IRC log for #koha, 2007-12-21

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

All times shown according to UTC.

Time Nick Message
12:52 fbcit g'morning koha
13:31 kados hiya fbcit, owen
13:31 owen Hi kados
13:31 kados hdl: I still haven't seen your patches
13:31 hdl: please send them again
13:32 fbcit hi kados
13:32 kados: I ordered it from posguys
13:32 kados fbcit: :-)
13:34 fbcit kados: I assume it has drivers for Debian?
13:34 kados it's just a keyboard wedge, shouldn't need any fancy drivers
13:35 you got the USB one?
13:35 any standard USB keyboard driver should suffice
13:36 fbcit USB
13:36 w/stand :)
13:37 kados: is self-checkout stable?
13:37 hdl kados : since when ?
13:50 kados fbcit: well, it was stable in rel_2_2, deleted from 3.0, added back to 3.0 with some api changes
13:50 fbcit: so it needs a good testing
13:51 hdl: since when what?
13:51 owen Here's a fun variable collision: The patron expiration date output to the screen in circulation.pl is used by the script as the due date for the checkouts
13:51 kados hdl: all patches you've sent to me I've applied, so any that haven't been applied should be re-sent
13:52 owen: *cough*
13:52 fbcit owen:heh
13:55 hdl owen : this seems a bug of WarnBeforeExpiry
13:55 owen chris just committed changes to that so I'm assuming the bug was introduced then
13:55 kados sorry I missed that one :/
13:57 owen Hard to spot unless you're paying attention to what due dates are coming through for new checkouts
13:58 kados yea, nice catch owen
13:59 hdl: http://git.koha.org
13:59 hdl: I see two patches applied from you last night
14:02 owen: 1673, looks like atz's refactoring has caused some issues ...
14:02 owen: can you try going:
14:02 git revert 14eb56e8dd78e38727363e97fbff20917380abaf
14:02 and see if that fixes the new bugs?
14:03 well ...
14:03 I'd do that in a new clone
14:05 owen No luck
14:05 kados k, we can't blame joe then :-)
14:06 well, maybe we can
14:06 owen Well, maybe we just don't know /why/ it's joe's fault ;)
14:06 kados there are some other patches from him earlier
14:06 hehe
14:15 hdl kados : patches sent
14:16 kados hdl: thanks
14:16 hdl But it is strange you did not received them the first time.
14:17 kados hdl: one is already applied unless I'm mistaken
14:17 two in fact
14:17 BugFix : authorities resultlist display was missing ...
14:17 hdl maybe I have not rebased this morning.
14:25 fbcit in opac adv search should I be able to select an itemtype, leave all keyword fields blank and expect the search to return all records of the itemtype selected?
14:30 owen That's how I would expect it to work
14:31 kados fbcit: yes
14:31 fbcit: if that doesn't work in the opac, test in the staff client, and if it works there, let me know
14:31 fbcit: well, let me know regardless
14:31 fbcit: well, if it's nozebra, let hdl/paul/mason know :-)
14:38 gmcharlt good morning #koha
14:40 kados hiya gmcharlt
14:55 fbcit hdl: what does the 'mc' "operator" translate into in plain English?
14:55 hdl material code ?
14:56 itype
14:56 fbcit it appears in the CGI string of an advanced search on itemtype ie. mc=BIO
14:57 tnx
15:20 owen kados: Got your email, those statuses all sound correct to me.
15:40 kados owen: sweet, thanks
15:45 hdl still around?
15:45 hdl yes
15:45 kados hdl: I could use some help understanding what french libraries would like to see in the staff-side results page for items
15:45 hdl: http://search.athenscounty.lib[…]ch?q=harry+potter
15:45 hdl: I will explain what that display shows and you can tell me what's missing
15:46 hdl: Copies available at:  Wells (Albany) (1),  Athens (1),  Glouster (1),  On loan (1), Lost (2)
15:46 hdl: the branches listed with counts mean that those items are available
15:46 hdl: the statuses marked in red "on loan, lost" mean those items aren't available
15:46 so in that example, there are 6 total copies
15:47 owen: if you have any thoughts on what the staff-side should show that'd also be helpful
15:49 hdl what was in rel2_2 was : return date of item checked out.
15:49 kados for the onloan items
15:49 how about call numbers?
15:49 hdl yes.
15:49 kados for all items, or just available ones?
15:50 hdl only available ones.
15:50 kados ok
15:50 anything else?
15:51 hdl I think it is a pity we donot know which library owned the book issued
15:52 kados we can know that
15:55 hdl: thanks
15:55 hdl: as I fix bug 670 I'll do my best to include the data you've mentioned
16:14 fbcit kados: some things are definitely screwed up with adv search in NoZebra...
16:15 :-(
16:15 paul hello world, from the TGV station in paris !
16:15 kados paul, hdl, maybe you can help?
16:15 hehe
16:15 hi paul!
16:15 TGV++
16:15 paul (connected through wifi)
16:15 kados oh, just the station
16:16 hdl it's free.
16:16 kados paul: do you have an account with one of the wifi providers? or did you find a free one?
16:16 oh!, great!
16:16 paul kados : I just sent some patches
16:16 kados TGV+++
16:16 paul kados : it's free if you have DSL at home through one of the provider that has an agreement with SNCF.
16:17 and I have a good provider ;-)
16:17 hdl fbcit: I have no problems with advsearch in NoZebra.
16:17 paul ++ : no problems to git fetch, which means their firewall is not too closed
16:17 so, now, the question, where can I help ?
16:17 hdl can you detail ?
16:18 kados fbcit: ?
16:18 fbcit hdl, paul: There are at least two cases so far...
16:18 first,
16:18 limit by itemtype and enter no keyword/etc params
16:18 second
16:19 enter a known good keyword/etc param and select an itemtype that does not match
16:19 the first case should return all records of the selected itemtype
16:19 it does not
16:20 the second should return no records found,
16:20 it does, but...
16:20 the query string appears to be all messed up in DEBUG
16:20 ie QUERY:ti,wrdl= alexander and (mc=ABIO)and (mc=ABIO) at /usr/share/kohaclone1/C4/Search.pm line 942.
16:20 when the limit is only entered once in the adv search page
16:21 which appears to cause not nice things like: EXECUTE : biblioserver, mc=ABIO, ABIO at /usr/share/kohaclone1/C4/Search.pm line 1382.
