Log of the #fcrepo channel on chat.freenode.net

Using timezone: Eastern Standard Time
* ajs6f1 joins01:37
* ajs6f joins02:28
* ajs6f1 leaves02:30
* dwilcox joins07:47
* mhwood joins08:12
* mikeAtUVa joins08:34
* ksclarke joins08:36
* dhlamb joins08:48
* ajs6f2 joins08:53
* cmmills joins08:58
* awoods joins09:01
* whikloj joins09:02
* awead joins
* osmandin joins09:10
* acoburn joins09:27
<whikloj>where is the best place to put a unit test for an Exception? There doesn't seem to be any, and I'm not sure fcrepo-kernel is correct?09:35
<awoods>whikloj: in general (with Exception being no exception), unit tests should be in the same project as the functional code, in a parallel source tree/package, but under "test" instead of "main".09:36
<whikloj>awoods: fair enough, having trouble as the function I am testing builds the Link fromUri and that seems to require a piece of glassfish that is not loaded down at the kernel level.09:50
<awoods>whikloj: feel free to push your commit and I can take a look.09:51
<whikloj>awoods: okay, let me clean up the IT as it works.
<awoods>whikloj: it seems like your test should not require more libraries than your functional code (except the test libraries like mockito).09:52
<whikloj>awoods: yeah, I am not clear on why it fails except that the end result is a ClassNotFoundException: org.glassfish.jersey.internal.RuntimeDelegateImpl
* mhwood leaves
<awoods>whikloj: let me know when you have something for me to look at.09:53
<ajs6f2>whikloj: Does your class use some kind of injection?09:58
<whikloj>ajs6f2: Mocked UriInfo
<ajs6f2>whikloj: I think that's Jersey trying to do your injection.09:59
<whikloj>ajs6f2: error seems to occur on this line10:00
return Link.fromUri(constraintURI).rel(CONSTRAINED_BY.getURI()).build();
ajs6f2: for context https://github.com/fcrepo4/fcrepo4/pull/808/files#diff-c14ecfb343036db02d3070d59bd47b0eR7210:01
<ajs6f2>whikloj: is that the first time you use the UriBuilder?10:02
<whikloj>ajs6f: you mean the builder in the Link.fromUri(), yes.10:03
<ajs6f2>Why is this exception in the kernel if it requires JAXRS types?
whikloj: lazy initialization means that that is where the UriBuilder will get pulled in.10:05
Or fail to.10:06
<whikloj>ajs6f2: I can remove the JAXRS types and place them in each of the ExceptionMappers, this was cleaner. But now seems less so.
<awoods>ajs6f2: good point regarding jaxrs. The exception itself is fine, it is the helper method (buildConstraintLink) which is the problem. Maybe it should be pulled up into an fcrepo-http-api helper class.
<ajs6f2>whikloj/awoods: i think both approaches could work. How about an ExceptionMapper superclass to stay DRY?10:08
Which would live in http-commons and feature that helper method.10:09
<whikloj>awoods/ajs6f2: I had it in the exception class to save passing the exception in, I could put it in the ConstraintViolationExceptionMapper class and pass 2 args in
awoods/ajs6f2: Or just have it return the String of the URI and build the link in ExceptionMapper superclass10:10
<awoods>whikloj: would working with a String of the URI require some ugly parsing?10:13
<whikloj>awoods: no, I can pass the String back from the Exception to the ExceptionMapper and have the ExceptionMapper build the Link. trying it out now.10:14
* osmandin leaves10:32
<f4jenkins>Project fcrepo4-T2 build #271: UNSTABLE in 4 min 30 sec: http://jenkins.fcrepo.org/job/fcrepo4-T2/271/10:52
* awoods: Add check for problematic SPARQL Update statement
* Aaron Coburn: removed maven.fcrepo.org repository
* awead leaves12:00
* awead joins
* awead leaves12:01
<whikloj>ajs6f2: when you said make an ExceptionMapper superclass, is this a holder for utility functions or an actual superclass?12:19
* github-ff joins12:27
[fcrepo4] awoods created fcrepo-1581 (+1 new commit): http://git.io/vLpWc
fcrepo4/fcrepo-1581 6333db3 Andrew Woods: Updates ModeShape -> 4.3.0, and ISPN -> 7.2.0...
* github-ff leaves
<whikloj>acoburn: ping12:37
<acoburn>whikloj: pong12:38
<whikloj>acoburn: If I change an old (Fedora 3.7) topic to a queue, will I lose any messages in the topic? or would they be "moved" to the queue?
