IRC log for #koha, 2008-07-08

Today | Next day → | Search | Index

All times shown according to UTC.

Time Nick Message
17:21 ricardo gmcharlt: Thanks Galen!  :)
17:39 thd nengard: check your email for a message with the subject line "using open source" in answer to your question
17:39 nengard thd was just wondering what the heck that was ;) I've actually read that before.  Thanks for sending it
17:59 thd I made some custom authorities frameworks for LibLime.  I do not see them in git.  What should I do about updating them for the latest MARC update, that is, where should they go?
18:00 gmcharlt thd: installer/data/mysql/en/marcflavour/marc​21/mandatory/authorities_normal_marc21.sql  , I would think
18:01 per comments in header of that file, that is a version that you edited for LL
18:03 ricardo Time to go home. Take care erveryone!  :)
18:09 thd gmcharlt: that is the default with the full standard, the others were two progressively smaller subsets.  Neither of them were exceptionally small so that they would not miss anything important.
18:10 ryan: are you here>
18:10 gmcharlt ok, then it's a question for ryan or kados: where did the subsets go?
18:10 kados hmmm, subsets?
18:12 thd: there have been changes to both authority and bib frameworks since you last contributed
18:12 thd: some patches to bugs in them ...
18:12 thd kados: I know: I am working from the git version and I made a diff
18:13 ryan should know about the subsets of the authorities framework.  Is he around?
18:13 kados thd: one concern I have about the current implementation of frameworks is that there is no mechanism to update frameworks on live systems
18:14 ie, post-installation
18:14 s/update/auto-update/
18:14 thd kados: drop the table and repopulate ;)
18:14 s/table/tables/
18:14 kados thd: doesn't work if they've made local changes
18:15 thd kados: that requires a sophisticated solution
18:15 kados yep
18:15 thd kados: that could be solved with two or three days thinking
18:16 ... and investigating.  yet, who has that much time at this point.
18:16 ?
18:16 gmcharlt thd, kados: another thing for 3.2 or 3.4, perhaps - frameworks should distinguish between standard, "app profile" a la OCLC's version, local fields, and decisions about which fields to display in a particular format view
18:17 kados *nod*
18:18 thd gmcharlt: ultimately we need a more flexible model for the frameworks.  Updating the variants with many changes is not very time efficient.
18:18 gmcharlt thd: agreed
18:18 also, now that LC's MARC website is providing the MARC21 format defines in something half-way parseable, drawing from that could be automated
18:19 kados definitely
18:19 thd gmcharlt: the problem is that only has the current LC standard which is vastly under-inclusive for the real world
18:21 there are records at some libraries which still have data in fields which were made obsolete when AACR2 was introduced in 1978
18:22 the current LC standard maybe OK for creating a new original record but not for interpreting existing records in the world
18:22 gmcharlt thd: ok, but automating the current version is a start, and it would actually be useful to have a mode that indicates that your records have obsolete fields
18:24 thd gmcharlt: I have noted all obsolete fields and subfields which only appear if already populated with the notation [OBSOLETE]
18:25 gmcharlt thd: that's good for display, but an enhancement framework structure might actually tie some functionality to those fields
18:25 e.g., "warning, warning, AACR1 in use"
18:25 thd gmcharlt: In addition to obsolete there are widespread non-LC conventions in use at OCLC, etc.
18:27 gmcharlt: yes, we need an enhancement with better functionality and where framework variants can be mere diffs
18:29 soon we will have warning AACR2 in use :)
18:31 gmcharlt thd: hehe - shouldn't that be "RDA in use, despair all ye who enter here"?
18:32 thd The RDA MARC working group proposed to reuse 011 which was made obsolete in 1992 and I think the MARBI committee roundly rejected them
18:33 gmcharlt good - I was not happy that there were proposals to reuse tag numbers
18:34 atz thd: sounds like OO/inheritance approach to frameworks
18:34 thd Despair especially much if you cannot afford to subscribe and there is no one time purchase price for a hard copy version
18:34 atz (variants as diffs_
18:35 thd atz: the problem is that even with clever use of regular expressions copying changes into every framework variation takes all day.
18:36 atz yeah, it must
18:36 gmcharlt thd: and loud complaining if we can't republish enough of the RDA spec to set up MARC frameworks (and presumably their RDF equivalents in the future)
18:40 thd gmcharlt: I intend to suggest on the RDA-L list that the co-publishers should look into funding publication in a manner similar to the work of wc3.org .  Compare the market success of library standards to web standards.
18:40 gmcharlt thd: good idea.  I would personally endorse that.
18:41 atz rch: added /home/ican/ican_sip_users.sql
18:42 kados atz: wrong channel ? :-)
18:42 thd s/wc3/w3c/
18:42 atz kados: true... i've got code4lib up today and it's throwing off my sense of tab-geography
18:43 thd gmcharlt: RDF implementations are not quite ready for MARC but we need to make them ready.
18:44 gmcharlt: Ed Summers LCSH.info is woefully inadequate because of the limitations of SKOS
18:46 kados hdl: you still around?
18:47 thd gmcharlt: DCMI must have a commitment to solving those issues because they have committed to mapping RDA/MARC in RDF.
18:47 kados paul: ?
18:47 thd ryan ?
18:48 ryan hi thd
18:48 thd hello ryan
18:49 ryan: I have been updating frameworks to the latest MARC update and want to know what to do about the subsets of the authorities frameworks which I made for you
18:49 ryan: I think they may never have been put in git because kados never chose a place for them
18:50 ryan thd: if they  are up to date, would be great if you could commit them
18:50 thd: do you see any reason not to include them in 3.0 ?