16:22 hdl fbcit: this is adressed with patches I sent kados today.
16:22 (I re-sent)
16:22 fbcit hrmm
16:22 I rebased this morning...
16:22 hdl I think
16:22 fbcit will again now...
16:22 if need be
16:22 hdl maybe he hasnot pushed them
16:24 kados hdl: which patches specifically?
16:24 hdl SearchNZ
16:24 kados Adapting things to new API ?
16:24 hdl BugFixing NoZebraSearch
16:26 No. this one could be redundant with  some of your devs. But is necessary to get all subtypes_unimarc translated
16:26 kados the  Adapting things to new API one I see some probs
16:26 hdl: for instance, availability limits
16:27 hmmm, maybe it's just moved
16:27 hdl yes : it is moving things down.
16:28 to get all the 'regular' stuff up
16:56 paul Leaving to take my TGV to go back home.
16:56 see you tomorrow
18:09 chris owne: you about?
18:09 owen even
18:09 owen Yes
18:09 chris 1427 thats by design
18:10 if you have the syspref turned on, thats what its designed to do
18:10 (issue to expiry date)
18:10 so might be, just need to turn the syspref off?
18:11 # if ReturnBeforeExpiry ON the datedue can't be after borrower expirydate                                                                          
18:11            if ( C4::Context->preference('ReturnBeforeExpiry') && $dateduef->output('iso') gt $borrower->{dateexpiry} ) {                                  
18:11                $dateduef = C4::Dates->new($borrower->{dateexpiry},'iso');                                                                                
18:11            }  
18:11 if you have a sec, can you try turning off returnbeforeexpiry and see if that is ok
18:12 if so we can reclose 1427
18:14 owen chris: I see, I expected the system to refuse to let me check out if ReturnBeforeExpiry is on, but the way it works makes sense
18:15 I guess maybe there needs to be a message that points out the non-standard due date?
18:15 chris yeah the sysprefs need better explaination for the most part
18:15 yeah, lets add that to the bug, and lower the severity
18:16 i can go update the bug if you want?
18:16 owen Of course
18:18 chris i wonder if we could put it in the same place as the alert they are about to expire?
18:18 this patron's card is due to expire ... since you have return before expiry on, the due date will be set to their expiry date
18:19 maybe?
18:45 atz owen: i was looking at some CSS
18:45 owen Yes?
18:55 atz firebug reports some warnings
18:55 some made sense to me  (supplying only 3 arguments to padding, for example)
18:56 but a lot did not ('Expected declaration but found "/"')
18:57 just wondered if you had seen them already, blame firebug, or what...
19:17 tim kados, you around?
19:24 owen, my problem yesterday wasn't a template problem, but I sure caused a few problems with the template before I got it working the way I wanted.
19:40 masonj morning #koha
19:41 fbcit hi masonj
19:42 what's it like in tomorrow?
19:42 masonj hiya fbcit
19:42 very forward thinking..
19:43 think blade-runner, but without daryl hannah
19:47 fbcit well, if you need it done yesterday send it to me... ;-)
19:49 gmcharlt: fix-perl-path.PL now behaves very badly on XP with very long filenames like auth_subfields_structure.pl
19:50 gmcharlt fbict: the version using attrib -r?
19:50 fbcit either version
19:50 when it hits such a file, it dumps the entire contents to the console
19:51 gmcharlt koha: not just open source: we'll *make* you look at the source
19:51 fbcit hehe
19:51 "in-your-face" source?
19:52 gmcharlt ha
19:52 fbcit hdl: u around?
19:57 hdl yes
19:58 fbcit:
19:58 fbcit hdl: it appears that some problems still exist after applying your patches
19:59 there seems to be some buggy recursion going on inside of NZanalyse
19:59 at any rate...
19:59 I see: [Thu Dec 20 15:47:21 2007] search.pl:  / = / NF
19:59 [Thu Dec 20 15:47:21 2007] search.pl: EXECUTE : biblioserver, , NF at /usr/share/kohaclone1/C4/Search.pm line 1377.
20:00 which I think causes $results to return empty
20:00 I do not yet have a full understanding of how NZanalyse works in all cases, though.
20:01 hdl what search have you done ?
20:01 fbcit it does need         $left='itemtype' if $left =~ '^mc$';
20:01 added
20:01 hdl yes.
20:01 It was called mt before.
20:01 fbcit just now I have done a search on an itemtype && title word that I know exist
20:01 and search says nothing found
20:02 masonj fbcit, ive been poking at the same bug yesterday
20:02 fbcit but that is after it performs the EXECUTE mentioned above
20:02 for some reason unknown to me yet, NZanalyse recurses after returning $return
20:02 for no apparent reason
20:02 masonj looks like hdl has a patch for the buggy AND search in  NZAnalyze()
20:06 fbcit hdl: can you send me that patch?
20:06 or is it commited already
20:07 I've got to go down for reboot... brb
20:13 hdl fbcit: sent.
20:13 fbcit tnx
20:14 hdl but it was already pushed today I think.
20:14 I turned  $left='itemtype' if $left =~ '^mc$'
20:14 fbcit hdl: I have an itemtype NF for Non-fiction
20:15 hdl by adding mc index for 200$b (UNIMARC)
20:16 Then you should not have NF alone. But along with an index.
20:17 masonj hdl, just testing your nozeb search pach , looks good :)
20:18 but i had to add it manually to Search.pm, as  it was too old to apply :/
20:18 hdl, ill send a new patch of your commit to chris now
20:27 fbcit hdl: try this:
20:27 in adv search
20:27 select an itemtype you know is in the db
20:27 search only on that
20:27 nothing else
20:33 gmcharlt fbcit: did first full test on strawberry: after removing checks on $^O and just using stat and chmod, fix-perl-path.PL is converting each *.pl to a 36-byte file containing only the shebang
20:34 masonj hmm, that search fails for me too fbcit
20:34 fbcit sounds like another form of 'rm -fr /' to me... :)
20:34 masonj: there is something broke in NZanalyse, I believe... :)
20:34 look:
20:34 [Thu Dec 20 16:24:03 2007] search.pl:  (mc=NF) at /usr/lib/cgi-bin/kohaclone1/catalogue/search.pl line 462.