<acoburn>whikloj: are you changing the activemq.xml configuration on the fedora 3.7 repo?12:39
<whikloj>acoburn: sort of, except in the fedora.fcfg.12:40
acoburn: here is the gist12:41
acoburn: we are running a topic right now, and want to move to a queue to add more GSearch subscribers12:42
<acoburn>whikloj: there should be no problem with that
whikloj: with queues, you aren't going to lose messages12:43
<whikloj>acoburn: but will changing the type from topic to queue mean items currently in the topic will be lost or never processed
<acoburn>whikloj: are there a lot of messages waiting to be processed?
<whikloj>acoburn: not sure as it is hard to browse a topic :)12:44
acoburn: here is the instructions we are thinking of implementing https://goo.gl/NHEcSH
<acoburn>whikloj: in order to make the change, you'll need to restart fedora, which means restarting the broker (unless you've externalized it)
<whikloj>acoburn: yes, it is still internal12:45
<acoburn>whikloj: are you seeing a lot of activity with GSearch at the moment?12:47
<whikloj>acoburn: yes, we still seem to have items being processed. I did a reindex of about 100,000 objects.
<acoburn>whikloj: I don't actually know how GSearch would behave during that sort of change12:50
<whikloj>acoburn: ok, thanks. We will have to decide if we risk it or not.12:51
<acoburn>whikloj: once you move it to a queue, you'll be able to handle these changes much more easily
whikloj: though it's even better if you externalize the message broker, as it doesn't eat into fedora's resources12:52
<whikloj>acoburn: agreed, but I'm taking baby steps here :)
<acoburn>whikloj: of course. one thing at a time!12:53
<awoods>mikeAtUVa: ruebot noticed that your code signing key has expired: https://wiki.duraspace.org/display/FCREPO/Fedora+Contributors12:59
* travis-ci joins
fcrepo4/fcrepo4#3726 (fcrepo-1581 - 6333db3 : Andrew Woods): The build passed.
Change view : https://github.com/fcrepo4/fcrepo4/commit/6333db3e546c
Build details : https://travis-ci.org/fcrepo4/fcrepo4/builds/68025332
* travis-ci leaves
<mikeAtUVa>awoods: should I make a new one and add the signature to the wiki? What expiration period are others using?13:00
<awoods>mikeAtUva: my expiration is unlimited13:01
mikeAtUVa: If you could use your existing key, that would be good... but if it is easier, feel free to add a new one.13:02
* awead joins13:10
* acoburn leaves13:43
* jmignault joins14:06
* acoburn joins14:36
* awead leaves14:54
* awead joins14:58
* dwilcox leaves14:59
* dwilcox joins15:02
* whikloj leaves15:05
* jmignault leaves15:15
<dhlamb>acoburn: think i fixed your osgi gremlin15:48
<dhlamb>acoburn: still testing on my end, but fcrepo-camel seems to be working in osgi again after bumping httpclient and httpcore up by a revision
acoburn: and the tests pass15:50
dhlamb: when i feel comfortable, i'll issue a pull request
<acoburn>dhlamb: that's great. thanks15:51
<dhlamb>acoburn: awoods: this passes my sniff test. is it ok to issue this PR without a ticket?16:01
<acoburn>dhlamb: can you create a Jira ticket for this?16:02
<dhlamb>acoburn: sure
* dwilcox leaves16:03
* github-ff joins16:11
[fcrepo-camel] daniel-dgi opened pull request #76: For https://jira.duraspace.org/browse/FCREPO-1598 (master...fcrepo-1598) http://git.io/vLjIt
* github-ff leaves
<dhlamb>acoburn: https://jira.duraspace.org/browse/FCREPO-159816:12
acoburn: and https://github.com/fcrepo4/fcrepo-camel/pull/76
acoburn: and i'm out for now. we can talk later if there's any issues16:14
ttyl folks
* dhlamb leaves16:19
* mikeAtUVa leaves17:10
* awead leaves17:12
* acoburn leaves17:34
* cmmills leaves17:39
* ksclarke leaves18:25
* awead joins20:28
* dhlamb joins21:06
* ksclarke joins21:39
* ksclarke leaves21:53
* ksclarke joins22:09
* awead leaves22:29
* github-ff joins22:32
[ontology] awoods deleted tech-metadata-cleanup at 8113976: http://git.io/vtexq
* github-ff leaves
* dhlamb leaves23:11
* dhlamb joins23:12
* dhlamb leaves23:33
* ksclarke leaves00:29
* ksclarke joins00:30
* ksclarke leaves01:01

Generated by Sualtam