IRC log for #koha, 2007-08-24

All times shown according to UTC.

Time S Nick Message
12:29 hdl hi there
12:56 kados hi hdl
12:56 kados hdl: owen has been good about filing bugs for 3.0 :-)
12:57 kados hdl: seen them?
12:57 dewey I haven't seen 'them', kados
12:57 hdl I did
12:58 hdl Are there "only 15" ?
12:58 hdl on rel_3_0 ?
13:10 kados I searched HEAD and rel_3_0 and I get 87
13:11 kados some are enhancements though
13:11 kados our bugzilla needs to be updated :/
13:43 kados foxnorth: you around?
13:59 foxnorth hey kados-- you still around?
14:02 kados foxnorth: yep
14:02 kados foxnorth:  should be around all day
14:02 foxnorth cool
14:03 foxnorth just checking your debugging dump from saving/resaving
14:04 kados yea, it's a pretty strange bug
14:04 kados almost seems random
14:05 foxnorth yeah the random part is what scares me :)
14:05 kados heh
14:05 kados have you been able to reproduce it at all?
14:06 foxnorth i tried saving/resaving yesterday when trying to reproduce it.  Let me try again and see what happens...
14:07 foxnorth given that it's the leader and 008 which were preserved in that last mangled record, the variable fields editor data must not have been picked up by the save function
14:07 kados ahh, maybe that's the key
14:07 foxnorth i.e. there was a problem accessing the yui editor's contents
14:07 kados yea, just happened again for me
14:08 foxnorth hhmm
14:08 kados slightly different this time though
14:08 kados it displayed properly in the Drafts folder, but clicking on it resulted in a one line record:
14:08 kados 008  981203s2000 cau e b 001 0 eng d
14:08 foxnorth it displayed properly in the drafts folder??
14:08 foxnorth hhm
14:08 kados yea, still is actually
14:08 kados but the preview pane has:
14:08 kados 000 00793nam a2200241 4500
14:09 kados 008981203s2000 cau e b 001 0 eng d
14:09 foxnorth hhmm
14:10 foxnorth ok my xml in plain english just displayed ok.  going to try saving/resaving
14:10 foxnorth so far repeatedly hitting "save" is ok.
14:11 kados try editing something in it
14:11 kados yea, that will work
14:11 kados edit something
14:11 kados click save
14:11 kados then click on 'Drafts' to get the record list
14:11 kados then double-click on the record again
14:11 kados edit it
14:11 kados save it, repeat
14:11 kados it's leaving the editor that does it I suspect
14:11 foxnorth ok editing
14:12 foxnorth sorry, i'm going to try editing...
14:12 kados *nod*
14:13 foxnorth can't reproduce!  arg
14:13 kados try again
14:13 kados :-)
14:13 kados sometimes it takes 3-4 times to happen for me
14:13 kados this is webcat-dev, right?
14:14 foxnorth yep
14:14 foxnorth just saved/resaved and edited fixed field and variable fields and no nice :(
14:14 foxnorth just out of curiosity, want to query the sqlite db manually?
14:14 foxnorth i forget where gears sticks it on mac?
14:21 foxnorth Google Gears for firefox on mac db location: Location: Users/<username>/Library/Caches/Firef​ox/Profiles/{profile}.default/Google Gears for Firefox
14:31 foxnorth brb
14:35 kados foxnorth: so you were able to reproduce the prob, or not?
14:36 foxnorth kados: not able to.  I saved/resaved as many times as i could!
14:36 foxnorth I'm just trying to think how we could debug this thing
14:36 foxnorth i forget, do you have linux running under vmware or something?
14:37 foxnorth that you could test on to make sure it's not ff for mac?
14:38 kados weird
14:38 kados I don't at the moment
14:38 foxnorth ok
14:39 foxnorth from your debug output I can't tell *why* the yui editor contents didn't get picked back up
14:39 kados yea
14:39 foxnorth you could run it under the firebug debugger
14:39 foxnorth if you've got the time, that is
14:41 foxnorth Click firebug script, choose edit.js and break at line 18 then try saving
14:42 foxnorth actually, you could break at line 86 instead since we know the fixed fields are ok
14:42 foxnorth and check the value of var_ed
14:42 kados hang on, phone :/
14:42 foxnorth np
18:35 kados hdl: are you still here?