20:34 [Thu Dec 20 16:24:03 2007] search.pl: dealing w/parenthesis. left :mc=NF at /usr/share/kohaclone1/C4/Search.pm line 1239.
20:34 [Thu Dec 20 16:24:03 2007] search.pl: itemtype / = / NF
20:34 [Thu Dec 20 16:24:03 2007] search.pl: EXECUTE : biblioserver, itemtype, NF at /usr/share/kohaclone1/C4/Search.pm line 1377.
20:34 [Thu Dec 20 16:24:03 2007] search.pl: result : nf == at /usr/share/kohaclone1/C4/Search.pm line 1383.
20:34 [Thu Dec 20 16:24:03 2007] search.pl:  / = / NF
20:35 [Thu Dec 20 16:24:03 2007] search.pl: EXECUTE : biblioserver, , NF at /usr/share/kohaclone1/C4/Search.pm line 1377.
20:35 why does it recurse the second time?
20:35 I cannot figure out what calls it.. :-\
20:35 masonj its doing right and left searches, i think
20:35 fbcit but the second time returns an empty $return
20:36 masonj its tricky/scary  for me too :)
20:36 fbcit :)
20:37 I had it 'warn'ed like mad earlier and it looks like it covers the same code twice.
20:37 masonj gimme a little while for a look at your  itemtype search
20:37 morning brain still warming up here
20:38 so fbcit, your other AND searches are working ok?
20:38 fbcit one moment... phone
20:38 masonj my other ones are looking ok...
20:38 1 results found for 'ti: love and au: caroline and su: emotional'
20:40 hdl it fails for me too.
20:41 will investiguate and fix this tomorrow.
20:41 too late today.
20:42 fbcit masonj hdl: they have been
20:42 nope
20:42 No results match your search for 'ti,wrdl: alexander' with limit(s): 'mc:NF'
20:43 scratch that...
20:43 1 result(s) found for 'ti,wrdl: alexander and kw,wrdl: greece'
20:49 hdl fbcit: ?
20:50 Is alexander greece an NF ?
20:50 fbcit yes
20:50 hdl: it appears to be a limit issue
20:51 masonj snap
20:52 fbcit hdl: actually the first pass through NZanalyse appears to retrieve the records ok, but the second pass munges up the query and returns no results.
20:53 look at this debug:
20:53 [Thu Dec 20 16:39:28 2007] search.pl: return : 2,thedecline-1;3,thebookoft-1;4,alexandert-1;5​,thesearchf-1;9,aliterarys-1;10,famousbrit-1;1​1,godismycop-1;1,alexandert-1;13,aparsinggu-1; for LEAF : mc=NF at /usr/share/kohaclone1/C4/Search.pm line 1443.
20:53 [Thu Dec 20 16:39:28 2007] search.pl: string : (mc=NF) at /usr/share/kohaclone1/C4/Search.pm line 1283.
20:54 line 1443 is right before we exit the sub.
20:54 am I overlooking the loop code?
20:55 what takes us back to line 1283..
21:02 hdl damned! Why can it work on an installation I have done last week .
21:09 masonj hdl: my 2 cents, i have made 2 commits to NZAnalyze() sub, in the last week that may have broken it
21:10 two 1 line commits..
21:10 fbcit just for my own understanding, why does $query get $limit appended to it?
21:10 if $query is empty, then $query eq $limits
21:10 would this cause a problem?
21:11 hdl masonj, fbcit : I send you the files which worked for me tonight.
21:11 same base.
21:11 masonj yeah, sounds good
21:11 hdl Then you can diff with your files.
21:12 fbcit ++
21:12 hdl time to leave you.
21:12 nite
21:13 fbcit g'night hdl
21:13 masonj au revoir
21:37 fbcit masonj: anything stand out?
21:37 masonj hmm not yet :/
21:38 ive got hdl's email ,and have tested this working files on my dev install
21:38 fbcit do his files work on your install?
21:38 masonj not yet
21:39 fbcit I diff'd them, but do not see any earth shaking changes..
21:40 gmcharlt fbcit: just e-mailed you a patch (against the main repo version) to fix-perl-path.PL that appears to work now on WinXP
21:41 please give it a try when you have a chance
21:41 fbcit I'll take it home tonight and try it after I get my eyes uncrossed... 8-P
21:47 gmcharlt but the bit of grit becomes so pretty eventually ;-)
21:56 fbcit masonj: it looks like buildQuery appends the limit to the query so...
21:56 if there is no query, the limit becomes the query...
21:56 then both $query and $limit are passed to NZanalyse
21:57 NZanalyse checks for both the $query and $limit separately in this case
21:57 only it checks $query contents last
21:57 because $query has a space appended before $limit is, this breaks the left split routine
21:58 and thus the last SELECT returns no results
21:58 so $results is blank and here we go...
21:58 ...how to fix...and not break other things?
21:59 masonj nice debugging
21:59 fbcit in this case $query eq " (mc=NF)" while $limit eq "(mc=NF)"
22:00 I *think* in this case $query eq "" would be correct?
22:07 interesting...
22:08 it appears that $limit disappears into the bit bucket somewhere after we call buildQuery... ??
22:15 masonj: are the limits applied after the result sets are obtained and before they are displayed?
22:17 why is the limit appended to the query?
22:17 gmcharlt? kados? anyone?
22:17 chris
22:18 masonj it looks like the limiting is handled by nzanalyze()
22:18 ti= love and (mc=BK)
22:18 fbcit ahhh, so that is why $query .= " $limit";
22:18 masonj looks like the limit is formatted with a (*)
22:18 fbcit but that breaks bad in this case
22:19 seems that really $limit should be passed into NZanalyse separately
22:21 masonj: does it appear to you that NZanalyse then turns the limit into a db query?
22:22 masonj yeah, NZanalyse doesnt seem to handle  limit arg any different to and 'AND'
22:22 s/and/an/
22:24 fbcit I notice in search.pl the following note:
22:24 # At this point, each server has given us a result set
22:24 # now we build that set for template display
22:25 maybe NZanalyse simply does SELECT's based on everything in $query and then it is up to search.pl to sort it out and apply limits?
22:33 I think the recursion logic in NZanalyse is faulty...
