August 2016 - Koha-devel - lists.koha-community.org

Koha 3.20.14 released
by Chris Cormack 23 Aug '16

23 Aug '16

The Koha community is happy to announce the release of Koha 3.20.14. This is a security release and users of 3.20.x are encouraged to upgrade. Please read the full release notes at https://koha-community.org/koha-3-20-14-released/ Chris

1 0

Call for News: August 2016 Newsletter
by kohanews 18 Aug '16

18 Aug '16

Fellow Koha users ~ I'm collecting news for the August newsletter. Send anything noteworthy to: k o h a news AT gmail dot com News criteria: --------------------------- ** For events **: * - Please include dates for past events. If I can't find dates I may not add it. * - Announcements for future events with dates T.B.A. are fine ...Eg., Kohacon * - For past events -- **** one month back is the cut-off ***. * News items can be of any length. * Images are fine * Anything and everything Koha. * Submit by the 26th of the month. If you are working on an interesting project or development related to Koha, please let me know and I'll include it in the development section. -- Chad Roseburg Editor, Koha Community Newsletter

1 0

Search by itemnumber
by Paul A 17 Aug '16

17 Aug '16

I must be missing something very simple -- maybe related back to bug#168 ;={ How to search (staff/admin) for [items.]itemnumber? Scenario: staff logs give adds, edits etc of itemnumbers, but without the biblionumber. Thanks and best regards -- Paul

2 2

Re: [Koha-devel] Koha-devel Digest, Vol 129, Issue 18
by Benjamin Rokseth 16 Aug '16

16 Aug '16

________________________________________ Fra: koha-devel-bounces(a)lists.koha-community.org <koha-devel-bounces(a)lists.koha-community.org> på vegne av koha-devel-request(a)lists.koha-community.org <koha-devel-request(a)lists.koha-community.org> Sendt: 16. august 2016 12:00 Til: koha-devel(a)lists.koha-community.org Emne: Koha-devel Digest, Vol 129, Issue 18 Send Koha-devel mailing list submissions to koha-devel(a)lists.koha-community.org To subscribe or unsubscribe via the World Wide Web, visit http://lists.koha-community.org/cgi-bin/mailman/listinfo/koha-devel or, via email, send a message with subject or body 'help' to koha-devel-request(a)lists.koha-community.org You can reach the person managing the list at koha-devel-owner(a)lists.koha-community.org When replying, please edit your Subject line so it is more specific than "Re: Contents of Koha-devel digest..." Today's Topics: 1. Re: What's on in Koha devel #3 (RESTful API) (Tomas Cohen Arazi) 2. Re: What's on in Koha devel #3 (RESTful API) (Katrin Fischer) ---------------------------------------------------------------------- Message: 1 Date: Mon, 15 Aug 2016 12:54:06 +0000 From: Tomas Cohen Arazi <tomascohen(a)gmail.com> To: Marcel de Rooy <M.de.Rooy(a)rijksmuseum.nl>, "koha-devel(a)lists.koha-community.org" <koha-devel(a)lists.koha-community.org> Subject: Re: [Koha-devel] What's on in Koha devel #3 (RESTful API) Message-ID: <CABZfb=WXTfeiMjO7zjccDTjSK7N=-E-MnQC3SPKNvR0VtZUttw(a)mail.gmail.com> Content-Type: text/plain; charset="utf-8" El lun., 15 ago. 2016 a las 4:28, Marcel de Rooy (<M.de.Rooy(a)rijksmuseum.nl>) escribi?: > > = RESTful API = > > I also would like to know who is willing to maintain this code? For > instance on bug 16330 (REST API: add routes to add, update and delete > > patrons) the delete route is calling > > # check if loans, reservations, debarrment, etc. before deletion! > > my $res = $patron->delete > > So no check is done from the API. > > But if bug 16907 (Move DelMember and HandleDelBorrower to > > Koha::Patron) is pushed, $patron->delete will take care of the holds, > etc. My question is simple: are we expecting from devs, signoffers and > > QAers to update and maintain these routes? > > Or a dedicated person will update them before each release? > > Should our scripts go via REST some day and should all business logic > remain in Objects? Etc. etc. > Exactly! What Jonathan noticed, is that no one realized there was a bug on the REST api, this should be reported, and fixed. The more we use the API the more bugs we will find, and the more we will enhance it. I think Marseille will be a good place to work on this. Regards -- Tom?s Cohen Arazi Theke Solutions (https://theke.io <http://theke.io/>) ? +54 9351 3513384 GPG: B2F3C15F

