13:12 kados hdl: can we schedule a time to do that next week?
13:13 hdl Can you give me the login to catch the lists and have it on our server.
13:13 hdl ?
13:13 hdl I would like to test it.
13:15 kados it's important that the transition be seamless, obviously
13:15 kados it was pretty simple when we moved from to savannah because the savannah guys took care of all the hard stuff :-)
13:15 kados like history, etc.
13:16 hdl Yes.
13:16 hdl hi fbcit.
13:17 kados hdl: did you address MJ's question about Mailman?
13:18 hdl When he said we should change our mailing list software ?
13:19 kados well, I think he asked a question about mailman
13:19 hdl I had no time setting up a new one on our box and test it.
13:19 kados s a better list software than Mailman possible?  mlmmj, ecartis,
13:19 kados sympa and eoc all appear better in some ways.
13:19 kados Can the List-Id values be preserved if lists moved?  Last I saw,
13:19 kados Mailman couldn't cope with that (it had to be listname.domain), which
13:19 kados breaks some good email filters.  I think that should rule out Mailman
13:19 kados if we're considering moving lists.
13:21 kados that was from slef
13:21 kados slef/mj :-)
13:21 hdl I cannot answer him until I try to move the lists from savannah to our server.
13:21 kados OK
13:22 kados I am trying to get SSH access now
13:22 hdl I donot know which mailman version they use.
13:24 hdl OK.
13:26 hdl So. Should we have bug squashing time now. For 2 hours and then have an other one at 22PM GMT ?
13:29 kados OK
13:31 hdl Should we review bugs we already see last time ?
13:31 kados hdl: sure
13:31 hdl Should we fix them.
13:31 hdl Now.
13:31 kados probably :-)
13:31 hdl
13:31 hdl So here it is.
13:37 kados hdl: I have a few bugs to push up ... they may affect your bugfixing
13:37 hdl Then Push. I will rebase
13:40 kados hdl: your fix for 1898, you want me to ignore the first one, right?
13:40 hdl Yes.
13:40 hdl I included the first one in the second.
13:50 fbcit why is opac so slow?
13:51 kados fbcit: ?
13:52 fbcit it seems to crawl lately
13:52 kados fbcit: anything in particular that it's slow at?
13:53 fbcit kados: page loads
13:54 fbcit kados: 'volume number' is not used in opac that I can see
13:55 kados fbcit: OK, it should probably be added
13:55 kados fbcit: are you using remote or local YUI?
13:56 fbcit kados: local
13:58 kados fbcit: that could be the prob :-)
13:58 fbcit ?
13:58 kados fbcit: I think YUI servers have some special cacheing settings on the js libraries
13:58 fbcit it was slow before, using remove
13:58 fbcit remote also
13:58 kados oh
13:58 kados :/
13:59 fbcit it improved when I went local initially
13:59 fbcit but slowed down over time
13:59 kados weird
14:00 fbcit I notice that we set "Cache-Control" to no-cache in the headers of all pages
14:00 kados yea, that was actually a bug report
14:00 fbcit any specific reason
14:00 kados and should likely be a syspref
14:01 kados lemme track down the exchange
14:01 fbcit maybe a time to expire would be better
14:01 kados I think that only caches the HTML portion of the page
14:01 fbcit although some pages we might not want cached
14:03 kados fbcit:[…]w_bug.cgi?id=1373
14:03 kados ouch
14:04 kados fbcit: see if it's faster when you remove the lines:
14:04 kados -Pragma => 'no-cache', -'Cache-Control' => 'no-cache',
14:10 kados :)
14:11 fbcit this is painfully slow... 30-45 secs per page load
14:12 kados hmmm
14:12 kados fbcit:
14:12 kados fbcit: is running latest git
14:12 kados fbcit: any faster?
14:13 fbcit on the order of a magnitude faster aamof
14:14 kados ok, so it's server-side
14:14 kados what's your server load?
14:14 kados top should give you a good snapshot
14:15 fbcit load average: 0.14, 0.23, 0.20
14:15 fbcit cpu is 99.8% idle
14:15 fbcit very strange
14:16 fbcit 1GB ram and no paging file used
14:16 kados OK, so it must be the network
14:16 kados :)
14:16 fbcit heh
14:16 kados ntop
14:16 fbcit 0.037ms rtt
14:17 kados weird
14:18 kados have you restarted your browser lately?