18:50 thd ryan: maybe tonight but where would I put them?
18:51 kados thd: these are optional marc21 authorities frameworks?
18:51 thd kados: yes
18:52 kados thd: if they compliment rather than supercede the existing marc21 authorities frameworks, they belong in:
18:52 installer/data/mysql/en/ma​rcflavour/marc21/optional/
18:52 ryan nstaller/data/mysql/en/marcflavour/marc21/optional
18:52 yeah, what he said
18:53 kados for 3.2 we may want to consider a mechanism for selecting one of many optional mandatory frameworks
18:53 if that makes sense
18:53 thd kados: has the issue of visibility using the hidden column been fixed for the authorities record editor
18:53 ?
18:53 kados thd: I'm not sure
18:53 thd: check bugzilla?
18:54 thd kados: was such a bug filed or do you mean file then check?
18:55 kados thd: check for a bug(I think there was one)
18:55 if there wasn't, please file one if the problem still exists
18:55 (feel free to check on koha.liblime.com
18:56 thd kados: in relation to bugs, are you going to announce a new release date or announce that a date will be fixed when the lingering important bugs have been fixed along with any newly reported important bugs?
18:57 kados thd: good question
18:57 thd: we really need to get this release out
18:57 thd: so we can start working on 3.2
18:58 thd kados: yet, it seems much less polished than 2.2.X for all its advances.
18:58 kados thd: are you serious?
18:58 thd: I completely disagree :-)
18:58 thd: it's much more polished than a koha release has ever been, in so many ways
18:59 thd kados: yes I ma very serious about some user interface issues which I never saw on late versions of 2.2.X
18:59 ryan thd: you are referring above to bug 2206 ?  yes, it's been fixed.
18:59 kados i agree we have a lot of improvements that could be made, but I really don't think you can compare it to a 2.2.x release
18:59 atz thd: there are bugs, but the install process alone is so much better than 2.2
19:00 kados 2.2.9 has a lot more blocker bugs IMO than 3.0RC1
19:01 thd kados: I mean things like when I tab through the fields for the Biblios record editor I get stuck and tab stops working.
19:02 kados: Many issues relating to using the keyboard do not work as well therefore it takes much longer to do real work.
19:03 ryan: 2206 was not the issue for controlling subfield visibility in the authorities record editor using the authorities frameworks.
19:04 ccatalfo thd: are you experiencing this tabbing issue in the Biblios editor in any particular circumstances?  I'm aware that tabbing is not working from authority controlled fields.  Is that what you've seen?
19:05 ryan thd: you are referring to the standard MARC editor, right?  not Biblios ?
19:05 thd ccatalfo: I was hoping that it was masking some function for the authority controlled fields.
19:06 ccatalfo thd: just to clarify, you do mean the Biblios editor, not the cataloging editor in Koha?
19:06 thd ryan: I think there are user interface problems for doing real work in all of the record editors which did not exist in 2.2.X
19:08 ccatalfo: I do not get stuck on the authority controlled fields in the semantically labelled editor.  I get stuck elsewhere in that editor but I expect less productivity from that editor in some ways.
19:09 ccatalfo: so yes, I was addressing you about the Biblios record editor.
19:09 ccatalfo thd: ok, thanks.  Are there other places tab isn't working, then?  Or it the authority controlled fields?
19:10 thd ccatalfo: I noticed the authority controlled fields but did not test extensively.
19:10 ccatalfo thd: ok, good to know.
19:11 thd ccatalfo: Is the authority controlled field block there as some anchor for a nifty authorised value filling feature which has not yet been implemented?
19:13 ccatalfo thd: basically, yes.  It's meant to query an sru authorities server for authorized values for that field.  The querying part is implemented but we don't have it pointing at a valid server at the moment (on koha.liblime.com)
19:14 thd ccatalfo: Yet I should be able to move quickly through the field even if I do not want to change a particular authorised value or if my connection is down, etc.
19:15 ccatalfo thd: you're right - that tab not working is definitely a defect
19:15 thd: it ought to tab right through it even thought it is an ajax dropdown
19:18 thd ccatalfo: Also have you seen that the latest MARC 21 update provides an official place to put an authority record control number such as 650 ## $0 (DLC) sh1356358 $a some subject ?
19:20 ccatalfo thd: no, I had not seen that.  that's very good to know.  So I can pull in that authority control number as well as the text, and not display the $0...
19:20 am i correct in remembering that koha is using a different subfield for linking, $9?
19:22 thd ccatalfo: Such nifty things were added for MAB2 format integration with MARC 21 for Germany and Austria
19:23 ccatalfo: You should still use $9 for Koha internally with the internal control number so the Koha code will use the authority correctly.
19:24 ccatalfo: However, if you also use $0 we can update authorised values using the LCCN from the NACO and SACO database.
19:27 ccatalfo thd: thanks, this is all good information.  FYI, at the moment, Biblios does not do any linking with koha $9.  So it's not suitable for authority work within koha as things stand now.  But, these are good things to keep in mind when we do implement better authority integration.
19:27 thd ... provided we have real authority records which have an LCCN or other number from a standards body
19:29 ccatalfo is the koha's $9 the control number from the authority record?  so that $0 would be the same as $9?
