Log of the #duraspace-ff channel on chat.freenode.net

Using timezone: Eastern Standard Time
<pivotal-bot_>Edwin Shin added comment: "As a note, the key thing about the version numbering is to make it fit the Maven conventions such that subseq..." https://www.pivotaltracker.com/story/show/5156988107:10
Edwin Shin added comment: "Also see: http://docs.codehaus.org/display/MAVEN/Versioning" https://www.pivotaltracker.com/story/show/51569881
Edwin Shin edited "Tag a 4.0.0-alpha-1 release of fcrepo4 and publish the artifacts" https://www.pivotaltracker.com/story/show/5156988107:13
Edwin Shin accepted "IllegalArgumentException: Invalid format: "2013-06-30T03:08:55Z" is malformed at "Z"" https://www.pivotaltracker.com/story/show/5259170507:14
Edwin Shin accepted "Add TX information to the graph response, or introspect the URL for a TX identifier and add the appropriate actions..." https://www.pivotaltracker.com/story/show/52316287
Edwin Shin accepted "Add reflection util to Testhelper for setting fields which lack a setter method" https://www.pivotaltracker.com/story/show/52055161
Edwin Shin accepted "Configure jenkins to build and publish site docs" https://www.pivotaltracker.com/story/show/5260085107:15
* travis-ci joins07:27
[travis-ci] futures/fcrepo4#831 (master - c9b3365 : Edwin Shin): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/b3d4ea75a3be...c9b336570264
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/8645769
* travis-ci leaves
* kaarefc joins07:35
* kaarefc leaves07:50
* kaarefc joins
* leenata joins08:51
* ksclarke joins09:20
<leenata>hey eddies09:28
* kaarefc leaves
<leenata>the junit tests for FedoraResources is kind of stuck in some testcases09:30
there are static methods in there which are difficult to mock and test
do you see this as a problem or you know any way around this?09:31
<eddies>leenata: can you pass me a link (on github) to show me what you're talking about?09:48
* pivotal-bot__ joins10:02
* pivotal-bot_ leaves10:03
* github-ff joins10:19
[fcrepo4] leenata created resource-unit-test (+1 new commit): http://git.io/Rx4y2Q
fcrepo4/resource-unit-test 4e7ef59 Leenata Bankhele: JUnit tests for FedoraResource UpdatePropertiesDataset
* github-ff leaves
<barmintor>leenata: There's some example code for mocking static methods in the unit tests for the kernel10:22
leenata: but if the static methods are defined on the class you're trying to unit test, you are in a pinch10:23
that's really a design principle we should articulate
<leenata>hey yes, I did refer some existing mocking of static methods in the FedoraResourceTest.java10:32
<barmintor>leenata: I'm on a roll for making outdated contributions to conversations this week.10:34
<leenata>but that didn't help. Also the method isn't from the class I am testing. The one that I am talking about is the DatasetFactory.create10:36
<eddies>leenata, barmintor: on a general note, i was saying (last week?) that we should treat mocks as a sign of code smell. it's one thing to need mocks for code we don't own, but if you find yourself needing to mock stuff in org.fcrepo, that's should be a sign for us to re-examine that code to support better testability
<barmintor>eddies: I think I disagree
<eddies>i like to entertain a variety of opinions on this project, however wrongheader ;-)10:37
<barmintor>It depends on the scope of the test, but I think of unit tests as just exercising a single class, so other objects would necessarily be mocked10:38
* barmintor shrugs
excessive mocking is definitely a smell, though
* unai_ joins
<eddies>you disagree as a general rule? i don't deny there will always be exceptions, but as a rule of thumb, i don't think of mocks as a good thing
<cbeer>i think i've lost the thread of what's being tested here10:40
<eddies>yeah. i think in FedoraResource, there's enough mocking going on that it's difficult to follow, let alone be sure what's supposed to be exercised is being appropriately exercised10:41
<cbeer>leenata: right, sorry. i mean why there's this logic to support a case where
i think it's the case when dataset.getContext() is null..
<leenata>its this static method that I was stuck at, and the reason for the discussion
<cbeer>i don't even see how this code is supposed to behave correctly either:
com.hp.hpl.jena.sparql.util.Context context = dataset.getContext();
if ( context == null ) {
context = new com.hp.hpl.jena.sparql.util.Context();
i assume we trapped the null for some.. good reason10:42
<eddies>(afk…have SG call in a few mins)
<cbeer>although i don't see any indication in the jena docs that getContext should return null for any good reason
<leenata>well, i got to understand that all the missed linecoverages from fedoraresourceTest had to be junit tested10:43
<barmintor>leenata: those ignored assertions aren't going to do anything10:44
b/c the mock doesn't actually have any behaviors
so does the null case have to be left the way it was?
<cbeer>oh, eddies: a little more morning grumbling: next time you, say, upgrade sonar.. make sure it can restart afterwards10:46
the sonar guys very helpfully replaced our config with some defaults, from what i can tell10:47
leenata: i'm questioning if there actually is a null case, or if that's a needless check10:49
<eddies>cbeer: i did restart sonar after upgrading (that was over a week ago)10:53
part of the upgrade was restarting sonar
<cbeer>eddies: then i'm even more worried.