22:34 causing double processing of $string at times
22:34 but I have to go to Wal-Mart... :(
22:35 masonj we'll get hdl and paul having a look later today
22:35 so it will be sorted real soon
22:49 fbcit methinks this regex is the culprit: $string =~ /(.*)(>|<|=)(.*)/;
22:52 got it...
22:52 heh
22:53 $string =~ s/-|\.|\?|,|;|!|'|\(|\)|\[|\]|{|}|"|&|\+|\*|\// /g;
22:53 should read
22:53 $string =~ s/-|\.|\?|,|;|!|'|\(|\)|\[|\]|{|}|"|&|\+|\*|\///g;
22:53 actually...
22:53 masonj: still there?
22:54 masonj yep
22:54 fbcit try this...
22:54 # it's a leaf, do the real SQL query and return the result
22:54    } else {
22:54        $string =~  s/__X__/"$commacontent"/ if $commacontent;
22:54        $string =~ s/-|\.|\?|,|;|!|'|\(|\)|\[|\]|{|}|"|&|\+|\*|\///g;
22:54        warn "leaf:$string" if $DEBUG;
22:54        # parse the string in in operator/operand/value again
22:54 look in your Search.pm about line 1355
22:54 make the change to the second regex as above
22:54 see if that fixes it on your install
22:55 not on hdl's Search.pm
22:55 I think the entire regex should be changed as I have it above
22:56 masonj will do
22:57 fbcit there may be one other change you have to make if that does not fix it
22:58 buildQuery introduces a space which should not be there
22:59 you will have to also change this
23:00 ~936      $query .= " ".$limit;
23:00 should now read
23:00    $query ? $query .= " ".$limit : $query = $limit;
23:00 to eliminate an offending space when there is no query, but only limit
23:05 gmcharlt: that commit on fix-perl-path.PL is in the installer branch?
23:05 gmcharlt fbcit: no, just a patch you can apply with git am
23:06 fbcit ok in my inbox...
23:13 it looks like searches on multiple limits only is broke also...
23:14 fixed...
23:17 hrmm..
23:19 masonj i gotta switch to something else for  a while fbcit
23:19 but thanks for having a look at this one :)
23:19 fbcit masonj: tnx for the help :-)
23:20 masonj oh no, thank YOU for the help!
23:22 fbcit gmcharlt: I wondered briefly about the line endings this morning... but very briefly ;-)
23:22 makes sense though.
23:35 chris i can think of cases when you wouldnt want to
23:35 kados yea
23:35 chris if for example there was something on order
23:36 kados ahh, well in that case, you'd likely have items, with a status of 'on order'
23:36 but in the case of electronic materials
23:36 you wouldn't have items ...
23:36 unless you're storing them at the item level (which is a new MARC feature)
23:36 hrmpf
23:36 chris depends on how your acquisitions work
23:36 kados true
23:37 chris but i cant think of any reason you'd want to hide biblio;s with no items
23:37 if they have no legit reason to be in the catalogue ... delete them :-)
23:38 so i should say, the only reason i can think of, is that you dont want to clean out orphaned biblios, just want the system to hide them
23:38 kados right
23:39 chris we could have a syspref "imtoolazytocleanstuffuphideitplease" :-)
23:39 kados hehe
23:39 hiding withdrawn items would make more sense I suppose
23:39 chris yeah i can see that
23:39 hiding in the opac at least
23:40 kados *nod*
23:40 chris i think we have a hidelostitems syspref
23:40 no reason we couldnt have one for withdrawn too
23:41 kados yea
23:41 yea, just fixed the hidelostitems syspref :-)
23:41 chris sweet
23:45 aklibrarian1 Good afternoon all.
23:46 Has anyone had to take a system that has overlapping barcodes between sites to a union catalog.
23:46 I am looking for ideas.  I had been told that we all had different barcodes.
23:46 kados yea, it's messy
23:47 one thing you can do is pre-pend the branchcode to the barcodes
23:47 and write a pre-parse line when reading barcodes to pre-pend the branchcode every time you read a barcode
23:47 but it's messy
23:48 aklibrarian1 messy yes, but less so than rebarcoding at all the branches.
23:48 kados *nod*
23:50 aklibrarian1 Thanks
00:39 atz aklibrarian1: you wouldn't have to rebarcode *all* of them
00:40 at most, half  :)
00:44 it is entirely worthwhile to have the barcode namespace standardized and enforced
00:47 when INFOhio migrated the state juvenile prison system on, they handled rebarcoding by having the staff converge at 1 library at a time
00:47 in rotation
00:47 kados atz: just started looking at your patches
00:48 atz: did you mean to truncate the sql for checkuniquemember function?
00:48 atz i think so... let me review
00:48 kados atz: $request="select count(*) from borrowers where surname=? and categorycode=?";
00:49 was changed to select count(*) from borrowers where surname=?
00:49 ahh, hang on, no it wasn't
00:49 atz yeah, look to context
00:50 kados did you verify that checuniquemember still works?
00:50 ie, in the interface?
00:50 atz do you see my note?
00:50 "this funtion is not legitmate"
00:51 kados yes, but clearly that's incorrect ... ie, the function is, in fact, being used
00:51 atz right, but it doesn't actually affirm/deny uniqueness
00:51 kados so my question is, did you test it in the interface?
00:51 you mean you don't like the algorithm?
00:51 atz it does not checkuniqemember
00:52 yes
00:52 kados so you verified that it does, in fact, check uniqueness from the interface
00:52 atz no.  that's why i commented it with FIXME
00:53 i verified it *originally* did not check adequately for uniqeness
00:54 i did not "fix" that
00:54 :q
00:54 (wrong window....)
00:56 kados atz: under what circumstances is the user warned that there is a problem with uniqueness with the current design?
00:56 aklibrarian1 Atz, I had to step away.  You are right.  Only half the items would have to be rebarcoded. :)
00:57 atz only 1 place: memberentry
00:58 kados ok, and what is a specific test case that will warn the user that there is a problem with uniqueness?
00:59 atz edit one user to have the same data as another existing user
01:00 kados which data?
01:02 atz: ?
01:03 atz first/last name, birthday
01:03 kados can you give me a reproducable example of this?
01:04 starting from /cgi-bin/koha/members/members-home.pl
01:04 atz of what?
01:04 you want to know how the old version warned?