1 0

Re: [Koha-devel] Git bz AttributeError: 'NoneType' object has no attribute 'group'
by Holger Meissner 15 Aug '16

15 Aug '16

Hi Katrin, problem solved, thank you! > Hi Holger, > I am not sure this is the reason, but did you set: > config --global bz-tracker.bugs.koha-community.org.https true > I think that was the last change we made to the configuration of git bz. > Hope this helps, > Katrin > Am 11.08.2016 um 13:10 schrieb Holger Meissner: >> Hello koha devs, >> >> I was just about to upload a small patch to bugzilla. git bz attach >> 14434 HEAD But git bz says >> AttributeError: 'NoneType' object has no attribute 'group' >> >> Any ideas how to fix it? I successfully fetched and rebased the fishsoup branch, but that didn't help. >> >> Kind regards >> Holger

1 0

Proposal to join the QA team for packaging related bugs
by Mirko Tietgen 15 Aug '16

15 Aug '16

Hello everyone, part of the idea of nightly packages is to constantly update the Debian related files that deal with depenencies. I would like to find and solve possible dependency problems early and minimize the delay between Koha release and related package release. The process to build nightlies is getting more stable now. I have tested automatic patches to the debian/control file -- that is the file that contains the package dependencies -- and I think it will be a helpful tool. I have created bugs labelled "Automatic debian/control updates" for master, stable and unstable, and added Kyle, Frederic and Julian respectively. https://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=17084 https://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=17108 https://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=17111 The bugs get automatic updates after each nightly build process, as long as they lead to changes in debian/control. To avoid spamming the bugs I have changed that now to "send patch only if bug status is 'Assigned'" (untested :P) There is not really anything to test here, besides a manual sanity check. But it will consume resources to go through the QA process every single time there is a change. For that reason I propose I join the QA team for packaging matters, do a signoff on the automatic patches (as long as they are ok), label them "Passed QA" and send them over to Kyle, Frederic and Julian so they can be pushed early. Any objections? Comments? Cheers, Mirko -- Mirko Tietgen mirko(a)abunchofthings.net http://koha.abunchofthings.net http://meinkoha.de

7 6

Git bz AttributeError: 'NoneType' object has no attribute 'group'
by Holger Meissner 14 Aug '16

14 Aug '16

Hello koha devs, I was just about to upload a small patch to bugzilla. git bz attach 14434 HEAD But git bz says AttributeError: 'NoneType' object has no attribute 'group' Any ideas how to fix it? I successfully fetched and rebased the fishsoup branch, but that didn't help. Kind regards Holger

