IRC log for #koha, 2014-03-08

All times shown according to UTC.

Time S Nick Message
00:48 maximep left #koha
00:57 joat joined #koha
01:13 wajasu joined #koha
01:30 trea joined #koha
01:33 Dyrcona joined #koha
02:07 rocio left #koha
02:35 mtompset Oh the joys. I'm back.
02:35 wajasu i'm fixing  bug2946 now
02:44 dpk joined #koha
02:59 mtompset oh. didn't notice that.
03:00 wajasu i made a patch.
03:00 wajasu if this bug 11909 patch get signed, QA'ed, pushed,  i will then fix all the bugs related to withdrawn status while its all fresh in my head.
03:00 mtompset it still happens in master?!
03:00 huginn Bug http://bugs.koha-community.org[…]_bug.cgi?id=11909 major, P5 - low, ---, matted-34813, Needs Signoff , Fix hidelostitems, OpacHiddenItems total count, prog them facet display
03:00 wajasu yes still happens.
03:12 mtompset okay... Going to down an energy drink.
03:12 mtompset Grab a mug of water.
03:12 mtompset And generate fresh data for testing.
03:13 wajasu also, note bug 3194, which i am still showing the biblio if all items are lost, opting for the availability facet to filter further.
03:13 huginn Bug http://bugs.koha-community.org[…]w_bug.cgi?id=3194 enhancement, P5 - low, ---, gmcharlt, NEW , Hide records in the OPAC if all items are lost and "hidelostitems" is on
03:13 wajasu i hope that was not what is wanted now.
03:14 mtompset This is why we need a good "standard" data set.
03:14 mtompset I'm thinking I choose colors and z39.50 60 books of red, green, and blue (exclusive to those colors).
03:15 mtompset green = "Green Eggs and Ham" (for example)
03:15 mtompset 60 books should put me over default result set and page sizes by 3x.
03:24 wajasu dis may help http://www.youtube.com/watch?v=-5m5Lva3yeY
03:31 mtompset Do you know the password for the sample patrons?
03:33 mtompset Nice music.
03:36 wajasu i picked lillie, and changed her username to lil and password apple.  and also for edna, etc.
03:39 mtompset I changed one to koha, and then update borrowers set password=...
03:39 mtompset But here is my problem.
03:40 mtompset None of the staff have access to the main page?!
03:42 wajasu you don't have superlibrarian access? to find a patron and edit ?
03:43 mtompset When you do a blank install and include the sample patrons, none of them do.
03:43 mtompset That's just weird and wrong.
03:45 wajasu i just work as superflybrarian  http://www.youtube.com/watch?v=0B6TKClPFQA
03:45 wajasu i guess they want us to create staff and assign privledges.
03:58 mtompset adding gets boring fast. :(
04:04 wajasu i spent 6 hours on this stuff.  understand lots of issues with search now.  this is the least intrusive/risky  patch to get things normal again.
04:06 mtompset maxRecordsForFacets -- check.
04:06 wajasu i made a patch against master for bug 2946.  though it was originally requested for rel3.0.  i wonder if i should change the version to master. if someone running 3.0 want to apply it, they can examine the patch and edit their runtime :)
04:06 huginn Bug http://bugs.koha-community.org[…]w_bug.cgi?id=2946 major, P3, ---, gmcharlt, NEW , Lost items (that are hidden in Normal View) still show up in the MARC View in OPAC
04:06 mtompset OPACnumSearchResults - check.
04:12 mtompset interesting... master with 13 biblios in midway. hide midway... one still visible, because... there are no items. ;)
04:15 mtompset sorry, wajasu. The interesting case is correct in master, and not in your 11909 patch.
04:16 mtompset The annoying part is that the visible item is on page 3.
04:16 mtompset I set my maxRecordsForFacets to 1 and my OPACnumSearchResults  to 3
04:17 mtompset Yes, I know those are crazy values, but with a small dataset, I needed to scale down. :)
04:17 wajasu ok
04:18 mtompset I just changed to maxRecordsForFacets =20. Still same problem.
04:20 wajasu does the one still visible have a non hidden item?  that is lost.  if there is a lost item that is hidden in availability, the biblio will still show.