19:30 gmcharlt thd: re $0, has anybody come up with a decent proposal to handle subject headings that are linked to one main authority record and N subdivision records?
19:32 thd ccatalfo: At French Koha libraries $9 taken from the Koha record ID and they build primitive authority records in most cases internally from existing values in authority controlled fields.
19:33 gmcharlt: $0 is repeatable but under review for how well that works.
19:34 gmcharlt: The UNIMARC equivalent $3 I discovered a week ago was made repeatable to conform to practise at French libraries for that very issue.
19:35 gmcharlt: I asked that very question citing divergent practise on the MARC list and while no one answered on list I received a private reply from the LC member of MARBI that the issue needs studying.
19:37 gmcharlt: http://listserv.loc.gov/cgi-bi[…]L=marc&T=0&P=6868
19:40 gmcharlt: German practise which I did not quote fully uses adjacent repeated fields instead of attempting to accommodate a complex expression in one subdivided field.
19:42 gmcharlt thd: gotcha.  thanks
19:42 ccatalfo if only we just put the linking control number in there instead of the text...would be so much easier...
19:44 thd ccatalfo: The DC-RDA group is working on doing just that using URIs but we still have to be backwards compatible until the revolution comes
19:47 ccatalfo: Also, the last MARBI meeting rejected all RDA working group proposals for using URIs for vocabulary such as relator terms specifying the role of an author as illustrator, editor, etc.
19:48 ccatalfo thd: interesting
19:49 thd ccatalfo: The conservative MARC people want textual strings, which are necessarily language specific, and make for poor cross-national record sharing.  However, the RDA working group proposal was technically deficient for accommodating both interests and deserved to fail.
19:59 kados paul: do you know if hdl is available?
19:59 paul hi kados.
20:00 kados hi :-)
20:00 paul nope. He's in holiday for 2 weeks
20:00 kados paul: so I guess we won't see any serials patches then :-)
20:00 ok, that answers that question
20:00 paul right
20:00 kados paul: have you seen the traffic on koha-devel about delaying 3.0 release?
20:01 paul: it seems some people don't want me to release it with blocker/critical/major bugs
20:02 paul yes, I've seen.
20:02 kados thoughts?
20:03 paul but my opinion is unchanged : we have delayed 3.0 really too too too much. and it start to be a real problem for us. + I strongly think we can continue debugguing for months or even years without releasing anything.
20:04 so, I'd said : let's go with 3.0, and continue with 3.0.1, even if it's 1 month after 3.0.0
20:04 + branch head for 3.2
20:06 hehe... hdl_laptop seems to be unable to be in holidays ;-)
20:07 gmcharlt that or the laptop is working, but hdl is still on vacation ;)
20:07 kados hdl_laptop: are you really here?
20:13 davi Trying to access the intranet.koha URL I get:
20:13  nothing on the browser
20:14 and the below error in logs:
20:14  DBD::mysql::st execute failed: Table 'koha.language_subtag_registry' doesn't exist at /build/kohaclone/C4/Languages.pm line 178.
20:14
20:15 I can not execute the web install because when I access to  intranet.koha , that is to say 8080
20:16 hdl_laptop kados yes.
20:16 davi I get such error with nothing on the browser.
20:16 hdl_laptop but on holidays.
20:16 paul hdl_laptop: i've written that already to kados ;-)
20:16 kados hdl_laptop: paul has confirmed, but I just wanted to check with you -- you will not have serials patches ready for 3.0, right?
20:17 davi Can I execute some script to do the web installation?
20:17 hdl_laptop kados: I have sent some.
20:18 kados davi: no, it has to be done via a web browser, but you could use lynx on the server if you only have command-line access to the server
20:18 davi: that's almost like running a script ;-)
20:18 gmcharlt hdl_laptop: are the subscription-add patches you sent me and ccatalfo potentially for 3.0, or do you think they'll need more work and should be for 3.2
20:18 hdl_laptop: I haven't had a chance to look at them yet
20:18 hdl_laptop gmcharlt: honestly,
20:19 i would have LOVED to get them ready for 3.0
20:19 And would jquery and all the javascript stuff not be such a new thing for me,
20:19 it would have been ready.
20:21 But since it adds 2 tables, changes subscription table and add a new dependency (Roman)
20:21 it may be safer.
20:21 chris morning
20:22 paul hi chris. time to go to bed for frenchies... (10:30PM)
20:22 gmcharlt hi chris
20:22 davi kados, lynx does not work neither. I am checking with    telnet localhost 8080
20:22 # telnet localhost 8080
20:22 Trying 127.0.0.1...
20:22 Connected to localhost.
20:22 Escape character is '^]'.
20:22 GET /
20:22 Connection closed by foreign host.
20:22 kados davi: maybe an issue with your apache conf?
20:22 chris goodnight paul
20:22 and hdl
20:23 davi   Note: opac-main.pl get 90% CPU
20:23 kados, I am going to check the apache conf. Thanks
20:24 hdl_laptop kados : I sent some patches on Friday, which really fix some broken things.
20:24 thd kados paul: If you would be OK releasing 3.0.1 in a month I would be understanding of your need to release 3.0 long ago and say that it is OK to release now or a week from now.
20:25 paul (sorry, tired & late here)
20:25 thd paul: I understand your need to release 3.0 now
20:27 paul: If you release 3.0 now and release 3.0.1 in a month, then I think releasing 3,0 now or within a week is good.