2 1

What's on in Koha devel #3
by Jonathan Druart 13 Aug '16

13 Aug '16

Hi devs, Lot of things have been going on during the last month and lot of patches are in the queue :) = Security issues = Plenty of security issues have been raised and fixed during the last month (and others are still coming). They are related to CSRF and XSS protections. See bug 14568 ([OMNIBUS] XSS in Staff Client) and bug 17096 ([OMIBUS] CSRF protections) for the entry points. Most of them are in the security project and most of you won't have read access, but if you are interested in testing them you can ask someone on #koha to have access. = Caching system = If you are familiar with plack and memcache, please have a look at the POC on bug 16579 (Use separate memcached namespace for caching koha-conf.xml). No need to test it if you don't want, you can just have a glance at it. = RESTful API = A lot of patches related to the RESTful API are in the SO and QA queues. The actors (authors/signoffers) are almost always the same. I'd like to know if someone (from the QA team or not) fancy to technically review these patches. I did not follow closely the discussions and the progress of the RESTful API and I think it could be good to have someone who's implied since the beginning in the QA process to have a look at how the architecture is done. I also would like to know who is willing to maintain this code? For instance on bug 16330 (REST API: add routes to add, update and delete patrons) the delete route is calling # check if loans, reservations, debarrment, etc. before deletion! my $res = $patron->delete So no check is done from the API. But if bug 16907 (Move DelMember and HandleDelBorrower to Koha::Patron) is pushed, $patron->delete will take care of the holds, etc. My question is simple: are we expecting from devs, signoffers and QAers to update and maintain these routes? Or a dedicated person will update them before each release? = Focus on blockers = When criticals or blockers are opened, usually it takes between 3 or 4 days to get a patch and see it pushed. Which is quite fast :) But most of the time, several bug reports are opened meanwhile and marked as duplicate later. So I think we should focus as much as possible on signing off these patches and 1/ ask on #koha if you find a software error and 2/ Search on the bug tracker in the most recent bugs. For instance bug 16960 (critical) has a patch waiting for a signoff for 3 days now... = Refactoring = Lot of things to say here :) == Accounts rewrite == I'd like to start QAing some parts of the accounts rewrite. For instance bug 15899 (Remove the use of recordpayment) is a good first candidate. But to do so I need all the dependencies to be signed off. == Koha::Patrons == Lot of things have been moved last month, the entry point is bug 16846 (Move patron related code to Koha::Patron) if you want to help or just have a look. == C4::Branch is (almost) dead == Bug 15758 (Move the C4::Branch related code to Koha::Libraries - part 4) removes completely C4::Branch. It is a big patch and has been widely tested. We need a quick but deep QA here, it's a nightmare to rebase! == Koha::AuthorisedValues == (I have c/p the following on bug 15799 - Move authorised values related code into Koha::AuthorisedValues - part 2, better to comment on the bug report) We have already started to move the authorised values related code to the Koha namespace, but some are still missing. To continue and move C4::Koha::*AuthorisedValues* subroutines to Koha::AuthorisedValues, we will need to join on marc_subfield_structure(.authorised_value which is actually an authorised value category). To do so we need to define a relationship (fk) between the two tables. But we have "special AV" like Asort1 (and friends) , LOST, DAMAGED, etc.) Which could not be in the authorised_values table [yet]. And 3 other "special special AV) like itemtypes, branches and cn_source. Others (undetermined so far) are linked to a marc_subfield_structure from installer/**/marcflavour/**/*framework*.sql, like STACK. My idea would be to move all the AV categories to a new table authorised_value_categories. It will contain only 1 column category_name refered by 1. authorised_values.category, 2. additional_fields.authorised_value_category (others authorised_value field should be renamed like that when they refer AV categories) and 3. iitems_search_fields.authorised_value_category. And, maybe later, auth_subfield_structure, auth_tag_structure, borrower_attribute_types, marc_tag_structure (some are varchar(10) et 20 when they should be 32!) We could imagine more attributes to this table, but don't ask me to add a authorised_value_categories.id as pk please :) We will have to use temporary tables to deal with the DB constraints, to add more fun. Finally the .sql will have to be updated to reflect these changes. I have just started but before to enter in this tunnel I'd like to know if there are a couple of volunteers to test? :) = Ubuntu 16.04/MySQL 5.7 support = Note that we are still waiting for bug 16554 to support Ubuntu 16.04 and MySQL 5.7 out of the box. A dev meeting is planned on August 24th at 7 UTC (https://wiki.koha-community.org/wiki/Development_IRC_meeting_24_August_2016) Please have a look at the agenda and the different topics we will discuss. Don't forget to register for the Hackfest in Marseille in October (http://lists.koha-community.org/pipermail/koha-devel/2016-June/042790.html) Cheers, Jonathan

2 1

Please help clean up bugs on unmaintained branches
by Owen Leonard 12 Aug '16

12 Aug '16

Recently I've been going through the list of bugs reports on unmaintained 3.x versions of Koha. I've closed bugs which I could not verify and updated to 'master' those which I could reproduce. The list is down to 45 bugs which I had questions about, didn't understand, or didn't know how to test. I would love to get help getting that list down to zero. Please take a look if you can. https://bugs.koha-community.org/bugzilla3/buglist.cgi?cmdtype=runnamed&name… Thanks, Owen -- Web Developer Athens County Public Libraries http://www.myacpl.org

1 1

Koha::Object[s] improvements - call for discussions
by Jonathan Druart 11 Aug '16