18:35 kados hdl: I have a question for you regarding how to handle English UNIMARC frameworks in the installer
18:35 kados fr 3.0
20:03 chris morning
20:06 kados hey chris
20:07 hdl hi kados
20:07 hdl fire.
20:08 hdl kados ?
20:08 dewey kados is champion alligator-wrestler in 7 counties
20:08 chris hehe
20:08 hdl hi chris
20:09 hdl has kados solved his question ?
20:09 chris i dont know, im back in nz, i cant look over his shoulder anymore :)
20:10 hdl OK.
20:10 hdl have a nice meeting then
20:10 chris i think he probably got distracted by a phonecall
20:10 chris *nod*
20:11 hdl kados : send me your question via email.
20:11 hdl I shall answer tomorrow.
20:15 chris push coming your way kados
20:17 kados cool
20:29 hdl kados : i am about for a little while
20:34 kados hi hdl
20:34 kados I just was wondering how you envision the installer handling a unimarc in english?
20:34 kados at what step should it be picked?
20:35 kados (since right now, it's unimarc vs marc21)
20:35 hdl Yes I was anxious you ask that.
20:36 hdl Indeed. no clear solution at the moment.
20:37 hdl Maybe have USMARC and UNIMARC in required directory.
20:38 hdl Or ask for the user to choose one among many.
20:38 hdl I thought taht we could have languages in our framework structure anyway
20:39 hdl may old xml temptation comes back.
20:39 kados right
20:39 kados hmmm
20:40 hdl maybe we could/should add a language field to marc_subfield_structure and have all the translations in sql file.
20:40 hdl marc_subfield_structure  and marc_tag_structure.
20:41 kados what about a separate table for translations of the name?
20:41 kados since it's a one to many relationship
20:41 kados one entry -> many translations
20:42 hdl well... from a strictly dbdesign point of view I agree.
20:42 hdl From a koha3.0 out soon I utterly disagree.
20:42 hdl (And I hope you understand)
20:42 kados yep
20:43 kados and I agree
20:43 kados not for 3.0, maybe for 3.2
20:43 hdl definitely
20:44 hdl For now, maybe we could add a directory that would need the user to choose MARC Flavor...
20:44 hdl That would require a simple change in installer.
20:44 kados good idea
20:45 kados it comes up because I have a unimarc english
20:46 hdl I understand. And I think we need multilanguage *marc structures in the future.
20:47 hdl But atm, the sole option is, imho, what I suggested you : adding a new marcflavour directory before language selection.
20:47 hdl Could you ask someone to do that ?
20:48 kados I don't have anyone who can
20:49 kados maybe for now it's just note that you ahve to pick one or the other
20:50 hdl OK I shall.
20:51 hdl maybe note in enhancements on buzilla ?
20:51 hdl and assign to me so that i donot forget.
20:52 hdl nite #koha
22:46 thd kados: are you there?
22:52 thd chris: I have a merge conflict in git.  Would you help me sort it?
22:58 chris can do
22:58 chris what file is it in?
23:03 thd chis: CONFLICT (rename/modify): Merge conflict in misc/marc_datas/marc21_en/structure_def.sql
23:03 thd Failed to merge in the changes.
23:03 thd chris: that file should not even be in HEAD
23:03 thd as far as I know
23:03 chris there is no HEAD anymore :) thats a cvs'ism
23:04 thd what is the non-CVSism for HEAD?
23:04 chris master
23:05 chris and you got this by doing a git fetch
23:05 thd chris: that file should not even be in master.  As far as I know it is a deprecated filename
23:05 chris git rebase origin
23:05 chris ?
23:05 thd yes, I hope that I did not put it there
23:07 chris yep that file exists for me still
23:07 chris ah ha
23:07 chris but when i do a
23:07 chris git checkout -f master
23:08 chris its gone
23:08 thd I definitely did not put it there but it has contents which should not be there
23:08 chris as it should
23:08 thd chris so what should I do to make it disappear
23:08 chris do you have any uncommitted changes?
23:08 chris if not
23:08 chris try
23:09 chris git checkout -f master
23:09 thd chris: I have no locally uncommitted changes
23:10 chris give the checkout -f a whirl then
23:11 thd chris: I do have 2 patches in the directory representing changes which I committed.  I sent those patches to you.  What should I do with those patches in the directory?