04:20 wajasu do you have hidelostitems on? if you turn it off, you might see it in availability field.
04:21 mtompset These aren't lost items.
04:21 mtompset It is a biblio with NO items attached.
04:22 wajasu do we want to hide biblios that have no items?
04:22 mtompset NO!
04:22 mtompset actually, I changed my opac search pref to 5 per page.
04:22 mtompset I get the pages display.
04:22 mtompset no results for pages 1 and 2, and 3 redirects.
04:24 wajasu did you rebuild_zebra after the OpacHiddenItems change?  rebuild_zebra uses Search to get get search_results to export for zebra.
04:24 mtompset I'll triple check that.
04:24 mtompset I reset to master.
04:24 mtompset I left OpacHiddenItems.
04:25 mtompset I reindexed fully.
04:25 mtompset master says 13 results (which includes the hidden count)
04:26 mtompset 11909 says no results, no results, and then redirects.
04:27 mtompset (for page 1, 2, and then 3)
04:27 mtompset 13 is wrong, but it is consistent.
04:28 mtompset why did you remove the hidden because lost code in C4::Search::SearchResults?
04:28 wajasu another thing to check.  i am not sure how the logged in branch is kept/cached.  if you try to search as anonymous user (log out) does that help?
04:29 mtompset I am not logged in.
04:30 wajasu i removed it so we show biblios with lost items, though the items are hidden.
04:30 mtompset This is why I am expecting 1 biblio visible.
04:30 mtompset The one without items. :)
04:30 mtompset But back to my question: why did you remove the hidden because lost code in C4::Search::SearchResults?
04:30 wajasu let me see if i delete my items, if what i get.
04:32 wajasu i made it so the OpacHiddenIems just doesn't return the biblio from Search.pm back to opac-search.pl so the counts could be correct.
04:33 mtompset I don't understand.
04:33 mtompset How does removing the hidden because lost code do that?
04:35 wajasu the code i removed was incrementing hideatopac for lostitems and i wanted to decouple that logic form the OpacHiddenItems.  I also hade to make that decision in Search.pm search_results and look at LAL the items to see if they ALL were hidden.
04:36 mtompset you are confused by the word item.
04:36 wajasu item = available copy to loan, checkout, etc
04:37 mtompset not in the code.
04:37 mtompset the foreach is marc::field
04:37 wajasu hey, did you check the availability facet to toggle biblios that have/have no items
04:39 mtompset -- maybe I'm getting confused... let's read again...
04:40 mtompset Okay... got it @fields is an array of 952 fields.
04:40 mtompset (which is items)
04:40 wajasu yes. i beleive so.
04:41 mtompset subfieldstosearch is a hash built from marc_subfield_structure (or something like that) for all the items.* entries.
04:43 wajasu yes
04:43 mtompset item is a hashref to keys based on item column names and their corresponding MARC values.
04:44 wajasu yes
04:45 wajasu corresponds to item table.  if you mysql describe items, you see the column names  for item->{'itemlost'} and such
04:46 mtompset right.
04:46 wajasu if you are looking in Search.pm goto line 1710.  that is the loop of the records (ie biblios).
04:47 wajasu they do lost of stuuf and by line 1800 they pull out items fields.
04:47 wajasu line 1847 has a hideatopac_count
04:47 mtompset 1802
04:48 mtompset Right 1847.
04:48 mtompset But that is because hideatopac_count is counted for each biblio.
04:49 wajasu right.  so i am only using that for OpacHiddenItems now.     in line 1869.
04:49 wajasu then we must loop through all the items and check each one to see fi they are hidden.
04:51 mtompset right... but why only OpacHiddenItems? Why remove the lostitems check?
04:51 mtompset why remove the 'next'?
04:51 wajasu then after that item loop in line 2003, hideatopac_count  correctly skips the following code and the biblio won't be in the newresults array when its returned later.
04:51 wajasu I'm explaingin hidden first.
04:52 mtompset actually, with the code removed, no it won't.
04:53 mtompset imagine a biblio with multiple items.
04:53 mtompset all but one are hidden via OpacHiddenItems.
04:53 mtompset the one that isn't is lost and the syspref says hide it.
04:54 mtompset You cut the code that properly counts it.