01:04 kados of first/last name, birthday data that caues the user to be warned that there is a problem with uniqueness
01:04 I want to see it work in the interface
01:04 ie, i want to see whatever warning the user gets
01:05 what did you use to test this?
01:05 atz i set everybody's birthday to the same day
01:05 and added users w/ the same name/bday
01:06 kados huh, I've done that about 5 times now with no result
01:07 atz Duplicate suspected Detail Duplicate ? Yes     No, it is not
01:07 The following fields are wrong. Please fix them.
01:08 there is a checkbox
01:08 (after "No, it is not")
01:10 Detail is a link to popup the "other" borrower_details
01:10 kados maybe patron birthdates aren't saving on my install ...
01:10 was that one of the memberentry bugs?
01:11 atz yes, that is the big'un at the end of my submission
01:11 kados k
01:11 guess I can't test the first refactor until I apply the last one
01:12 so how ya feeling about the members code these days?
01:13 atz well... it's bulky and not fun to work with
01:14 and it needs a robust test suite
01:14 but today is the first time I've seen (almost) everything working right
01:15 kados cool
01:15 atz (categories are still broken for me... the record data isn't reflected on memberentry)
01:15 kados really? categories are broken?
01:15 how so?
01:16 atz when you go to edit, the first field is selected, not what the record has
01:16 so it ends up editing that field every time you go to save some *other* change
01:17 kados ahh
01:17 that sounds relatively easy to fix
01:17 atz agreed
01:17 i figured i'd come back to that as a separate task
01:18 btw, any feedback re: xplanner?
01:18 looks like a dead project.
01:18 kados it's java, so I'm again' it :-)
01:19 atz java's great.   it runs (crappily) everywhere!
01:19 masonj yeah, consistently crappy performance
01:20 atz a lot of the stuff works well on sparc workstations... for what that's worth
01:21 the last xplanner release was from 20 months ago
01:24 kados atz: still no luck getting the check for duplicates stuff working on my install
01:24 I add myself a couple times, and nothing prevents it
01:24 Surname Ferraro, Firstname Joshua, DOB 04/21/1978, zipcode 45701
01:25 atz: all patches have been applied
01:26 atz: and DOB isn't preserved either
01:26 atz mind if I try on your koha?
01:26 kados please do
01:26 atz if DOB is still getting dropped, then that's the culprit
02:07 chris i have found the problem
02:07 [Thu Dec 20 20:06:39 2007] [error] [client 121.73.68.100] [Thu Dec 20 20:06:39 2007] memberentry.pl: Illegal Date '2007-12-11' does not match 'metric' format: dd/mm/yyyy, referer: http://staff-crc.dev.kohalibra[…]rs/memberentry.pl
02:07 so we converted it to iso earlier
02:09 which is why its no longer metric
02:10 atz yeah, i had to clean up some double conversions already
02:10 chris yep, i think some might still be happening
02:11 atz trying to follow the rule that modules should only pass ISO date strings between each other
02:11 chris *nod*
02:11 in AddMember tho we have
02:11 $data{'dateofbirth'}  = format_date_in_iso( $data{'dateofbirth'} );
02:11 and since we pass it an iso already it borks
02:12 at least thats what i think is happening
02:12 atz yeah, that's it.
02:12 chris yep that seems to be it
02:12 if i comment out those 3 lines .. all good
02:13 atz i think they've gone comment in/out several times now
02:13 chris :)
02:13 ill check the modify too, and do a patch
02:13 atz modify I am more certain of
02:13 but pls feel free to verify
02:14 chris will
02:14 do
02:14 heres my comment in the code
02:14  # WE SHOULD NEVER PASS THIS SUBROUTINE ANYTHING OTHER THAN ISO DATES                                                                                  
02:14    # IF YOU UNCOMMENT THESE LINES YOU BETTER HAVE A DARN COMPELLING REASON
02:14 atz sweet
02:14 you're my dude.
02:20 chris modmember looks fine
02:52 kados ok, time for a brief break
02:52 then back to work on the opac
03:13 [K] *** join #koha@FreeNode: Arsenic n=Arsenic@modemcable008.130-81-70.mc.videotron.ca
03:14 <Arsenic@FreeNode> anybody is up to date with the head of koha 3.0?
03:20 atz Arsenic: yeah, we've been patching and updating a ton today
03:22 [K] <Arsenic@FreeNode> ok.  Did you try to install it using perl Makefile.PL?
03:23 <Arsenic@FreeNode> Did you change anything related to LDAP ? :)
03:28 kados ldap's been enhanced quite a bit
03:31 [K] <Arsenic@FreeNode> it seems that the version of Net::LDAP::Filter is wrong in the prerequist
03:32 <Arsenic@FreeNode> actually, not LDAP, but Net::LDAP::Filter.  The latest version is 0.15, not 0.34.  .34 is the version of the distribution
03:34 fbcit kados: u around?
03:36 [K] <Arsenic@FreeNode> he was here few minutes ago
03:36 kados fbcit: yea
03:37 fbcit looking over that NoZebra commit I made earlier, it appears I deleted one line too many. :(
03:37 I can't fix it until after 9am tomorrow est...
03:37 kados which line?
03:38 fbcit @@ -1321,7 +1334,6 @@ sub NZanalyse {                 my $value=$_;                 $value=$1 if $value=~m/(.*)-\d+$/;                 unless ($rightresult =~ "$value-") {-                    $finalresult .= "$_;";                 }             }             return $finalresult;
03:38 sorry
03:38 -                    $finalresult .= "$_;";
03:39 this did not affect the searches I was performing, but may affect others
03:41 kados hehe
03:41 yea, search gets tricky really fast
03:42 [K] <Arsenic@FreeNode> :)
03:42 fbcit its hard to run perl -d on Search.pm...
03:42 kados here's a strange one
03:42 $query ? $query .= " ".$limit : $query = $limit;
03:42 causes zebra to bork
03:42 while $query .= " ".$limit;
03:42 is fine
03:42 fbcit I wonder why?
03:42 maybe it should be in standard if...else form...
03:43 kados the first one yields a query that looks like this:
03:43 () and ( (lost,st-numeric <= 0)
03:43 hehe
03:44 yea, weird
03:44 standard if works
03:44    if ($query) {
03:44        $query .=" ".$limit;
03:44    }
03:44    else {
03:44        $query = $limit;
03:44    }
03:44 that's the intent, right?
