IRC log for #koha, 2007-12-25

All times shown according to UTC.

Time S Nick Message
12:48 arsenic good morning!
13:35 fbcit g'morning koha
13:36 arsenic hi fcbit!
14:11 fbcit kados ping
14:14 fbcit morning arsenic
14:22 arsenic fcbit: How goes the windows' port?
14:53 kados morning all
15:00 fbcit arsenic_: there lack maybe 4 deps of having it up
15:00 fbcit one is a real challenge
15:00 fbcit libxslt
15:01 fbcit seems nobody has XML::LibXSLT running under strawberry
15:02 fbcit and I have not been able to get the activeperl version to work under strawberry
15:06 gmcharlt good morning #koha
15:06 arsenic 'morning gmcharlt
15:06 kados hey galen
16:05 fbcit morning gmcharlt
16:05 fbcit I sent you an email about libxslt
16:05 gmcharlt morning fbcit -- got the e-mail, thanks
16:24 fbcit can anyone recommend a quick and accurate way to test regex's before hand?
16:28 gmcharlt fbcit: test regexes for what?
16:28 fbcit results
16:29 gmcharlt fbcit: sorry, not understanding -- so you have a regex, and you want to test it -- throwaway script?
16:29 fbcit ie when I apply $foo =~ /[ ,].*$/ what is the regex returning for any given $foo
16:30 gmcharlt i.e., what string it's matching?
16:30 fbcit ie something like
16:31 fbcit yes
16:31 fbcit specifically that is
16:33 gmcharlt fbcit: you could use $& to see what was matched
16:33 gmcharlt but note that this should never appear in production code -- according to man perlvar, it causes a performance penalty
16:34 fbcit right
16:34 fbcit I'm looking for the equiv of sed -n -e '/foo/p'
16:34 fbcit I think that's it
16:34 fbcit I'm looking for the equiv of sed -n -e '/foo/p /foobar.txt'
16:34 fbcit I'm looking for the equiv of sed -n -e '/foo/p' /foobar.txt
16:34 fbcit even
16:37 fbcit gmcharlt: tnx that does exactly what I wanted it to
16:37 gmcharlt you're welcome
16:43 fbcit kados: are there a fixed number and nomenclature of zebra keywords?
16:43 kados it's configurable
16:43 fbcit :(
16:44 kados nozebra searching is supposed to suck ...
16:44 kados if that's any consolation :-)
16:44 kados that's why we switched to zebra :-)
16:49 owen Merry Christmas Eve, #koha
16:54 fbcit Merry Christmas Eve owen
16:55 gmcharlt hi owen
16:56 owen So I'm not the only one thinking about Koha today? :)
16:59 gmcharlt owen: no, just can't tear ourselves away :)
17:03 fbcit a NoZebra title search on 'Alexander the Great (NF)' results in a search limited by non-fiction...
17:04 fbcit NZanalyse has some serious problems methinks.
17:04 fbcit and lunch.
17:05 kados hehe
17:05 kados hey owen
17:05 kados owen: working on 670 today
17:06 owen A timeless classic.
17:07 owen I thought sometime we ought to talk in detail about the advanced search form.
17:07 kados owen: sure
17:09 kados owen: I'm thinking a mix between and 's advanced searches
17:09 kados maybe what we should do is have two separate templates
17:09 kados one 'clean' and one 'traditional'
17:10 kados (traditional woudl have boolean, all the subtype limits, etc.)
17:10 owen A simplified one for the general public, and a complex one for librarians and die-hards
17:10 kados even better
17:11 kados so basically it's just a matter of skinning opac-results.tmpl
17:11 owen Results?
17:11 kados oops
17:11 kados opac-advsearch.tmpl
18:49 fbcit gmcharlt: if I put Koha 3 into production w/installer->standard
18:49 fbcit is updating just a matter of running installer again w/same options?
18:49 fbcit ie non-git
18:50 kados yea
18:50 kados so long as you haven't modifed any files in the tree
18:50 kados install tree I mean
18:51 kados it would be worth testing though
18:51 kados if it overwrites the files
18:51 kados ie, make a change and see if it overwrites the file
18:51 kados long-term of course, we want to be all fancy, back up the user's install, warn them, force them to sign in blood before upgrading, etc. :-)
18:51 fbcit right :-)
18:52 fbcit I'm mainly wanting to get on with my production install while keeping up with changes
18:53 fbcit before beginning
18:53 kados hehe