20:28 paul: If 3.0.1 would take three months to issue on some schedule, then I think that you should wait for some more bug fixes before releasing 3.0.
20:30 paul kados: When I made the start of an argument like that earlier, slef and owen did not think that standards should be lowered to satisfy business needs.
20:30 paul thd: that's exactly the kind of things that will make koha 3.0 NEVER be released... after standard, we will find something else to improve, then something else, then ...
20:31 many project died from "search for perfection"
20:31 kados paul: examples?
20:31 chris i think we need to get back to the release early release often mantra
20:31 paul I don't speak as a business man here.
20:31 chris: ++
20:31 chris heck apache 2.0 is at .63
20:31 and 2.2 is at .9 already
20:31 thd paul: I agree with you
20:31 paul look at kde 4 !
20:31 gmcharlt if we do a bugfix 3.0.1 in a month's time, that might be managable, but it will give us at least one complicated DB schema change to work through, if we consider the serials bugs that hdl is working on to be blockers that should be fixed for 3.0.1
20:31 paul it has been released stable, and it was really unuseable !
20:32 kados chris: did apache release a product with blocker bugs?
20:32 chris yes
20:32 thd paul: my only worry is that once 3.0 is released 3.0.1 would not come out for months
20:32 kados blockers? really? like 'virtual hosts don't work', something that major?
20:32 davi kados, Accessing to the koha:8080 via lynx I have got:
20:32  production mode - trapped fatal error
20:32 chris yep, they have dev branches and they release from them
20:33 kados chris: ahh, well we've done releases too by that metric
20:33 chris well only in the last few months
20:33 thd paul: I think that you should issue many minor releases quickly until the serious bugs have been fixed.
20:33 chris there was years in between
20:33 kados chris: but I seriously doubt they would have released stable versions of apache with blocker bugs
20:33 paul kados: except 2297, all other blo are "long time" blo bugs, that are a problem only in some circumstances.
20:33 davi kados, mainpage.pl and install.pl was executed before I got such error at lynx
20:33 paul for example ':' character blocks search
20:33 or circulation reports too big
20:34 gmcharlt paul: well, as some larger and/or more sophisticated libraries adopt Koha, those bugs become more critical
20:34 paul some other one just need a warning in release notes to be workarounded, for example : 2178
20:34 thd normalise the search query and you solve that bug quickly for some practical purpose even if you have not traced it
20:35 davi kados, You were right about my Apache configuration was broken.
20:35 kados, I get "production mode - trapped fatal error" after fixing the Apache configuration. Now it is better but do not work yet.
20:35 kados davi: what's the rest of the error?
20:35 davi: anything in the apache logs?
20:35 paul gmcharlt: I agree. It don't mean those bugs must be ignored. Just that they should not prevent us from releasing.
20:36 thd paul kados: so release and keep the pace of fixing the serious bugs and bugs which scare librarians
20:37 paul OF COURSE !
20:37 when I was 2.2RM, I used to release a subversion once every quarter.
20:37 that was my goal, and I succeded to do it !
20:37 thd paul kados: In the real world no one trust an X.0 release of anything anyways.
20:37 kados thd: *nod*
20:37 paul thd++
20:38 kados paul++ for releasing on schedule
20:38 gmcharlt thd: that's no excuse, though
20:38 thd paul: I think that 1 quarter is too long to wait for 3.0.1 though.
20:38 paul to continue feedback with my experience : the DB stabilisation is the 1st thing to do to get a stable software
20:38 thd: I agree.
20:39 gmcharlt for 3.2, I'm happy releasing checkpoints releases more often
20:39 paul thd: once a quarter is OK once it's really stable
20:39 so, the rule is :
20:39 gmcharlt but there's a question: how long do we want to keep up 3.0.x?
20:39 thd exactly I think paul just said it perfectly
20:39 kados because I think it will be too much overhead for this community
20:40 paul I agree to have BibLibre take that role.
20:40 thd kados: you want to go directly to 3.2 from 3.0 with only 3.1 between them?
20:40 gmcharlt paul: which role? 3.0 maintainer?
20:40 paul yep
20:41 kados I sincerely hope that as we move forward as a community we not hide our heads in the sand about major issues like this
20:42 chris are you suggesting we are now?
20:42 kados chris: yep :-)
20:42 paul yes, but is you want to move forward as a community, you should not ignore other's position ?  (sorry to be rude)
20:42 kados I think there's a lot of pressure to release software that's not ready
20:42 chris ah well, your entitled to your opinion
20:42 thd The caveats should be noted prominently without scaring people just so they do not get the wrong idea and assume the claim is that all works perfectly and never look again when they find that it does not.
20:42 kados as can be seen on the list traffic
20:42 chris: not just my opinion anymore :-) others on-list have expressed similar sentiments
20:43 atz if the warnings were there in the interface that might matter
20:43 otherwise everybody will expect it to work
20:43 kados thd: true, that may be what we have to do for 3.0
20:43 atz: *nod*
20:43 atz regardless of what INSTALL says
20:43 paul mmm... kados, i've seen only 2 ppl expressing their opinion. I could list 10 french libraries that want 3.0 now...
20:43 + when releasing you can express the limits of the release & the caveats.
20:43 davi kados,  (1) I execute   lynx http://localhost:8080   (2)  mainpage.pl and install.pl get CPU  (3)  lynx shows "production mode - trapped fatal error"
20:43 kados paul: and I could list 100 US libraries that want 3.0 to be stable ;-)
20:43 thd My concern is only too keep the good reputation of Koha.
