IRC log for #koha, 2008-05-08

All times shown according to UTC.

Time S Nick Message
12:57 owen Hi #koha
12:57 hdl hi owen
12:58 hdl owen : do you know of jquery problems with large dataset management ?
12:58 owen Hi hdl. I'm just getting caught up after being out of the office for a couple of days. I'm just now reading about that in your bug report
12:59 hdl sorry to catch you up.
12:59 owen It's just a matter of the time it takes for the javascript engine to process all that data.
12:59 owen It's definitely not designed for really huge datasets
13:00 owen Could you save the HTML of your huge overdues page and send it to me for testing?
13:00 hdl *nods*
13:00 paul owen & hdl : an interesting point is that is works with Opera.
13:00 paul so it really seems it's a javascript/firefox problem
13:01 owen And I suspect it also depends on the computer the browser is installed on, and available memory at the time of execution
13:04 hdl owen : sent
13:05 owen Thanks
13:16 paul couldn't we just remove jquery for this script ?
13:17 hdl It is a quick fix.
13:17 owen We could. We should also think about whether there is a better way to present this page if there will always be thousands of results
13:18 owen For instance: paged results and a link for a full printable version?
13:18 hdl owen it is likely to be.
13:18 paul owen: right. In fact, some/many/most libraries will use this page to upload the CSV for mailing ppl
13:18 paul so the CSV is the most important button I think ;-)
13:18 paul maybe :
13:18 paul - display the page without any result
13:19 paul - let the user clic to "display" or "upload"
13:27 owen Hmmm... Downloading a CSV file of overdues isn't working for me: "impossible d'ouvrir le fichier de relances at ../misc/cronjobs/ line 81"
13:31 gmcharlt_ greetings #koha
13:31 paul hi gmcharlt_
13:32 hdl gmcharlt: mail sent to report some errors from upgrade script.
13:32 gmcharlt hdl: got it.  if it's not too huge, could you send me a mysqldump of the 2.2.9 DB you were testing?
13:33 hdl on the way
13:55 owen Should I file a bug report about not being able to download a CSV of overdues? Or should I be checking for a problem on my system?
13:59 owen nengard?
13:59 nengard hiya owen
13:59 owen I'm just thinking about the RSS icon on the search results page
14:00 owen That should appear even if there are no results, dont you think?
14:01 atz owen: yeah
14:01 atz in case you want to know when they finally get something that matches the query
14:02 owen atz, that's what I was thinking
14:02 atz seems like the *most* valid use of RSS
14:05 atz owen: do we have something like a green checkmark (OK indicator) icon?
14:07 owen atz, intranet-tmpl/prog/img/approve.gif
14:07 atz cool, thx
14:13 nengard owen - sorry got distracted - i agree with atz it should show up no matter what because there may be results later
14:22 hdl owen : I am working on that csv problem
14:26 gmcharlt_ odd_internet_connection_problems--
14:26 gmcharlt_ heck, even_internet_connection_problems--
14:28 hdl owen : about jquery : is this only a problem of line counts ?
14:28 hdl We could write HTML table markup on one line.
14:28 owen I believe it's about the number of rows in the table
14:28 owen So whitespace doesn't matter to jquery
14:28 hdl *nod*
14:29 owen Would my vendors not appear in the filter form if I don't have any late orders?
14:30 hdl yes.
14:31 owen Okay
14:34 owen[…]d=255&action=view
14:35 owen What is the purpose of the 'acquisition claim' dropdown?
14:36 owen Oh...That lets you select which letter to use when submitting a claim?
14:36 nengard owen - yes
14:37 nengard it's coming from Tools > Notices
14:39 atz owen: btw, remember when we tried to use H:T:P built-in loop variables (like counters, even/odd indicators, etc.) and it didn't work?  I've got a patch that fixes that.  The switch for that has to get passed to the template constructor (not as a template param).
14:40 owen Huh, interesting
14:42 atz so we can eventually kill off all the dumb parts of internal code that preprocess loops to add counters, "highlight" or "toggle" variables, etc.
14:42 owen nengard:
14:43 nengard owen you are awesome!!!!
14:43 nengard only one comment
14:43 owen Please
14:43 nengard you might want to have the bullets up where the filters are - just to be consisten from page to page - and maybe move the filter and list over to the left
14:44 nengard k - hang on
14:47 nengard owen email sent to gmail
14:48 owen Hmmm.. I see. But I disagree. Other pages use the left-hand area for filters as well
14:49 owen /circ/, for instance
14:49 nengard owen - okey dokey - which page? I haven't seen one yet and may have to update some screenshots
14:49 nengard okay
14:49 nengard owen doesn't have menu items though ....