11 Aug '16

Hi devs, At the KohaCon16 some of us have briefly talked about the limitations of Koha::Object[s] we encounter when implementing new modules. Koha::Object[s] has been pushed to master less than 18 months ago and we already have more than 30 classes using it. It permits to uniformise our way to code, use powerful objects and is very easy for refactoring. I have been moving code from C4 to the Koha namespace and it's very handy to just have to write a concise module and automatically have the basic methods (CRUD and more). However it seems that we need to clearly define what we need in a near future and adapt it quickly if it's needed. = Rooms for Improvement = There are 3 main rooms for improvement: - DBIx::Class basic methods - DBIx::Class relationship accesses - Koha::Objects vs DBIx::Class objects == DBIx::Class basic methods == If you want to access a method provided by DBIx::Class, like ->store, ->delete, ->is_changed, we need to create a silly method which just calls the same method on the DBIx::Class object. For instance, let's take Koha::Objects->is_changed sub is_changed { my ( $self, @columns ) = @_; return $self->_result()->is_changed(@columns); } Possible solution: It would be possible to modify our AUTOLOAD method and check if it's a DBIX::Class method. If it is a valid one, let's call it, otherwise carp something. == DBIx::Class relationships == >From a Koha::Objects, the relationships provided by DBIx::Class will not be automatically generated. For instance, from a virtual shelf, you will not have access to the items on the shelf. You will have to create a method which will call the DBIx::Class relevant relationship. sub get_contents { my ( $self ) = @_; my $contents = $self->{_result}->virtualshelfcontents; return $contents; } This piece of code adds a get_contents method to Koha::Virtualshelf and call the virtualshelfcontents relationship created by DBIx::Class. Note that this is finally good here, it is better to call have a get_contents method than virtualshelfcontents. But we may prefer to redefine it in the schema file (Koha::Schema::Result::Virtualshelve). Possible solution: Same as previously, it would be possible to modify our AUTOLOAD method and check if it's a valid relationship. == Koha::Objects vs DBIX::Class objects == We'd like to always manipulate the same kind of objects, to avoid confusion when reading the code. At the moment we already manipulate 2 different entities: hashref from C4 and objects from Koha::Object. It's confusing and not so easy to manage in the code. To add flexibility we have added a Koha::Object->unblessed method to transform the object to an hashref, it's not ideal but it ease the transition. Anyway, from a Koha::Object if you want to retrieve data linked to this object, you would like to get a Koha::Object object base, not a DBIx::Class object base. If we continue with the previous example, we will need to modify get_contents like this: sub get_contents { my ( $self ) = @_; my $rs = $self->{_result}->virtualshelfcontents; my $contents = Koha::Virtualshelfcontents->_new_from_dbic( $rs ); return $contents; } Basically ->new_from_dbic will just create an empty Koha::Virtualshelfcontents object and set the internal resultset. No query or extra processing should be done here. I do not know if the changes suggested (and already implemented in some areas and patches in the queue) is consistent. It's not very elegant to read, but it looks safe on the performance side, despite the encapsulation problem we could see at first glance. = Concerns = == API == Some people raised concerns about the new API to lean. But finally we stick on DBIx::Class method names, so it should not be to hard to learn for new devs. And... the modules in the Koha namespace are much more easier to read/understand than the legacy code in C4. == ResultSet == It seems that writing the following is not obvious for everybody: # Create and store my $biblio = Koha::Biblio->new( { title => "a title", [other => $data] } ); $biblio->store; # Retrieve data my $biblio = Koha::Biblios->find( $biblionumber ); my $biblio = Koha::Biblios->search( { title => "a title" }); # Access relationship my $items = $biblio->items; warn ref( $items ); # Koha::Items warn ref( $items->first ); # Koha::Item This looks very easy to read for me, feel free to suggest something else. = A complete example = At the moment, the more complete example in master (and 16.05) is the Koha::Virtualshelf and Koha::Virtualshelves modules. It includes several search methods, overwrite of the store method and some complex methods as well as complete test coverage (t/db_dependent/Virtualshelves.t) = Alternatives = If you think you can provide an alternative or an idea to improve/replace our Koha::Object[s] implementation, do not hesitate, I would be glad to review or even implement it! If you want to discuss one of these point, add a new one, or have a specific question, please answer this email. We could talk about that during the next dev meeting as well. Cheers, Jonathan

4 5