> > 3) Dictionary encoding bug (Rubenciño) - MH-9307 > Ruben @ TelTek > Installing the spanish dictionary. > Purified it by removing illegal characters. > Saw in the logs that certain characters didn't exist or couldn't > be loaded in the database > Took a look at the dictionary scanner to see and the language must > be specified at loading time, right now set only to english. > One line change probably. > Switch to UTF-8 as default? > > Tobias Wunden @ Entwine > Make sure that the dictionaries are currently UTF-8 if we are > switching to this format. > Also add a warning that the dictionary needs to be UTF-8. >
Just to clarify: * The "purified" dictionary (where we eliminated most illegal words) was committed some time ago. I just hadn't noticed the loading errors until now. * What it needs to be specified is the character encoding when you read a file using an InputStreamReader. Otherwise, it may or may not use UTF-8 by default,depending on the machine configuration. * We are not switching to UTF-8. I believe it is an implicit convention (read: "I don't know if there was an official decision about it"), that all the files in MH are UTF8-encoded. However it's good to make sure the dictionary files stick to this convention, and that people building their own dictionaries are aware of that. Best regards Rubén Pérez Vázquez <http://www.teltek.es> www.teltek.es 2013/2/12 McKenzie, Adam <adam.mcken...@usask.ca> > Hi Greg, > > I took the minutes for the meeting and have pasted them below: > > 1) 1.4 QA (Tobias) > Tobias Wunden @ Entwine > Entwine has started doing QA on 1.4 > James from Manchester and Karen from Harvard discussion on list -> > losing half the metadata > Metadata form doesn't keep all of the data from before, strips out > anything it doesn't recognize. > QA is not where it should be, finding too many new issues. > > Rubencino @ TelTek > For 1.4 it is difficult, being a long time for this release. Lot > of stuff to review > For the next release we could try to get more adopters involved. > Those without commit rights could help by testing. > We might not be detecting all of the possible failures. > > James Perrin @ Manchester > Multiple dublin core packages in the media package was not > something expected. > We don't know what the expected behavior should be. > Just accepted that that was how it was meant to be. > > Tobias Wunden @ Entwine > Do we need more documentation? > > Jaime Gago @ Entwine > If we have the documentation and someone is QAing as not a > developer at least they can compare the documentation with how the > application is actually behaving and report if there is an issue. > Every release we need new QA tests to cover the new features. > The QA test table is good for knowing whether something in > particular has happened. > Maybe we need to create an open space to easily put QA results? > Something like a wiki page. > > James Perrin @ Manchester > A lot of the basic design for matterhorn isn't documented and > exists in people's heads. > Things like use cases are there but not how the components are > tied together and how the bigger pictures fits together. > Where some of the basic information is. What information is > attached to a mediapackage? What happens to it? How is it used, processed & > passed around. > Developing some paper documentation on what they have discovered > so far. > > Jaime Gago @ Entwine > Tried to work on this documentation at the conference. Working > with the entwine developers tried to workout what is actually happening but > in a couple of hours only got to scheduling. > We could still put something together to understand how to QA the > new features to try to make QAing easier. > > Tobias Schiebeck @ Manchester > Was only able to discover the issues they are having with the RSS > feeds by going to production and actually trying to use the system. > > Tobias Wunden @ Entwine > I think it is obvious that test cases are missing and need details. > That we need document some of these technical details. > UCB & Manchester are running 1.4 instances. > There are other institutions running their own QA. Could we make a > more formal process or some mechanism to get people to funnel information > back about how their QA process is going to the community at large. > If we aren't sharing our findings then the issue is only getting > solved at the instituional level and not for the whole community. > > Stuart @ Manchester > Something more formal would help > > Tobias Wunden @ Entwine > It would be good if more institutions run the local QA process > through the matterhorn ticket system to help publisize some of the results > from the QA > > Jaime Gago @ Entwine > Not knowing whether something is functioning properly is a large > part of the problem. It needs to be clear about what exactly is expected. > It is not easy as an adopter to know what version of everything is > running on a particular instance for example what database with what > version. > > > 2) Metadata bug (Tobias) > Tobias Wunden @ Entwine > Ticket created in jira. Talked about the issue above when talking > about QA. No one has started working on it yet. > > 3) Dictionary encoding bug (Rubenciño) - MH-9307 > Ruben @ TelTek > Installing the spanish dictionary. > Purified it by removing illegal characters. > Saw in the logs that certain characters didn't exist or couldn't > be loaded in the database > Took a look at the dictionary scanner to see and the language must > be specified at loading time, right now set only to english. > One line change probably. > Switch to UTF-8 as default? > > Tobias Wunden @ Entwine > Make sure that the