14:50 owen True :)
14:50 nengard and i do like the way it looks ... i'm just concerned that the bullets look out of place on the lateorders page
14:50 nengard i do have to update my screenshots for overdues :)
14:50 owen I'm concerned that we have those bulleted links on that page because we can't come up with a better way to organize those links
14:53 owen I see we've got some terminology inconsistency here... Is "late issues" correct rather than "late orders" ?
14:54 nengard owen i think it should be late orders
14:54 nengard good catch
14:55 nengard owen as for the bullets - i'm up for anything that makes it more usable ... I personally don't know where else to put them ....
14:58 nengard owen should i report the language issue - or are you doing that with your redesign?
14:58 owen I'm fixing it now
15:00 nengard k
15:42 nengard on koha/circ/ there is nothing in the action column ... how do i actually mark the item as transferred
15:42 hdl check it in
15:47 tim I copied the prog directory to a new one to make changes to the opac templates and set the opacthemes system preference to the new directory, but the changes I make there don't make a difference.
15:47 hdl which changes ?
15:48 tim Then I tried making a change in prog and it was updated on the opac.  Does anyone know why the systempreference isn't setting the opac to use the new directory?
15:48 nengard thanks hdl
15:49 hdl np : it is not straight-forward.
15:49 hdl owen : can you answer tim ?
15:50 tim I thought maybe it was a permissions problem.  I just checked and they're ok.
15:51 owen tim, I'm curious what kind of changes are you making?
15:52 owen Once you start down the path of template customizations you're looking at future upgrade problems.
16:00 owen tim: switching opac themes seems to be working for me. You're testing 3.0, right?
16:02 atz i know koha has ways of hiding items from the opac, marking them notforloan, or whatever
16:02 atz is there currently a mechanism that hides a biblio record from OPAC display/indexing?
16:06 owen atz, Liblime worked up something for NPL's dev_week install, I don't know if that has made it into 3.0
16:07 atz I'm guessing not, mainly b/c I don't see anything in the biblio table that could record "hiddenness"
16:08 owen we have an authorized value called "suppress" which can be applied to a biblio. That hides the biblio from OPAC results
16:08 owen You have a customer that has a similar requirement?
16:13 tim owen: some things aren't displaying and I get no errors in the logs, so I was marking spots on the templates to make sure those parts are getting displayed.
16:14 tim owen: I'm testing 3.0.  Did a git update yesterday.
16:15 owen When you view the source of an opac page, do you see the /prog/en/ path or your custom template path for things like stylesheets in the <head>?
16:16 tim I wanted the call number displayed.  I'd like itemnumber displayed but understand why that might not be a good idea, but nothing is getting displayed there now.  I'd like to find out why.
16:18 owen The call number display question is one that is still open. I think it must be resolved before the official release
16:18 owen If you're working with a new set of templates just for that, I'd advise you to wait
16:21 owen ...not that that helps explain why your custom set of templates isn't working properly
16:22 tim I read where that hasn't settled.  It just looks like something should be displaying there and it's blank.  Actions is also blank.  I don't even know what that is for.
16:23 owen That might contain "place hold," "save to lists" and/or "add to cart"
16:24 owen Those might be turned off in your system preferences?
16:24 owen We need to hide the label if all three are off.
16:26 tim I'll need to find the preferences.
16:26 tim I found noOPACholds and it's set to OFF.
16:31 gmcharlt_ atz: there's an OPAC suppression - implemented by setting 942$n to 1 in a bib, which is indexed in the 'Suppres' index in zebra. then adds the 'and not suppress=1' to searches if an OpacSuppression syspref is ON
16:32 atz gmcharlt_: thx, kindly
16:33 hdl 942$n is linked to which item field ?
16:33 gmcharlt_ hdl: it isn't linked to any
16:33 hdl is it hard coded ?
16:33 hdl 942$n ?
16:33 gmcharlt_ also, note that feature is not fully integrated - syspref is not in install scripts, e.g.
16:34 gmcharlt_ hdl: subfield is hard-coded in zebra config - not sure if a SUppress index is defined in the Unimarc config by default or not
16:34 hdl gmcharlt_: brilliant.
16:35 hdl I will look at that... If i have time.
16:39 tim RequestOnOpac is set to ON.  noOPACUserLogin is set to OFF.
16:41 tim No item status is displayed in the search results.  Is that still missing or is it just my install?
16:42 tim If I click on Limit to currently available items I get no rusults.
16:57 paul could anyone test circ/
16:57 paul for me it's always empty, even if it should not be.
16:57 paul I think it's due to a date handling problem.