04:54 wajasu the biblio will still run through.
04:54 wajasu the item is lost and won;t should, but we still want it to pass through.
04:55 wajasu my view.
04:55 wajasu if all the items are hidden, and no other non-hidden items exist, the count will match and the biblio will be skipped and not returned.
04:56 wajasu if one lost or non-lost item exists, the count will not match and the biblio will be returned.
04:56 wajasu so we get to see the biblio if it had lost items.
04:57 wajasu if hidelost items is on, the biblio will still show, and the items just won't show inthe availability field.
04:59 wajasu if we hide the branch, there cannot be a lost or non-lost  item from another branch or else the biblio will still show.
05:00 wajasu i deleted all the items in my test DB for a biblio and no it shows inthe results but No copies available.
05:01 mtompset bug 7960
05:01 wajasu now if i click facet "currently avialable items: , that one disappears.
05:01 huginn Bug http://bugs.koha-community.org[…]w_bug.cgi?id=7960 enhancement, P5 - low, ---, oleonard, NEW , choice to not show the text labels for item types
05:01 mtompset bug 7690
05:01 huginn Bug http://bugs.koha-community.org[…]w_bug.cgi?id=7690 minor, P5 - low, ---, fridolyn.somers, CLOSED FIXED, hidelostitems not working in Normal Display
05:05 wajasu hiding lost items is not really what i had to fix.  i was untangling OpacHiddenItems behavior because i had to check all items to see if they were hidden.
05:06 mtompset if OpacHideItems hides all but 1, and the 1 is lost and the syspref is set, the biblio will show. That is correct?
05:07 mtompset if OpacHideItems hides all but 1, and the 1 is lost and the syspref is set to hide, the biblio will show. That is correct?
05:07 mtompset if OpacHideItems hides all but 1, and the 1 is lost and the syspref is set to show, the biblio will show. That is correct.
05:08 wajasu if there is one item for a biblio that is not hidden, the biblio will show.
05:09 mtompset right, but what about the first case I said?
05:09 mtompset if OpacHideItems hides all but 1, and the 1 is lost and the syspref is set to hide, the biblio will show. That is correct?
05:10 wajasu the hidelostitems syspref  just hides the items ( Copies available  Centerville (1), Midway Lost(1))
05:10 wajasu not the biblio.
05:11 mtompset whereas the OpacHideItems is (in my understanding) intended to hide the biblio for 100% hidden items.
05:11 wajasu YES!  Thats what I coded.
05:11 wajasu thats why i decouple Hidden from lost in the code.
05:12 mtompset AH... that's why you removed the hidden by lostitems block
05:13 wajasu if i put  all my branches in OpacHiddenItems as  homebranch: [CPL, MPL]   i'll get no results.
05:13 wajasu yup.
05:13 mtompset The hideatopac_count didn't apply to hiding the biblio.
05:14 wajasu because i could not short circuit the "next" to keep examining ALL the items to get the count correct.
05:15 wajasu and the "next" outside the items/fields loop is for when I want to hide the biblio.
05:15 mtompset Okay... but I don't understand why you removed the next in the OpacHidden block?
05:16 wajasu for each item, the C4::Items::GetHiddenItemnumbers seems to get one item  ($hi)  and we need to check each item to incrment $hideatopac_count
05:17 mtompset Yes, but your inside the foreach item loop.
05:17 mtompset next goes to 1859.
05:17 mtompset (as far as I know)
05:17 wajasu only when that count matchs  items_count after the loop, can we know all items are hidden
05:18 wajasu line 1864
05:18 mtompset next doesn't break out of the foreach loop, does it?
05:18 wajasu next doesn't break.  it jumps to beginning.
05:19 wajasu you might be right.
05:19 mtompset I'm pretty sure the next should be there.
05:19 mtompset And it won't skip items.
05:19 wajasu i'll add it real quick. and check
05:20 wajasu it will save some processing
05:21 wajasu it will also keep from enriching the $item, which won't matter for a hidden one.
05:22 mtompset that was the point. :)
05:22 wajasu at one time i had code at the end of the loop that i needed to process, but don't need to now.