20:43 kados thd: *nod* ... mine too
20:44 davi: no errors in the apache logs?
20:44 davi kados, No error at /var/log/apache2/error.log  or at  koha-dev/var/log/koha-opac-error_log   !!!
20:44 paul So, release, & specify what is know to be unstable or working poorly.
20:44 that's better than not releasing imo
20:44 atz right now I'm servicing client tickets for 3 reports in 3.0 that just break
20:44 thd wait too long to release and you also loose reputation
20:44 paul thd++
20:45 kados davi: check the other koha log
20:45 atz davi: your errors should be in the locations specified by your koha .conf file as ErrorLog
20:46 davi koha-error_log
20:46 kados davi: that's the one
20:47 davi  install.pl: DBD::mysql::st execute failed: Table 'koha.language_subtag_registry' doesn't exist at /build/kohaclone/C4/Languages.pm line 178.
20:47  install.pl: Can't locate object method "generate_id" via package "CGI::Session::ID::" (perhaps you forgot to load "CGI::Session::ID::"?) at /usr/local/share/perl/5.8.8/CGI/Session.pm line 74.
20:47 kados ahha, there's the culprit I bet
20:48 davi source code bug   or   installation bug?
20:48 Maybe I missed the installation of some Perl package?
20:49 Let me try to install it
20:49 atz looks like unpopulated code trying to make the serializer a variable?
20:49 paul davi : there's a problem with CGI::Session package, that has been splitted recently by the package maintainer.
20:50 you must also install another one, I don't remember which one, sorry (11PM in France... i'm tired)
20:50 atz CGI::Session::yaml perhaps ?
20:50 paul that one being a problem that could delay the release...
20:50 davi thanks paul, and good night
20:51 paul mc has more infos about that
20:51 atz yeah, that one is not directly our fault, but a serious problem anyway
20:51 paul yep
20:51 that at least need some warnings in Release Notes !
20:51 thd Anyways, I will have the MARC 21 frameworks up to date by tomorrow and I assume gmcharlt is fixing the unreported general data loss bug which had concerned me.
20:52 paul really time to go to bed for me.
20:52 I'll be back tomorrow afternoon (FR afternoon)
20:52 thd good night paul
20:54 davi Will work if I download and overwrite installing an specific version of the CGI::Session package?
20:55 atz davi: it might
20:55 davi I will ask  mc  tomorrow for the CGI::Session package version to install.
20:55 atz like the slightly older version
20:56 davi Shit!  I have already installed it from   http://search.cpan.org/CPAN/au[…]ssion-4.30.tar.gz
20:56 as the Ubuntu guide advices
20:56  http://wiki.koha.org/doku.php?id=ubuntu_gutsy
21:02 I have installed CGI::Session 4.30
21:02 It is correct as documented at the Ubuntu installation guide.
21:03
21:03 The first error line I get at   koha-error_log  is:
21:03  mainpage.pl: DBD::mysql::st execute failed: Table 'koha.auth_types' doesn't exist at /build/kohaclone/C4/Koha.pm line 340
21:03 So maybe the source of the error is at another location
21:04 gmcharlt davi: now that you have CGI::Session 4.30, there are a couple additional modules you can install from CPAN that shoudl complete things
21:05 davi: one being CGI::Session::Serialize::yaml
21:05 davi let me do it, thanks
21:06 kados gmcharlt: the cgi-session issues alone probably need to be documented explicifly
21:06 gmcharlt: shoudl we add notes to the installation docs?
21:06 gmcharlt kados: agreed
21:06 yes
21:07 kados gmcharlt: so installation of CGI::Session::Serialize::yaml with version 4.30 fixes the issues, should we just require that now?
21:08 gmcharlt yes, that would be a minimum
21:08 there may be something more - lemme double-check quickly
21:13 ricardo Hi all
21:17 kados davi: did it work for ya?
21:17 ricardo: hiya
21:17 ricardo kados: Hi Joshua!  :)
21:17 davi kados, I Qemu virtual machine is very slow :-)
21:18 My*
21:18 I report to all you with the result, after installing CGI::Session::Serialize::yaml
21:18 I will*
21:20 ricardo davi: That solved my CGI::Session problems for me (installing CGI::Session::Serialize::yaml is what I recommend in my Installation Tutorial of Koha 3 in openSUSE 11.0 at http://wiki.koha.org/doku.php?id=opensuse_11.0
21:21 kados: BTW - I'm reviewing that tutorial (and doing some minor changes) with a fresh installation of openSUSE 11.0 final (I'll use the latest Koha development version for it).
21:21 kados ricardo++
21:21 ricardo :)
21:23 I'm guessing that I won't have the problem with the missing dependency for Algorithm::CheckDigits this time since MJ Ray promptly corrected that (so that section will probably be deleted from the wiki / INSTALL.opensuse file)...
21:23 davi I take note of your guide ricardo!  Maybe I will move from Ubuntu to OpenSUSE
21:24 ricardo davi: LOL! Well, I was not expecting my Guide to have an impact *that* big, eheh... But thanks, I will take that as a compliment  :)
21:24 davi I am tiny :)
21:26 ricardo davi: eheh
21:27 On the other hand, I seem to recall hdl talking about some CGI::something::mysql  perl module that would be useful. I'll search the IRC logs for that now
21:30 davi ricardo, You guide is a lot longer than the Ubuntu one at http://wiki.koha.org/doku.php?id=ubuntu_gutsy
21:30 Your*
21:31 ricardo davi: True... I'm usually pretty detailed in my documentation (you might call it "boring"  ;-)
21:31 gmcharlt ricardo: "complete" is a friendlier word ;)
21:31 ricardo gmcharlt: Eheh... Thanks!  :)
21:33 (I had one internal documentation that included ALL the output of CPAN Module installations, file extractions and such for Koha AND had some screenshots... that went over a thousand pages, believe it or not...)