16:57 paul I'm wondering if it's my DB, the update process, or a true bug.
16:58 paul, line 123 reads :
16:58 paul    next unless ($num->{'waitingdate'} && $num->{'waitingdate'} ne '0000-00-00');
16:58 paul and waitingdate is empty for me
16:58 paul thus, all entries are removed from the list & the list is empty :\
16:59 owen paul, I get some results
16:59 paul :\
16:59 atz probably date formatting on the input
16:59 paul do you have an access to your DB directly ?
16:59 paul could you tell me what is in waitingdate field ?
16:59 paul (for me it's always '0000-00-00')
16:59 owen Is that the reserves table?
16:59 paul yes, sorry
17:00 owen For me it's either NULL or has a valid date. I don't have any '0000-00-00'
17:00 paul mmm... sorry, for me it's null too
17:11 paul owen : for reserves lines with priority=0 and found='W' do you have a waitingdate ?
17:12 owen Yes
17:13 paul then my problem is here.
17:13 paul the next question being to understand why I don't have a waitingdate when a reserve is filled...
17:30 tim owen: I missed one of your questions.  Yes.  /prog/en is in the paths in <head> for scripts, css and the shortcut icon.
17:32 tim paul: My reserves table has the same problem as yours.
17:32 tim but I hope some day I'll get up to the status of really poor hacker.
17:43 atz paul: you should probably replace 0000-00-00 with null in waitingdate field.
17:43 atz all (or conservatively, almost all) of the "0000-00-00" code should be deprecated by now
17:54 tim I was wrong.  Mine is not exactly like pauls.  Mine are all null.  No 000-00-00
18:00 atz there was old code that seeded "0000-00-00" as default data
18:00 atz but newer installs shouldn't see any of that
18:33 owen Are icons for CCODE working? If so, what am I missing?
18:56 paul OK, I understand my problem with reserves table : the waitingdate field did not exist in 2.2 It is created NULL on upgrade. I'll investigate for a way to fill it with something when upgrading
19:20 paul atz : in fact, there is no 0000-00-00 in reserves.waitingdate, just NULL. but the NULL is a problem for reserves with priority=0 & type=W
19:20 paul a newly created reserve has a date. It's a migration problem.
19:20 paul i'll investigate tomorrow
19:44 tim My install has a system preference called noOPACUserLogin.  I had to add one called opacuserlogin and set it to yes to get a login on my opac.
19:44 tim Has anyone else had to do that?
19:48 tim That fixed my actions links.
20:17 acmoore is koha supposed to be able to run on a version 4.something of mysql or just 5.something?
22:20 chris hmm[…]oha-project-news/
22:22 chris acmoore: the unicode/utf8 support in mysql 4 is not nearly as good as 5, but you can run it on 4
22:26 thd chris: the Vermont Koha project fact sheet indicates an interest in scaling up Koha to their needs,[…]ha_Fact_Sheet.pdf .  What do you think they mean by that?
22:29 chris i think they havent been keeping track of koha development
22:29 chris and their assumptions are based on 2.2.9
22:29 chris also i dont think they understand free software/open source :)
22:30 chris ie they say this
22:30 chris                                                              This means there are likely
22:30 chris hundreds or thousands of people, in Koha’s case, with library and/or computer
22:30 chris programming experience working at any given time to improve the software.
22:30 chris but then they want to splinter off, and lose that benefit
22:32 chris i suspect the same thing that will happen with the south american libraries
22:33 chris sorry, the same that happened with the south american version of koha, will happen with the vermont version
22:33 gmcharlt chris: they need to be strongly encouraged to contribute back to consensus Koha
22:34 chris yep
22:34 chris and engage with the community
22:34 gmcharlt but I'm all for libraries getting together to employ a programmer to hack on open source library software
22:34 chris oh definitely
22:35 gmcharlt libraries need to encourage and develop organic programming capacity
22:35 chris but i think if you talk with the community, you might find out the features you want arent as vermont specific as you thought
22:35 gmcharlt and not just rely on steenking vendors ;)
22:35 chris yep, im all for that
22:35 chris its the creating a new version i think is a bad idea
22:36 gmcharlt forks often are - particular ones made from a lack of understanding about how open software projects work
22:36 chris yup
22:37 chris theres even less need for a fork with git
22:38 chris back in the old cvs days, if you didnt have commit access, well a fork was about all you could do
22:38 chris now you can cherry-pick, branch, merge, to your hearts content :)
22:39 thd tumer's fork was probably necessary for him to make the progress he needed in a reasonable time frame
22:39 chris yep, he did commit it all though
22:40 chris as was the liblime dev_week
22:40 chris they were more branches, but with cvs its a nightmare to merge back in
22:41 thd yet, tumer was working for a long time making great improvements before he gave the community any notion of what he was doing
22:41 chris hmm
22:41 chris he was pretty open on irc and the devel list with what he was doing
22:42 chris i wonder if nicole wants to go to vermont? :)
22:42 thd chris: I mean before that he had a few messages about Turkish character support without commenting upon what he had been working towards
22:44 thd chris: One can imagine that others have efforts at making improvements in more modest ways and do not try to communicate well about them to the larger community
22:46 chris yep
22:46 gmcharlt thd: which is actually one of things I think we'll need to be careful about as we think about things like expanded unit test coverage - don't want to make the barrier to entry so high that somebody who makes a small change feels too intimidated to contribute it
22:46 chris good point
22:47 chris the barrier for koha has always been pretty low, which i think is a good thing
22:49 thd gmcharlt: unit tests should not be designed to do perhaps not much more than ensure that valid data at input is still valid data at output and has not been corrupted by the system along the way.