03:44 fbcit I think that Search.pm really needs some refactoring...
03:44 right
03:45 kados how would you approage that refactoring?
03:45 approach I mean
03:45 fbcit I'd leave it to someone else... hehehe
03:46 kados :)
03:46 well, I'm sure it could use refactoring
03:46 fbcit seriously, I cannot put my finger on it, but it seems a bit unwieldy
03:46 [K] <Arsenic@FreeNode> Hey!  Which version of Net::LDAP::Filter does you CPAN mirror have?
03:46 fbcit no offense to those who have hours of work in it..
03:47 kados well, it's not OOP
03:47 fbcit kados: I'm not sure it will handle zebra && nozebra gracefully
03:47 kados and I'm sure there's some tightening up that could be done
03:47 but in general it's the zebra stuff is pretty sound IMO
03:48 yea, nozebra was kinda bolted on as an afterthought
03:48 fbcit it seems to be groomed for zebra
03:48 I'll try to work through somemore of the issues over the next several days
03:48 kados well, I can't take credit for all of it
03:48 fbcit now that I'm getting a better feel for it.
03:49 kados I wrote buildQuery and it's calling internal subs
03:49 as well as getRecords
03:49 the rest was either already there or bolted on afterwards
03:49 if you have suggestions, let me know
03:50 fbcit I suppose the nozebra has not had a good shakedown?
03:50 kados I need to update the perldoc one of these days
03:50 it's a little rusty
03:50 I've never looked at the nozebra code ... a couple glances and I realized there were some major nomencalture issues with the variable names and I ran away screaming :-)
03:50 fbcit well, a zebra that can't run dies quick, so that explains why zebra work so well :-)
03:50 kados hehe
03:51 [K] * Arsenic@FreeNode is ignored..
03:51 <Arsenic@FreeNode> :)
03:51 kados hehe
03:51 sorry Arsenic
03:52 fbcit I'm about 6 modules away from having koha 3 running on XP...
03:52 kados Arsenic: I don't think I've got that one installed, lemme check
03:52 fbcit hi Arsenic
03:52 [K] <Arsenic@FreeNode> hehe.. np... I'm not hurry...
03:52 kados fbcit: cool
03:52 [K] <Arsenic@FreeNode> fbcit.. cool!
03:52 kados fbcit: library.neu.edu.tr runs on windows 2000 and IIS :-)
03:53 fbcit I wondered if anyone was using IIS
03:53 kados cross_platform_architecture++
03:53 hehe
03:54 fbcit some of the pm's are a challenge to compile due to the outside reqs
03:54 kados atz: what version of Net::LDAP::Filter do we recommend?
03:54 fbcit if we go to maintaining a truly cross platform package, we may have to house some of this stuff as it is hard to find/build.
03:55 kados atz: Arsenic is wondering
03:55 fbcit: yea, I'm all for that
03:55 atz checking
03:55 kados I think we're getting enough userns now that we can afford to maintain multi-dbms and multi-platform
03:56 fbcit 0.34
03:56 atz I've got 0.15 of Filter
03:56 [K] <Arsenic@FreeNode> fbcit: This is Net::LDAP, right?  what about Net::LDAP::Filter... it seems that the latest version is 0.15
03:56 fbcit Makefile.PL wants 0.34 on both
03:56 kados fbcit: did you see a new version of zebra came out?
03:56 atz Net::LDAP is 0.34
03:57 [K] <Arsenic@FreeNode> ok.  the Makefile.PL request 0.3
03:57 <Arsenic@FreeNode> ok.  the Makefile.PL request 0.34
03:57 fbcit makefile needs to be fixed then
03:57 kados fbcit: 2.0.20 ... has support for ICU unicode (yay!) and native support for facets
03:57 fbcit kados: I downloaded the win32 version of zebra last evening
03:57 kados fbcit: which will drastically speed up search results and allow us to do some nifty counting of the facets across the whole result set
03:57 bad kados, focus on the release
03:57 [K] <Arsenic@FreeNode> alright, so i'm not crazy... yet ;)
03:57 fbcit hehe
03:58 atz the versioning confused me considerably... the .15 version of FILTER is in the the .34 tarball for LDAP
03:59 [K] <Arsenic@FreeNode> Tell me guys, what is your set up when your are working.  I mean you are probably editing the installed version of the script, right?
03:59 atz and the CPAN pages show the .34 in the title pages for the .15 Filter
03:59 kados Arsenic: apart from ldap, you've got 3.0 up and running?
03:59 or just can't install because of the wrong version of the perl mod?
03:59 [K] <Arsenic@FreeNode> Almost... some problem with Algorithm::CheckDigits too
04:00 fbcit Makefile.PL says:  Warning: prerequisite Net::LDAP::Filter 0.34 not found. We have 0.15.
04:00 needs to be fixed...
04:00 gmcharlt...
04:00 kados I can fix it
04:00 [K] <Arsenic@FreeNode> can't install the v0.48 (Algorithm::CheckDigits)
04:00 kados should be 0.15?
04:00 fbcit concerning Algorithm it says: Warning: prerequisite Algorithm::CheckDigits 0.48 not found. We have unknown version.
04:01 kados hmmm
04:01 [K] <Arsenic@FreeNode> fbcit: that's it
04:01 kados should I un-name the version?
04:01 atz Arsenic: that has been strange... the original perl installer was buggy about CheckDigits
04:01 fbcit you have to use a funny name for Algo
04:01 atz i think i DL'd the tarball from cpan for that one
04:01 kados 'Algorithm::CheckDigits::M43_001' => 0.48,
04:02 does debian have a package for that one?
04:02 atz yeah, the weird package name on that one
04:02 Arsenic: LDAP should be broadly configurable just through the main KOHA_CONF xml
04:03 fbcit atz: have you run LDAP against an M$ AD domain by chance?
04:03 atz the perldoc on C4::Auth_with_ldap is rather robust... and there are sample LDIF you can load and a test script to run against the sample data
04:04 kados Arsenic, just pushed up a fix for the first one
04:04 Arsenic: does Algorithm::Checkdigits warning cause complete failure?
04:05 atz fbcit: actually, we have some clients for whom we setting up just that
04:05 no MS Server in house, though
04:05 fbcit I currently have an AD backend...