05:24 wajasu i started to pass a hashref of @hiddenitems back to opac-search.pl but dindn;t need it.   I fixed the total count using @newresults in opac-result.pl
05:25 wajasu so you see that all i am doing now is not returning the biblio if all items are hidden.
05:25 wajasu and the counts are correct.  no 404 not found.
05:26 mtompset Actually... no the counts are wrong.
05:26 mtompset They are a different wrong than currently, though. :)
05:27 mtompset currently: page 1 = 13, page 2 = 13, page 3 = 13.
05:27 wajasu well maybe we just have a paging counter issue to fix now.
05:27 mtompset 11909: page 1 = 0, page 2 = 0, page 3 = redirect.
05:27 wajasu Oh.  we can fix that.
05:27 mtompset wajasu: No you can't.
05:28 wajasu why not
05:28 mtompset Because we only return maxRecordsForFacets records.
05:29 mtompset There is no way to actually count all the records and determine if they are hidden or not unless we return ALL the records.
05:30 mtompset and returning all the records is a bad thing... could be millions of records which would cause agony.
05:30 wajasu do you need to determine if they are hidden?
05:31 mtompset to have an accurate count: yes.
05:32 mtompset now, if there was a way to create an index and update it every time
05:32 mtompset OpacHiddenItems is modified...
05:32 mtompset and then somehow use that ... perhaps it would be possible.
05:33 mtompset But I figure the people working on getting Zebra to return the facets would be more likely to solve the problem.
05:33 wajasu when Search.pm returns  @newresults, those won't have hidden ones.  but the paging code, which is in opac-search.pl, I think, just needs to page the newresults.
05:33 wajasu so maybe i need to explore the facets
05:34 mtompset Because if Zebra accurately returns the facets (hidden/not hidden), then counting is most definitely possible.
05:35 wajasu now i see what you are talking about.  i've never used/seen dsiplayFacetCount.
05:36 mtompset That's why I don't think the last part of your patch is right.
05:37 wajasu hey, thats new to me.  counters onthe facets.
05:37 mtompset I rather have it say 13 consistently wrong.
05:37 mtompset than a changing number each page.
05:38 mtompset Also... with the 0,0,redirect... what if there was 1,0,1? It would redirect on page 1, and you'd never know about page 3.
05:38 wajasu what i am finding is alot of code in opac-search.pl probably should be in Search.pm.
05:39 mtompset perhaps.
05:40 mtompset Also, the second part of your patch... remember facet size and page result size can be different. So the second part is wrong too.
05:40 mtompset I do, however, agree that the hidden because lost should be removed.
05:41 mtompset or at least a different counter.
05:41 mtompset then we could hide based on 100% opachiddenitems, 100% lost items, or some mix.
05:42 mtompset Because surely someone somewhere will want a different scheme.
05:42 mtompset putting the section back with the next will also be an optimization as well. ;)
05:42 wajasu i was coding things so that  @hiddentiems would be returned to opac-search.pl.   then one can get the count
05:43 mtompset That would be good, but you could key it on reason.
05:43 mtompset $reason->{'OpacHiddenItems'} = @ArrayOfHiddenItems;
05:44 mtompset $reason->{'HiddenLostItems'} = @ArrayOfHiddenItems;
05:45 wajasu in Search.pm  we could put  $oldbiblio->{hideatopac_count} = $hideatopac_count;  and grab it off of any biblio.  messy, but possible.
05:46 mtompset Whatever floats your boat. It's approaching 1am... (12:46). I think I should call it a day.
05:46 wajasu what counts do we need?  Authors, librarires, itemtypes, locations, series, topics?
05:46 wajasu thanks for the effort.
05:47 wajasu put it to failed, and i'll see about facets followup patch.
05:47 mtompset I moved it to ASSIGNED.
05:47 mtompset It's much happier that way. ;)
05:47 wajasu ok
05:48 mtompset I don't like FAILED, since I didn't test it, but rather eyeballed it. :)
05:48 mtompset Well, did test... just not enough.
05:49 wajasu i just did not concern myself about facets, more to learn.
05:50 mtompset Don't stare too hard into C4/Search.pm -- it will drive you crazy.
05:50 mtompset Have a great day/weekend, #koha wajasu.