21:33 davi ricardo, It is lot of work but it is better when well detailed
21:34 ricardo davi: Well, at least I appreciate detailed info that/when it is relevant  :)
21:35 davi I get some failed tests when installing such package
21:35 t/g4_dbfile_json................ok 64/101       (in cleanup) Dumping circular structures is not supported with JSON::Syck at /root/.cpan/build/CGI-Session-4.20/blib/lib/CGI/Session/Serialize/json.pm line 18.
21:35 ...
21:36 t/g4_dbfile_json................dubious
21:36        Test returned status 255 (wstat 65280, 0xff00)
21:36 DIED. FAILED tests 84, 90, 92-101
21:36 ...
21:36 ricardo right
21:36 davi Can be go on?
21:36 Can I go on?
21:37 ricardo davi: Try to install CGI::Session::ID::MD5 first (I think... let me check)
21:37 davi ok, thanks
21:38 ricardo Yep. At least, it has that name ->  CGI::Session::ID::MD5
21:40 davi: Try to install that (from CPAN)...
21:41 davi I have installed all the packages from CPAN
21:41 ricardo BTW: In my tutorial I have these lines (that seem to have no bad consequences, but...):
21:41 davi    perl -MCPAN -e 'install Package::...'
21:41 ricardo davi: Oh OK :)
21:41 gmcharlt / kados (other):  in my tutorial I have these lines (that seem to have no bad consequences, but...):
21:41 davi you must correct me, please
21:43 ricardo "  t/Items.....................ok 2/28Use of uninitialized value $db_driver in concatenation (.) or string at /root/Koha/blib/PERL_MODULE_DIR/C4/Context.pm line 628.Can't connect to data source 'dbname=__DB_NAME__;host=__​DB_HOST__;port=__DB_PORT__' because I can't work out what driver to use   "
21:43 gmcharlt / kados: Is this really harmless.... or NOT?
21:43 gmcharlt ricardo: it will, in fact, be going away due to a patch from acmoore
21:44 ricardo gmcharlt: Ah great! Thanks for the feedback!  :)
21:44 gmcharlt: Do you know if that patch has already been submitted & approved?
21:44 davi: OK. Did you install CGI::Session::ID::MD5 already?
21:45 gmcharlt ricardo: the patch moving the items test was just approved
21:46 davi ricardo, My Qemu virtual machine is slow
21:46 ricardo gmcharlt: Very nice... Thanks for the info. So I'll grab again that latest snapshot development version. That will also improve (and "ease up" a bit) my tutorial.
21:46 davi: Hmmm.... Qemu runs under Linux or Windows?
21:46 davi GNU/Linux
21:46 Debian stable 'etch'
21:47 ricardo davi: OK... I use VirtualBox in Windows (for running openSUSE 11.0) and it is reasonably light-weight (much lighter than VMware, at least). In Linux, I use Xen. I never tried Qemu, though
21:48 davi ricardo, CGI::Session::ID::MD5    installed rightly. I am going now with  CGI::Session::Serialize::yaml  again.
21:48 ricardo OK. Good luck!  :)
21:49 davi ricardo, I will go with Xen when time allow me.
21:49 Your positive comment re-force my intention to try Xen.
21:51 ricardo davi: Well... Xen is included with SLES (SUSE Linux Enterprise Server), so that's what I ended up using  :)
21:51 But yes, I like Xen
21:52 davi I have and old AMD Athlon, so I think I have some problems to boot
21:52 the Xen hipervisor
21:53   with PAE or without PAE support.
21:53 ricardo davi: Maybe... I don't have any info regarding that (I run it on a typical Intel Pentium 4 - *playing mandatory Intel advertising sound*  ;-)
21:54 davi
21:54 Same error again:
21:54 t/g4_dbfile_json................ok 64/101       (in cleanup) Dumping circular structures is not supported with JSON::Syck at /root/.cpan/build/CGI-Session-4.20/blib/lib/CGI/Session/Serialize/json.pm line 18.
21:54 ...
21:54 t/g4_mysql......................skipped
21:54        all skipped: Couldn't establish connection with the MySQL server: Can't connect to data source '' because I can't work out what driver to use (it doesn't seem to contain a 'dbi:driver:' prefix and the DBI_DRIVER env var is not set) at t/g4_mysql.t line 44
21:54
21:55 The CGI::Session::ID::MD5 installation did not solve the CGI::Session::Serialize::yaml installation problem.
21:55 ricardo davi: One user reported that he managed to solved that JSON problem by downloading CGI::Session and installing it directly (instead of using CPAN):
21:55 http://lists.katipo.co.nz/publ[…]8-May/013842.html
21:55 See if that solves the problem for you
21:57 davi I have already installed an specific version of CGI::Session
21:57 I think the above error is different than the one exposed at such link.
21:58 Maybe I should set the CGISESS_MYSQL_DSN  environment variables before starting the installation of CGI::Session::Serialize::yaml ?