04:05 [K] <Arsenic@FreeNode> kados: no, only a warning prerequisite Algorithm::CheckDigits::M043_001 not found.  We have unknown version
04:05 fbcit all student accounts are there
04:05 maybe I'll give it a whirl
04:05 kados Arsenic in that case, if you do a git fetch, git rebase origin, you should be able to install now
04:06 fbcit kados: how important is 'at' to koha?
04:06 atz fbcit the perldoc C4::Auth_with_ldap is a good starting point
04:06 kados fbcit: it's used for the task scheduler only
04:06 fbcit CPAN says: Module  = Algorithm::CheckDigits::M97_002 (MAMAWE/Algorithm-CheckDigits-0.48.tar.gz)
04:07 kados: I'm having a time replacing it on Win32
04:07 kados fbcit: is there a replacement?
04:07 fbcit windows at is not as robust and does not play with Schedule::At at the present
04:08 kados bummer
04:08 fbcit I thought about modifying S::A to talk to task scheduler
04:08 kados fbcit++ # even for thinking it
04:08 fbcit but that may be a bit more than I can chew
04:08 I wonder how the current win32 port does it?
04:09 kados I doubt they use the task scheduler
04:09 ie, the koha one
04:09 fbcit S::A is a lot of regex stuff
04:10 Irma Hi all!
04:10 masonj hiya irma
04:10 [K] <Arsenic@FreeNode> still have to learn to work with GIT :)  How can I force to get the latest version of a file?
04:10 fbcit hi irma
04:10 [K] <Arsenic@FreeNode> Hi Irma!
04:10 kados Arsenic:
04:10 git fetch
04:11 git rebase origin
04:11 Irma Hi Mason...can I ask you a questio?
04:11 kados Arsenic did you see the git intro page on the koha wiki?
04:11 Arsenic: http://wiki.koha.org/doku.php?[…]lopment:git_usage
04:11 hi Irma
04:11 [K] <Arsenic@FreeNode> kados: Yep.  I did some local modif on my machine and I want to revert to the one you just pushed
04:11 Irma How does Koha 2.2.9 handle hyphenated words?
04:12 masonj for searching?
04:12 fbcit arsenic: git fetch
04:13 git rebase origin
04:13 kados Arsenic: git checkout -f fileyouchanged
04:13 Irma We are still using 2.0.0.and some entries have been catalogued with hyphens some not. So searching is often not successful as the user thinks it is hyphenated and gets no results.
04:13 atz or git checkout filename
04:13 kados Arsenic, then git fetch, git rebase origin
04:14 Irma My question is how does this get best handled in Koha 2.2.9?
04:15 kados Irma: 2.2.9 doesn't do any special handling of dashes
04:15 [K] <Arsenic@FreeNode> ok.  It worked.  It seems that the -f option can not be used on the master branch
04:15 kados Arsenic you should be working off of a branch
04:15 Arsenic: http://wiki.koha.org/doku.php?[…]ew_for_developers
04:15 Arsenic take about 2 minutes to read that section
04:16 Arsenic: it has 90% of what you need to know about Git :-)
04:16 [K] <Arsenic@FreeNode> I did already :)
04:16 kados Irma: 3.0 treats dashes as spaces
04:16 Irma: at least the zebra version does
04:17 [K] <Arsenic@FreeNode> brb guys!
04:18 Irma2 Irma is now Irma2 and is back again.
04:18 Can I have some more commnets on this issue?
04:19 kados 20:15 < kados> Irma: 2.2.9 doesn't do any special handling of dashes
04:19 20:16 < kados> Irma: 3.0 treats dashes as spaces
04:19 20:16 < kados> Irma: at least the zebra version does
04:20 Irma2 so NZ does not?
04:21 kados Irma don't think so
04:23 Irma2 Thank you and and cheers for now.
04:23 fbcit kados: have you run a search on a term with hyphens in it since I worked on Search.pm?
04:23 zebra that is?
04:24 Irma2 would this be on the LibLime demo ?
04:25 If so, can you suggest a hyphenated teerm that might be in the index?
04:25 kados fbcit: no, why?
04:25 fbcit: it's not a code thing, it's an index thing :-)
04:26 Irma2 sure.
04:26 kados irma: I don't know of any examples
04:26 fbcit good...
04:27 kados http://search.athenscounty.lib[…]h?q=mother-in-law
04:27 http://search.athenscounty.lib[…]s/search?q=mother in law
04:27 fbcit I modified a regex that sub'd spaces for hyphens (among other things) to sub nothing for them.
04:28 kados fbcit: hmmm
04:28 fbcit: just for nozebra I hope
04:28 fbcit: :-)
04:28 fbcit hard to tell...
04:28 kados yea, wouldn't want to do that really
04:29 that'd turn mother-in-law into motherinlaw ?
04:29 fbcit sp
04:29 probably so
04:29 it may need to be cut into two regex's
04:30 kados http://search.athenscounty.lib[…]earch?q=top-notch
04:30 fbcit $string =~ s/-|\.|\?|,|;|!|'|\(|\)|\[|\]|{|}|"|&|\+|\*|\// /g;
04:30 became
04:30 $string =~ s/-|\.|\?|,|;|!|'|\(|\)|\[|\]|{|}|"|&|\+|\*|\///g;
04:30 kados http://search.athenscounty.lib.oh.us/search?q=top notch
04:30 hmmm
04:30 fbcit: what line?
04:30 fbcit ~1382
04:30 } else {
04:30 kados yea, that's all nozebra stuff
04:30 fbcit afterward
04:30 kados so you're safe
04:31 well ...
04:31 fbcit as I say, it may need to be changed to two
04:31 kados safe is probably the wrong term
04:31 fbcit it was the () that should not be swapped with white space
04:31 their used to bound the limit
04:31 for some reason or other
04:32 kados *shrug*
04:32 the recursive bits of that code couldbe improved
04:32 fbcit I notice parens are constantly being stripped off w/ regex's though.
04:33 kados huh
04:33 fbcit is the object of the recursion to walk from right to left?
04:34 kados yea, it's basically a syntax parser
04:34 [K] * Arsenic@FreeNode is back.
04:35 fbcit re: huh
04:35 kados takes the CCL generated from buildQuery and turns it into SQL basically
04:35 fbcit    #process parenthesis before.  