22:51 chris we do a bunch of work here with moodle
22:51 gmcharlt thd: necessary but not sufficient, IMO - unit tests should also provide some indication that the desired functionality is actually occurring; regression tests should verify that bugs don't reoccur
22:51 chris i wonder if we could do some classes, via moodle
22:51 chris how to write a test case
22:51 chris etc
22:52 gmcharlt chris: interesting idea
22:53 chris something a bit more interactive, than reading a wiki page
22:53 thd gmcharlt: just as importantly, tests should show that invalid data that manages to sneak in does not break the system. However, programming for robustness modelled on biological systems is still too inefficient in terms of CPU time if not programming time
22:54 chris koha dev 101 :-)
22:54 gmcharlt chris: you mean you've never had a wiki page try to edit you?
22:55 chris in soviet russia maybe :)
22:56 gmcharlt thd: I'm curious - modelled on biological systems?  genetic algorithm type approaches?
22:58 thd gmcharlt: Gerald Sussman and colleagues at MIT teach a course about true system robustness.
22:59 thd Gerald Sussman usually gives a presentation about that work at FSF associate member meetings each year.
23:00 thd I made a summary of what he presented a couple of years ago.  I thought it was fascinating.
23:01 thd If you have a cut on your hand you do not expect to die.  Your code should also be that robust but no ones code is.
23:04 gmcharlt thd-fr-lesson: thanks - you've given me something interesting to pursue
00:05 atz nobody's code is that robust... but on the other hand, very few ppl's code needs to be.
00:05 atz i.e., nobody dies because of OPAC errors.
00:06 atz at least, not directly
00:51 chris
00:52 chris its currently extruded a new bracket for itself
00:52 chris extruding even
00:52 chris its very tempting to just go stare at it instead of poking at oracle
00:54 atz chris: your link crashed my firefox/chatzilla
00:54 atz :\
00:55 chris doh, sorry
00:55 chris[…]/darwin-small.jpg  
00:55 chris hopefully that one doesnt
00:55 chris you can get one, its licensed under the GPL :)
00:55 atz it also crashes if I try to print preview on certain docs... but only on mac
00:56 chris weird
00:56 atz is that a 3D printer?
00:57 chris yep
00:57 chris RepRap is short for Replicating Rapid-prototyper. It is the practical self-copying 3D printer shown on the right - a self-replicating machine.
00:57 atz sweet
00:57 chris This technology already exists, but the cheapest commercial machine would cost you about €30,000. And it isn't even designed so that it can make itself.
00:57 chris So what the RepRap team are doing is to develop and to give away the designs for a much cheaper machine with the novel capability of being able to self-copy (material costs are about €400).
00:57 chris Following the principles of the Free Software Movement we are distributing the RepRap machine at no cost to everyone under the GNU General Public Licence. So, if you have a RepRap machine, you can make another and give it to a friend...
00:58 atz how can it remake the steel bars?
00:58 chris   <-- if that link doesnt crash
00:58 chris that tells ya
00:58 atz i.e., what counts as "materials"?
00:59 atz cool, loads ok
01:00 chris as i said, its fairly distracting to have in the office :)
01:00 atz it's like a poor man's CNC
01:01 chris pretty much yep
01:05 masonj heya chris
01:06 chris hey mason
01:06 chris hows things?
01:06 masonj telstra has been wiggy all week ;(
01:07 chris hmm mine has been ok
01:07 chris but im on a different subnet
01:07 chris (i run this irc client in a screen on my machine at home)
01:07 masonj nice
03:32 chris hmm is bugzilla down?
07:37 chris evening
07:40 lloyd_ hola

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