21:58 It reads:
21:58 ricardo davi: Hmmm.... It is related - http://lists.katipo.co.nz/publ[…]8-May/013815.html - but yes, it may be different
21:58 davi The simplest method is to use the standard "DBI_DSN/DBI_USER/DBI_PASS"
21:58 environment variables.
21:58 Otherwise, you can set these variables:
21:59 ricardo davi: CGISESS_MYSQL_DSN? First time I hear about that one
21:59 davi OK, so not needed. thanks.
21:59 So do you think  CGI::Session::Serialize::yaml  is actually needed?
22:00 gmcharlt yes, the YAML serializer is required
22:00 davi thanks
22:00 ricardo davi: Try to install the YAML Perl module... Maybe Ubuntu has a package for that?
22:02 acmoore announcing this again FYI: There's a news bot in #kohanews on freenode. It announces things like updates to git and bugzilla. I find it more convenient than using a RSS reader for those things. some developers may be intereseted in idling there. feel free to.
22:03 kados acmoore: probably would make sense to announce this on koha-devel too
22:03 davi That is the output of the installation:  http://rafb.net/p/jnik8F61.html
22:05 ricardo davi: Reading... Yep, sounds familiar, all right. I think that when I got this, I ended up doing a "force install CGI::Session". No, I don't like it either but it seemed to solve it  :-/
22:05 This is why I now install CGI::Session::Serialize::yaml "instead" of "CGI::Session"
22:07 davi Such paste is the output of CGI::Session::Serialize::yaml  not CGI::Session
22:07 so, you advise force it?
22:07  "force install CGI::Session"
22:07 ?
22:07
22:07 I have located this Ubuntu package:   libyaml-perl - YAML Ain't Markup Language (tm)
22:08 What do you think is better: install the Ubuntu package or "force install CGI::Session" ?
22:08 ricardo davi: I would try the libyaml-perl path first
22:08 davi thanks
22:13
22:13 not solved
22:13 I get again:
22:13 ricardo davi: Darn!  :(
22:13 davi  install.pl: DBD::mysql::st execute failed: Table 'koha.language_subtag_registry' doesn't exist at /build/kohaclone/C4/Languages.pm line 178
22:13  install.pl: Can't locate CGI/Session/Serialize/yaml.pm in @INC (@INC contains: /build/kohaclone /etc/perl /usr/local/lib/perl/5.8.8 /usr/local/share/perl/5.8.8 /usr/lib/perl5 /usr/share/perl5 /usr/lib/perl/5.8 /usr/share/perl/5.8 /usr/local/lib/site_perl .) at (eval 15) line 3.
22:13 ricardo davi:  Hmmm... But you don't get anymore the JSON error?
22:13 davi: Try now to install CGI::Session::Serialize::yaml
22:14 davi I just installed the  Ubuntu package
22:14 I have not tryed it install it form CPAN again
22:14 ricardo davi:  I'm guessing the libyaml-perl is for the YAML::Syck package, but I may be wrong. Let me see
22:14 davi I do not know if besides the libyaml-perl ubuntu package it is needed some other one
22:15 I have:
22:15  libyaml-perl - YAML Ain't Markup Language (tm)
22:15  libyaml-syck-perl - Fast, lightweight YAML loader and dumper
22:15  libyaml-tiny-perl - Read/Write YAML files with as little code as possible
22:15 no more YAML perl packages.
22:15
22:16 ricardo davi: I would install also the libyaml-syck-perl package
22:16 davi So, was it right and enough the installation of the libyaml-perl one?
22:16 ok, thanks
22:17 ricardo davi:  Just checked http://packages.debian.org/sou[…]/sid/libyaml-perl  - it seems to install the YAML Perl Package (and not CGI::Session::Serialize::yaml). So I guess that after installing libyaml-perl and libyaml-syck-perl, it's perfectly OK to install CGI::Session::Serialize::yaml
22:18 davi Let me try to install CGI::Session::Serialize::yaml now
22:18 ricardo OK. Good luck
22:19 davi  perl -MCPAN -e 'install CGI::Session::Serialize::yaml'
22:20 ricardo davi: Yep, that's the correct line. You do realize that you also have to paste it in the Linux shell, don't you?  ;-)
22:21 davi Oops! thanks. I forgot it  ;-)
22:21 ricardo Eheh
22:21 davi :)
22:25
22:25 Same problem with JSON
22:25 t/g4_dbfile_json................ok 60/101       (in cleanup) Dumping circular structures is not supported with JSON::Syck at /root/.cpan/build/CGI-Session-4.20/blib/lib/CGI/Session/Serialize/json.pm line 18.
22:25
22:27 Why is it referencing  CGI-Session-4.20 ?   I have an older version of CGI-Session.
22:27 ricardo davi: Really? Interesting. What's your version?
22:27 davi
22:27 MySQL fails too. So the same than before installing such packages:
22:27 t/g4_mysql......................skipped
22:27        all skipped: Couldn't establish connection with the MySQL server: Can't connect to data source '' because I can't work out what driver to use (it doesn't seem to contain a 'dbi:driver:' prefix and the DBI_DRIVER env var is not set) at t/g4_mysql.t line 44
22:28
22:28 CGI::Session::VERSION  3.95
22:29  as adviced at the Ubuntu guide: http://wiki.koha.org/doku.php?id=ubuntu_gutsy
22:29
22:29 I fear I am messing the Ubuntu installation with some of the instruction only right for OpenSuse
22:30
22:30 ricardo davi: Hmmm... Makefile.PL lists the required (minimum) version as being 4.10:
22:30 davi ricardo, Do you have a VirtualBox image with Koha 3.0 which I could download?