23:11 chris you can safely ignore them
23:11 chris or type git clean
23:11 chris and that will get rid of them
23:12 thd chris: so I will run git clean and then git checkout -f master
23:12 chris yep, its worth a try :)
23:16 thd chris: should git fetch and git rebase origin work now?
23:16 chris im hoping so
23:25 thd chris: everything seems good now.  I had to remove .dotest recursively because git complained that it already existed
23:25 chris ah yep
23:25 chris cool it sorted
23:26 thd chris: I have one other question
23:26 chris yep?
23:27 thd chris: when I first tried setting up gitweb using a symlink as per the instructions in the documentation did not work
23:27 chris right, probably needed to let apache know its ok to follow symlinks
23:28 thd chris: gitweb gave a 403 forbidden error no matter what the user and permissions were and apache was well informed with a directive for the directory
23:29 thd chris: I presume that you are not using a symlink for
23:35 chris yep we are
23:36 thd chris: what version of Apache are you using?
23:36 chris lrwxrwxrwx 1 root root 26 2007-08-03 15:55 Koha -> /pub/scm/koha-web.git/.git
23:37 chris 2
23:38 thd chris are you using the MPM-prefork version which can manage non-threaded libraries?
23:41 chris pass
23:41 thd :)
23:41 chris probably just worker
23:41 chris thats what i usually install
23:44 thd chris: is there anything wrong with my directive ...
23:44 thd    <Directory "/var/cache/git/">
23:44 thd        AllowOverride None
23:45 thd        Options  +ExecCGI -MultiViews +FollowSymLinks
23:45 thd        Order allow,deny
23:45 thd        Allow from all
23:45 thd    </Directory>
23:45 chris shouldnt be, i dont even have that
23:46 chris i just have gitweb.cgi in /usr/lib/cgi-bin
23:46 chris and
23:46 chris ScriptAlias /cgi-bin/ /usr/lib/cgi-bin/
23:48 thd chris: you allow FollowSymLinks globally?
23:48 chris i misled you
23:48 chris apache doesnt need to follow the symlink
23:48 thd :)
23:48 chris its gitweb.cgi
23:48 chris that uses it
23:48 chris so you should be hitting http://yourserver/cgi-bin/gitweb.cgi
23:49 chris and it does the rest
23:51 thd chris: I told Apache where to find gitweb.cgi in my internal domain ...
23:51 thd <VirtualHost git.agogmel1.lan:80>
23:51 thd    ServerAdmin webmaster@hal7000m.agogmel1.lan
23:51 thd    ServerName git.agogmel1.lan
23:51 thd    HeaderName HEADER
23:51 thd    # bogus but safe DocumentRoot
23:51 thd    DocumentRoot /var/cache/git
23:51 thd    # ErrorLog /var/log/apache-ssl/git.agogmel1.lan-error.log
23:51 thd    # CustomLog /var/log/apache-ssl/git.agogmel1.lan-access.log combined
23:51 thd    Alias /robots.txt /var/www/cvs.robots.txt
23:51 thd    Alias /gitweb.css /var/www/gitweb.css
23:51 thd    Alias /git /var/git
23:51 thd    # Alias /koha_thd/ /home/thomasd/software/bibliographic/​ils/koha/checkout/HEAD/git/koha_thd/
23:51 thd    # ScriptAlias / /usr/lib/cgi-bin/gitweb.cgi
23:51 thd    Alias / /usr/lib/cgi-bin/gitweb.cgi
23:51 thd    RedirectMatch permanent "^/~(.*)$" "http://agogmel1.lan/~$1"
23:51 thd    <Directory "/var/cache/git/">
23:51 thd        AllowOverride None
23:51 thd        Options  +ExecCGI -MultiViews +FollowSymLinks
23:51 thd        Order allow,deny
23:51 thd        Allow from all
23:51 thd    </Directory>
23:52 thd </VirtualHost>
23:52 chris right theres you problem
23:52 thd where?
23:52 chris Alias / /usr/lib/cgi-bin/gitweb.cgi
23:52 chris but you havent allowed +ExecCGI
23:52 chris on /usr/lib/cgi-bin
23:53 thd chris: that works if the actual files are in /var/cache/git
23:53 chris right
23:53 chris cos you have +ExecCGI for /var/cache/git
23:53 chris but not for /usr/lib/cgi-bin
23:53 thd oh
23:53 chris so apache says it doesnt have permission to exec it in that dir
23:54 chris that make sense?
23:54 thd yes thanks
23:54 chris cool

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