dictionaries are currently UTF-8 if we are > switching to this format. > Also add a warning that the dictionary needs to be UTF-8. > > 4) CAS Regression due to MH-9307 fix > John King @ UCB > Changing empty strings to null strings in MH-9307 breaks CAS since > the passwords are empty in that authentication. > Needs to write it up with a proposed fix. > 1) The central project fix for > http://opencast.jira.com/browse/MH-9307, svn 13542 change in > UserAndRoleDirectoryServiceImpl about line 190, broke CAS because this > line was removed - String password = user.getPassword() == null ? "" : > user.getPassword(); This breaks the instantiation of > org.springframework.security.core.userdetails.User because CAS users have > no passwords. > Secondary issues that will cause breakage > 2) Some but not all of the Spring Security libraries in > the project had their versions bumped from 3.0.2.RELEASE to 3.1.0.RELEASE. > 3) security_sample_cas_xml was not updated to reflect the > change in Spring Security version. > 2 & 3 caused the error "user _cas_stateful_ unknown" > Upping the version and the quick fix seems to have solved the > issue. > > Tobias Wunden @ Entwine > Shibboleth & CAS - User logs in through the admin ui, starts a > workflow and then it is processed and published on different server. > We don't have access to the user anymore causing issues. > > John King @ UCB > Using LDAP & CAS. > CAS is enabled on everything. Any machine would have access to > that user in LDAP. > Internal authorization with matterhorn system account, how does > that work in a CAS enabled environment? > Able to get CAS enabled with the fixes. > > Adam McKenzie @ University of Saskatchewan > The matterhorn system account goes through the authenticaion with > a different user provider. > > John King @ UCB > svn 13830 - for consistency, making sure all spring-security > libraries are updated to 3.1.0.RELEASE, changing security_sample_cas_xml > and mh_default_org_cas_xml security config files to reflect the update as > documented here - > http://static.springsource.org/spring-security/site/docs/3.1.x/apidocs/index.html?org/springframework/security/cas/web/CasAuthenticationFilter.htmland > here - > http://static.springsource.org/spring-security/site/docs/3.1.x/reference/cas.html. > Adding one necessary OSGI import to matterhorn-security-cas/pom.xml and > fixing one library name inconsistency in matterhorn-security-cas/pom.xml. > Replacing deprecated class in UserAndRoleDirectoryServiceImpl with > SimpleGrantedAuthority > Going to open two seperate jiras. > > 5) Ongoing discussions (Tobias) > - devlopment proposal > Tobias Wunden @ Entwine > Will be writing down what was discussed at the > conference and then will be sending out an email. > > - capture agent codebase > Tobias Wunden @ Entwine > A lot has been said. Olaf has summed up a lot of > the discussion thus far. > Please contribute your voice on list about how you > feel about the reference capture agent. > - rpm version of matterhorn > Jody Fanto @ Harvard > What is the state of the rpm for the most recent > release candidate? > > Jaime Gago @ Entwine > Entwine tries to have an rpm at the latest version. > There is a github repo to make a matterhorn rpm. > Would prefer this as an installation method. > Blog post describing how to roll your own: > http://entwinemedia.com/2012/rpm-based-matterhorn-deployment/#more-1448 > One rpm that can be configured to install any type > of host e.g. admin, worker or engage node all same rpm, different config. > > 6) Number of 1.4 Tickets > Tobias Wunden @ Entwine > Still less than 100 tickets open in 1.4, please try to > look at your tickets so that we can decrease this number further! > > Moved to be discussed Next Week: > 7) Database Schema updates (strategy for maintenance updates of production > systems aka db patching, versioning?) (Jaime Gago) > 8) 1.4 Tickets (All) > > > Cheers, > Adam > > Adam McKenzie > Programmer Analyst > Applications, Information and Communications Technology > Ph: (306) 966-8241 > > > > ________________________________________ > From: matterhorn-boun...@opencastproject.org [ > matterhorn-boun...@opencastproject.org] on behalf of Greg Logan [ > greg.lo...@usask.ca] > Sent: Monday, February 11, 2013 5:53 PM > To: Opencast Matterhorn > Subject: [Opencast Matterhorn] Dev Meeting 2013-02-12 > > Hi folks, > > I'm on holidays this week, so I probably won't be at the meeting. That > being said, I would like to draw everyone's attention to the bugs at > > https://opencast.jira.com/secure/IssueNavigator.jspa?resolution=-1&pid=10010&fixfor=10190&priority=6&reset=true&show=View+%26gt%3B%26gt%3B > (our remaining 1.4 release blockers). I would like to hear status > updates on all of the open issues there, even if it's 'no progress yet'. > I'm sure there will also be lots of discussion around the future of the > CA as well, but I would ask that someone take notes and forward them to > the list. I do not want us discussing things that aren't shard with the > general userbase! > > Thanks, > G > > _______________________________________________ > Matterhorn mailing list > Matterhorn@opencastproject.org > http://lists.opencastproject.org/mailman/listinfo/matterhorn > > > To unsubscribe please email > matterhorn-unsubscr...@opencastproject.org > _______________________________________________ >
_______________________________________________ Matterhorn mailing list Matterhorn@opencastproject.org http://lists.opencastproject.org/mailman/listinfo/matterhorn To unsubscribe please email matterhorn-unsubscr...@opencastproject.org _______________________________________________