22:30 ricardo http://git.koha.org/cgi-bin/gi[…]lob;f=Makefile.PL
22:31 davi: I'm afraid not
22:31 davi
22:31 Oops! you are right!
22:31 Let me install it again
22:31  wget http://search.cpan.org/CPAN/au[…]ssion-4.30.tar.gz
22:31  ...
22:32 ricardo davi: I'm always right... except when I'm wrong!  ;-)
22:33 davi I use to be wrong except when I am right
22:34 ricardo LOL
22:34 davi :)
22:35 When I grow I want to be a nerd, due to nerds are the more happy people in the world
22:35 ricardo davi: You really should NOT believe the marketing brochure so much...  ;-)
22:35 davi :)
22:44 Now I have the version of 4.30 CGI::Session
22:44  I had forgot execute:  # make install
22:45
22:45 But I get yet:
22:45   mainpage.pl: DBD::mysql::st execute failed: Table 'koha.auth_types' doesn't exist at /build/kohaclone/C4/Koha.pm line 340.
22:45   ...
22:45   install.pl: DBD::mysql::st execute failed: Table 'koha.language_subtag_registry' doesn't exist at /build/kohaclone/C4/Languages.pm line 178.
22:45   ...
22:45   install.pl: Can't locate object method "generate_id" via package "CGI::Session::ID::" (perhaps you forgot to load "CGI::Session::ID::"?) at /usr/local/share/perl/5.8.8/CGI/Session.pm line 74.
22:46 gmcharlt davi: once you have the 4.30 CGI::Session, try installing the YAML serializer again
22:46 davi  perl -MCPAN -e 'install CGI::Session::Serialize::yaml'
22:46 thanks
22:51
22:51 Shit! same error again:
22:51 t/g4_dbfile_json................ok 70/101       (in cleanup) Dumping circular structures is not supported with JSON::Syck at /root/.cpan/build/CGI-Session-4.20/blib/lib/CGI/Session/Serialize/json.pm line 18.
22:51 #   Failed test 'Previously stored object loaded successfully'
22:52 gmcharlt it's pulling from CGI-Session-4.20 - try clearing your CPAN cache under /root/.cpan/build
22:53 davi What command to clean it?
22:53   rm -rf
22:53 ?
22:55 rm -rf CGI-Session-3.95
22:55 rm -rf CGI-Session-4.20
22:55 done
22:55
22:55 perl -MCPAN -e 'install CGI::Session::Serialize::yaml'  # again
22:56 ricardo davi:  Crossing fingers here, for what it's worth...   :)
22:56 davi :D
23:01
23:01 Failed again:
23:01 t/g4_dbfile_json................ok 57/101       (in cleanup) Dumping circular structures is not supported with JSON::Syck at /root/.cpan/build/CGI-Session-4.20/blib/lib/CGI/Session/Serialize/json.pm line 18.
23:01 ricardo You've got to admire JSON::Syck for its persistence!  ;-)
23:02 gmcharlt davi: try install manually
23:02 1. download and unpack http://search.cpan.org/CPAN/au[…]ize-yaml-4.21.tgz
23:02 davi The directory   /root/.cpan/build/CGI-Session-4.20/  which I removed has been regenerated.
23:03 gmcharlt 2. follow INSTALL
23:03 ricardo davi: Yep... That's weird (why would it show 4.20 instead of 4.30? puzzled...). Oh well. Follow gmcharlt's advice
23:03 davi OK, I am going to follow gmcharlt advice
23:11
23:11 lynx localhost:8080 works now! thanks
23:11 gmcharlt yay!
23:11 ricardo GREAT!!! :D
23:11 davi :)
23:12 ricardo I must admit I think I'll never understand WHY that works. Oh well..
23:15 davi I would like to learn why it was failing, but not now :)
23:16 ricardo Eheh... I understand
23:16 davi I am just a robot, a nerd
23:16 :)
23:18 ricardo davi: Well, I like to believe that nerds / geeks are curious by nature.
23:19 davi It could be a way to follow
23:20 ricardo Yep
23:29 davi Now, I want to understand the Koha source code. Is there any architecture document or similar?
23:29 Or else an 'analysis' or 'design' document?
23:30 ricardo davi: "Abandon hope all ye who enter here"  ;-)
23:31 davi :) I see
23:31 ricardo Seriously though, there's a Koha Architecture Guide for v2
23:31 davi That can be an start. Any URI?
23:32 ricardo Koha Architectural Study
23:32 http://www.kohadocs.org/Koha_architecture.html
23:33 davi thanks
23:34 ricardo davi: You're welcome  :)
23:35 davi let see what we get out from this...
23:37 ricardo davi: Sure
23:38 davi night
23:38 ricardo davi: Good night!
01:27 kados: are you still here?
01:33 kados: In the case you read this - I just updated my Installation Instructions for Koha 3 in openSUSE 11.0 in the Koha Developer Wiki - http://wiki.koha.org/doku.php?id=opensuse_11.0
01:34 However, it's already 2H30 AM (!) here in Portugal, and I will only be able to updated those instructions for the INSTALL.opensuse file tomorrow night (night here in Portugal, probably afternoon there in the US)
01:34 Take care!  :)
01:35 ( s/updated/update    gee...)
01:35 Bye

Today | Next day → | Search | Index

koha1