18:56 fbcit kados: is there a way to get apt-get to tell what has been installed?
19:14 gmcharlt fbcit: should work as long as you haven't been playing with timestamps in your production tree
19:27 fbcit hrmm
19:27 fbcit kados
19:27 fbcit Errors were encountered while processing:
19:27 fbcit libxml-sax-perl
19:27 fbcit libxml-libxml-perl
19:27 fbcit libxml-libxslt-perl
19:27 fbcit libxml-simple-perl
19:27 fbcit E: Sub-process /usr/bin/dpkg returned an error code (1)
19:28 fbcit using 'dpkg --set-selections < install_misc/debian.packages'
19:28 fbcit ok
19:29 fbcit yup
19:29 kados that's gonna be problematic
19:29 kados I should add anote to the install doc
19:29 kados libxml is really a problem!
19:29 kados someome oughta fix it :-)
19:30 kados basically, you want to run the package maintainer's versions of libxlm2 and libxxml-libxlm-perl (and associated)
19:30 kados because they are compatible
19:30 kados the cpan version isn't compatible with debian etch's version fo libxlm2 (C)
19:35 kados s/bib/big/
19:37 fbcit kados: I added a note to README.debian you want the patch direct?
19:38 gmcharlt kados: now have converting from MARC-8 to UTF-8 correctly
19:45 kados awesome
19:45 kados fbcit: you can route it through patches@ if it's non-critical
19:45 kados gmcharlt: what was the trick?
19:45 gmcharlt basically, forcing use of NFC
19:46 kados oh, interesting
19:46 kados did you find that bit of code .../me can't remember where I hid it :-)
19:46 kados maybe?
19:46 gmcharlt that gets the german example displaying in Firefox correctly
19:46 kados _entity_encode
19:46 kados in
19:46 kados FWIW
19:47 kados y'know, I think there's a problem with firefox on OSX
19:47 kados for certain fonts
19:47 kados but maybe your NFC fixes that ...
19:48 gmcharlt yeah, ran into something saying the Arial does not handle characters with decomposed diacritics correctly
19:48 gmcharlt hence why NFC helps
19:48 kados I always check with safari
19:48 kados cool
19:48 kados it'd be good to document that
19:48 kados[…]ncodingscratchpad
19:48 kados is as good a place as any
19:49 kados Web Browser section
19:50 kados gmcharlt: so you convert to NFC before import?
19:50 gmcharlt yes -- I revived the fMARC8toUTF8 functin that someone (thd?) had started
19:51 gmcharlt it's bulkmarcimport only, for now
19:51 kados *nod*
19:59 gmcharlt icky, icky -- prior to this change, biblioitems.marc would be left in MARC-8 but biblioitems.marcxml would be in UTF-8 (presumably because MARC::File::XML was doing an automatic character conversion when the XML was generated)
19:59 kados ok, here's an SQL question gmcharlt  ... say I have the cn_sort value from one item, and I want to find the two before and after ?
19:59 kados yea, that's messy
20:00 gmcharlt select cn_sort from item where cn_sort <= X limit 2?
20:00 gmcharlt select cn_sort from item where cn_sort <= X order by cn_sort desc limit 2?
20:01 kados brilliant, thanks
20:02 kados gmcharlt: I notice there's no imposed uniqueness on cn_sort, is that on purpose?
20:03 gmcharlt yes -- can't guarantee it, since nothing prevents a library from having duplicate call numbers
20:03 gmcharlt i.e., a lot of small public libraries don't insist on uniqueness
20:03 gmcharlt they're probably a bit misguided IMO, but that's the practice
20:03 kados :)
20:04 gmcharlt false_economy--
20:04 kados I guess it means there's no way to truly identify shelf order
20:04 kados unless we just 'fake it'
20:04 gmcharlt kados: yeah, you pretty much have to fake it, at least in the general case
20:05 gmcharlt a library could provide lots of specific info about which location codes correspond to which physical shelves, but that's not an instant thing to implement
20:05 kados *nod*
20:06 gmcharlt although a feature to ease management of such library maps would be nice (and distinctive)
20:07 kados *nod*
20:08 kados VM acting up again ... :/
21:03 fbcit Merry Christmas koha.
21:04 gmcharlt Merry Christmas fbcit

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