04:35 [K] <Arsenic@FreeNode> Is there a list of best practices available somewhere?
04:35 kados it could use a re-write
04:35 fbcit    if ($string =~ /^\s*\((.*)\)(( and | or | not | AND | OR | NOT )(.*))?/){
04:35 kados yea, that's just wrong
04:36 well, certainly not from the zebra query string
04:36 I dunno about the nozebra one
04:36 [K] <Arsenic@FreeNode> I would like to know how you are setup for your development.  You are probably the installed version of the perl script.  Are you copying them back afterward under your GITed directory?
04:36 kados in the zebra query string they are used to specify order of operations
04:36 fbcit does buildQuery do zebra queries also?
04:37 kados buildQuery builds several versions of the query the user has entered
04:37 one for zebra, one for humans, one in CGI, etc.
04:38 it doesn't actually do any zebra queries
04:38 fbcit I see that
04:38 kados that's handled in getRecords
04:38 fbcit but the syntax for both queries is established there, right/
04:38 ?
04:38 kados don't quite understand the question
04:39 Arsenic take a look at the 'make dev' target
04:39 Arsenic if you want to run directly off of a git repo
04:39 fbcit: what do you mean by 'both queries'?
04:40 fbcit: the syntax generated conforms to CCL, which has a formal syntax
04:40 Irma Sorry kados and fbcit...I must be on my way.  I will read the log later and see if I can gleam some new wisdom and answer to my querry. Thanks for all you guys do.  Cheers.
04:40 kados fbcit: is that what you were asking?
04:40 fbcit but the return data from buildQuery will be passed to getRecords or NZgetRecords depending
04:40 kados oh, right
04:41 yes, and then NZanalyze or whatever goes and converts the CCL into SQL
04:41 fbcit so one must be careful how one munges with buildQuery code.
04:41 kados yes, please don't touch buildQuery :-)
04:42 fbcit too late... :-)
04:42 kados hehe
04:43 fbcit I think that the fact that both z and nz use the results is what breaks nz queries...
04:43 but the syntax can be cleaned up in the NZ subs I would think for the most part
04:44 kados well, nz was bolted on as an afterthought
04:44 fbcit but that extra whitespace had to go... hence the ternary
04:44 kados I think that paul wrote the original code for nz
04:44 fbcit why even offer nz support?
04:44 kados I dunno
04:44 not my idea :-)
04:44 easier installation procedure
04:44 fbcit really?
04:44 kados easier to maintain (in theory)
04:45 yea, you don't need to setup zebra
04:45 we've really streamlined that process now
04:45 you shoulda been here  a year ago
04:45 it was insane to get a koha + zebra system going :-)
04:47 [K] <Arsenic@FreeNode> I didn't install zebra yet.  And there a lot of library who doesn't really need it, right?
04:48 <Arsenic@FreeNode> I mean, a library with less than 20000 items do not really need zebra...
04:48 kados Arsenic: right
04:48 Arsenic: but zebra is nice even for small collections if you don't mind managing it
04:49 ok, well I've got to get some dinner
04:49 [K] <Arsenic@FreeNode> You are right.  I forgot the other advantage than the perfomances one..
04:50 <Arsenic@FreeNode> diner... I'm about to go to sleep :) 23h50 here in Montreal :)
04:54 <Arsenic@FreeNode> Hmmm... i just ran into another installation problem.  During the make, I got a "Bad arg length for Socket::inet_ntoa, length is -, should be 4 at rewrite-config.PL line 67.  Any clue?
04:55 fbcit ?
04:55 [K] <Arsenic@FreeNode> lenght is 0, no -
04:57 <Arsenic@FreeNode> hmmm probably related with the gethostbyname.  was this added lately^
04:57 <Arsenic@FreeNode> ?
04:58 fbcit your isntall appears not to like the gethostbyname call
04:58 can you 'echo $WEBSERVER_IP' at the command line?
04:59 [K] <Arsenic@FreeNode> nothing, nada, rien
04:59 <Arsenic@FreeNode> Is that a variable that we have to set?
04:59 fbcit if not maybe 'export WEBSERVER_IP=whatever-it-is'
05:00 then re-run make
05:00 that should avoid the offending call
05:00 [K] <Arsenic@FreeNode> actually, by looking at the code, if gethostbyname returns nothing, it should use localhost
05:02 <Arsenic@FreeNode> sorry, if hostname returns nothing
05:04 fbcit I've never seen that error before... try the export
05:06 [K] <Arsenic@FreeNode> make in progress...
05:08 <Arsenic@FreeNode> it works...
05:09 <Arsenic@FreeNode> Weird. I didn't have to export this variable before when I tried the koha-installer repo.
05:10 fbcit not sure there, AFAIK rewrite has not changed
05:11 at least that line of it
05:12 the dir structure underwent significant revision a week or so ago
05:15 [K] <Arsenic@FreeNode> ok... I guess the version on the koha-installer repository was change.  I'll check that right now.
05:15 <Arsenic@FreeNode> fbcit: Are you the only one working on making the Windows version work?
05:16 fbcit there is a windows port of 2.x
05:16 [K] <Arsenic@FreeNode> ok, but on 3.0?
05:16 fbcit gmcharlt & I have been working on 3
05:17 mainly porting the installer first
05:17 some of the pm's are a challenge to port over
05:18 [K] <Arsenic@FreeNode> i guess so..
05:19 fbcit i'm off to sleep, g'night
05:19 [K] <Arsenic@FreeNode> where are you based?
05:19 fbcit NC, USA
05:19 00:20
05:19 [K] <Arsenic@FreeNode> same time here in Mtl
05:19 <Arsenic@FreeNode> montreal
05:20 fbcit hope your install goes well
05:20 [K] <Arsenic@FreeNode> I hope too :)
05:20 <Arsenic@FreeNode> thanks for your help!
05:20 fbcit np g'night
05:20 [K] <Arsenic@FreeNode> thx, you too
05:48 arsenic How does koha know that it is running for the first time and he has to launch the webinstall?
05:58 chris by the version systempreference .. or in this case, the lack of it
06:01 arsenic thanks chris... :) this solve my problem.. Note to myself: don't forget to recreate your db after reinstalling ;)
06:01 chris :)
08:19 hdl chris there ?
08:23 paul hi

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

koha1