06:48 drojf joined #koha
07:00 cait joined #koha
07:06 drojf good morning cait ;)
07:06 cait good morning drojf :)
07:08 cait @wunder drojf
07:08 huginn cait: Error: No such location could be found.
07:08 cait oups
07:08 cait @wunder Marseille
07:08 huginn cait: The current temperature in Realtor, CABRIES, France is 3.7°C (8:08 AM CET on March 08, 2014). Conditions: Clear. Humidity: 60%. Dew Point: -3.0°C. Windchill: 4.0°C. Pressure: 30.21 in 1023 hPa (Steady).
07:08 drojf i am not a location :D
07:09 cait yousure?
12:25 Dyrcona joined #koha
14:34 tcohen joined #koha
14:37 tcohen morning
14:37 tcohen need a Perl guru
14:37 tcohen :d
14:38 jcamins tcohen: I need vanilla paste and glucose.
14:38 tcohen i'm trying to mock something, and have the situation that the function i mock is cached by the use of "our" in the package
14:39 tcohen should I unload the module and load it again?
14:39 tcohen is there a way to circumbent that?
14:39 jcamins I think you can prevent Test::MockModule from loading the package.
14:39 jcamins But I'd vote for just fixing the package, because that sounds like bad coding.
14:40 pastebot "tcohen" at 127.0.0.1 pasted "jcamins:" (30 lines) at http://paste.koha-community.org/153
14:42 tcohen what do u think of it jcamins?
14:42 jcamins Oh, I see.
14:43 jcamins I thought you meant a function pointer was being stored in a package variable.
14:43 jcamins Yeah, unload and reload.
14:53 * cait waves
15:00 tcohen does it even make sense to cache that jcamins? isn't it just called once each script run?
15:01 jcamins tcohen: probably not, actually, no.
15:01 tcohen i'd rather cache it using memcached for instance
15:02 jcamins Agreed.
15:03 jcamins There's a Koha::Cache class that might help.
15:05 cait is that the one gmcharlt wanted to revive?
15:05 * cait thought she had seen a bug
15:05 jcamins That's the one that I wasted my time on making it work.
15:06 * tcohen belives that 'our' might have been introduced by a Plack user?
15:06 cait :(
15:07 cait tcohen: yeah, all the ours came with the first plack introduction
15:07 jcamins tcohen: yeah, prior to that all the module-level "caching" didn't do anything anyway.
15:07 jcamins "our" is the wrong solution, however.
15:08 cait someone tell me to pack please?
15:08 tcohen cait, you need to stop what you're doing and pack
15:08 jcamins Go pack.
15:08 wajasu pack it up!
15:08 * tcohen sends cookies to people who packs
15:08 wajasu plackup!
15:09 tcohen heh
15:10 cait hehe
15:10 * wajasu just read through all the search + facets + federated code. "Who  knew?"
15:10 cait thx :)
15:12 * wajasu ponders about facet counts and result of OpacHiddenItems
15:13 tcohen wajasu: you're about to enter into a life of pain and obscurity?
15:18 wajasu i did not realize that folks search across multiple servers, i.e. federated.
15:20 * tcohen is about to git reset --hard and forget about this Biblio.pm tests
15:34 jcamins I always use git stash for those kinds of things.
15:35 jcamins Then I don't have to feel bad about having decided to entirely abandon a project, because by the time it's fully abandoned, I've forgotten.
15:36 * wajasu Search.pm + opac-search.pl      O what a tangled web we weave.
15:41 tcohen heh
15:45 wajasu so i guess we only build facets for our own server? i.e. biblioserver
15:51 tcohen wajasu: i don't think so
15:51 wajasu ok. that helps.
15:52 tcohen there's a big loop on each server
15:52 tcohen loop over the servers array
15:53 tcohen but I have the feeling that we only use [0] in searchResults (which is used for the final rendering)
15:54 tcohen leaving for lunch
15:54 tcohen bye!
17:05 AmitG joined #koha
17:06 AmitG hi
17:06 AmitG alex around
17:06 AmitG hi
17:11 thd-away joined #koha
20:45 trea joined #koha
21:19 cait left #koha
22:53 NateC joined #koha
23:06 sophie_m joined #koha
23:21 tcohen joined #koha
23:31 sophie_m left #koha
23:44 tcohen joined #koha

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