14:18 fbcit so what am I looking for w/ntop?
14:18 fbcit no on browser
14:19 kados if a browser restart works, that indicates perhaps a memory or variable leak?
14:20 kados ntop I haven't used in 10 years :-)
14:20 kados well, 5 at least
14:20 kados so I can't recall ... I think it has some kind of web front end
14:20 kados wheren you can see trafic patterns
14:21 fbcit sounds interesting
14:24 fbcit ttyl
14:24 fbcit still slow btw
14:24 kados :(
14:24 fbcit I'll keep working on it
14:24 fbcit tnx
14:25 kados it could be mysql needs query cacheing adjusted
14:25 kados that'd be in the my.cnf file
14:25 kados key is to find the bottlekenck
14:25 kados bottleneck
14:27 fbcit tnx... on the phone
16:52 martinmorris afternoon all
16:52 martinmorris can anybody here advise if future editions of Koha will support FRBR?
19:01 frederic hello
19:03 atz_is_elsewher martinmorris: yes, FRBR is currently being developed
19:10 atz_is_elsewher gmcharlt might be able to tell you more details
21:08 hdl hi
21:12 atz hdl: you may be interested to see this:
21:12 atz
21:12 atz using Google Book jacket images (very slickly applied w/ js)
21:14 hdl atz : cool but will it be ok with French Books + what about :
21:15 atz I think Google's reach will be less than Amazon initially... but we will see
21:15 hdl a) books without ISBN or LCCN
21:15 atz what do we do presently w/ Amazon?
21:15 atz nothing, i think
21:16 hdl we display a bad no image for that book.
21:16 atz yeah, looks rather out of place
21:18 hdl Librarything code for displaying "fake" cover images could be good for that.
21:19 atz i have to read more about the Google API, but just looking at the HTML source for that demo page... it is very clean
21:20 atz w/ the Amazon module falling out of date (Amazon started enforcing new API), now would be a good time to look at this stuff again
21:20 atz hopefully I will be able to do so next week
21:24 frederic hi guys! It's funny you're talking about Google Book Search API. I played with it this afternoon.
21:24 frederic It's very easy to use.
21:24 frederic hdl: Est-ce que tu es intéressé à avoir les facettes en français en recherche ?
21:25 frederic J'ai proposé qque chose à kados. Je n'ai pas de nouvelle...
21:25 hdl Allez épate moi.
21:25 hdl ;)
21:25 frederic Et je ne vais pas rester bien longtement derrière mon écran :)
21:25 frederic J'ai envoyé qque chose mais je ne l'ai plus
21:25 hdl (english speaking here.)
21:26 hdl can you point URL ?
21:26 frederic yes naturlich.
21:26 frederic Concerning Google Search:
21:26 frederic[…]Aparkin%2C+robert
21:27 frederic kados: do you plan to integrate something to allow facets translation?
21:27 ryan hello #koha
21:27 atz facets currently operate via Zebra, as I understand it
21:28 frederic atz: I already sent a patch to allow translation of facet labels
21:29 hdl hi ryan.
21:29 frederic But it's not so easy to do since it's require someting like a configuration file which has to be copied in the appropriate directory with other zebra config files
21:30 atz ah, i see... not part of the usual translation process
21:31 frederic In other languages than English, it's definitely awful to have a result page with facet lables hard coded in English...
21:32 frederic hdl: There is another aspect: facets can be expand on each site. One site could decide to allow facet search on a new field. This only could be done via a config file (my opinion :))
21:35 hdl Well : good idea. But eventually, we should provide a way for the user to do this without having troubles with configuraiton files.
21:35 hdl have you sent your patch to paul ?
21:36 frederic Not sure but I think so
21:41 hdl atz : seems that facets are still not calculated with zebra today.
21:43 hdl ... But i may have overlooked the code.
21:45 frederic hdl: I confirm facets are not generated using zebra zebra but with a loop through biblio records result set
21:49 frederic I have to go. Good night (day).
22:08 hdl good night
22:25 hdl is there anyone ?
22:45 hdl kados ?
23:18 hdl Going to bed now.

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