* unai_ leaves10:54
<cbeer>maybe it has evil first-run logic?
anyway, it was down (explaining e.g. the fcrepo-jmeter failure)
<eddies>eh. we're on sonar 3.6...10:56
which was just released last week
i don't remember upgrading sonar last week
i wonder if sonar is config'ed to autoupgrade?!
* github-ff joins10:58
[fcrepo4] cbeer pushed 1 new commit to master: http://git.io/f_uW_A
fcrepo4/master 8e5f9b5 Chris Beer: prune some dead logic for null dataset contexts
* github-ff leaves
<cbeer>leenata: i could find no signs that that branch was actually being used, and it didn't break any of the other tests
<eddies>fyi: i'm temporarily disabling kitchen-sink on jenkins while the SG call is going11:00
<cbeer>they're actually going to look at fcrepo4?!11:01
* kaarefc joins11:10
* ajs6f joins11:12
I heard those SG crazies are actually looking at the app?!?11:13
<cbeer>"no, no, you're doing it all wrong! the left margin should be 3px!"11:14
* travis-ci joins11:15
[travis-ci] futures/fcrepo4#832 (master - 8e5f9b5 : Chris Beer): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/c9b336570264...8e5f9b558a05
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/8660015
* travis-ci leaves
_That's_ how I feel about it this morning.
One beer in PEI for whoever remembers where that came from.11:18
afk bbi511:19
* ajs6f leaves
* ajs6f joins11:23
* github-ff joins11:30
[fcrepo4] ajs6f pushed 1 new commit to OAuthIntoWebapp: http://git.io/Spikhg
fcrepo4/OAuthIntoWebapp eea2ad6 ajs6f: Bringing fcrepo-auth-oauth under fcrepo4
* github-ff leaves
<cbeer>ajs6f: do you have any time today to (help with/think about/write) https://wiki.duraspace.org/display/FF/High-level+architecture? i'm drawing a blank about how to write (usefully!) about the architecture without going too far into the impl11:52
<ajs6f>cbeer: That's because we haven't done the arch. We just did kernel-ization, not modularization.
We don't have a formal arch.
But yes, I can help.11:53
You want to talk (voice)?
<cbeer>sure, we can do that. in about 5m?
<ajs6f>Too soon— I'm not where I can talk loudly. Maybe 30 min? Too late?
<cbeer>sure, works for me
<ajs6f>It's on like Donkey Kong.11:54
One thought: rather than describe a vague and emergent arch, maybe a "Platform Description" with capabilities and a discussion of layers (such as you've already written most of)?
(high-level capabilities)11:55
<cbeer>sure. at least that'd get rid of some loaded terms and expectations11:58
Fedora is boozy enough as it is. We don't want to get any more loaded.
cbeer/eddies: That PR (90) has the Oauth module under fcrepo4 now. It's waiting for barmintor's review.11:59
<ajs6f>Thanks, but I'm still not convinced. But then, convincing me of anything is rarely worth the effort.12:00
<cbeer>i'm convinced, but not necessarily thrilled. but until we have something pluggable, we might as well have that.
<ajs6f>Yeah, I'm not convinced we need something pluggable in core, but I'm certainly not going to make a big stink over it. I'm more concerned immedaitely to (ASAP after OR) to break apart fcrepo4.12:01
<eddies>fyi, i've re-enabled kitchen-sink12:02
<ajs6f>eddies: You mean CI-deployment?
<eddies>ah yes =)
that wasn't terribly clear
<ajs6f>Nothing about Fedora is terribly clear. It's more like… horribly ominous.12:03
eddies: I just noticed that Cargo has a cargo:package goal— you ever try that?12:14
Might be an easy way to produce a quick-runnable (no build) artifact...12:15
<barmintor>Fedora 4: Horribly ominous!
<ajs6f>The Lemony Snicket repo framework.
Architected by Edward Gorey.
afk bbi512:19
* ajs6f leaves
* ajs6f joins12:24
<cbeer>ajs6f: ok, i've promoted that copy to the https://wiki.duraspace.org/display/FF/Features%2C+FAQs+and+How-Tos page12:26
(letting us get away with saying even less!)
<ajs6f>I like.
If someone really wants to get into the architectural questions, send 'em my way. I will bore them to tears, and when I'm done, I'll sick Dan Davis on 'em.12:27
sic, not sick
<cbeer>then here's the other page i have lots of doubts about: https://wiki.duraspace.org/display/FF/The+Fedora+4+object+model12:28
i basically lifted the JCR model description and junked the stuff we don't care about
<ajs6f>It's all true, but didn't you write a few sentences somewhere about the diffs between fcrepo3 and 4?12:29
so, right. it's not clear how all these pages interrelate
<ajs6f>Right, as long as it's up there somewhere.
That's true.
Maybe just a link or two.
I'll put a link to https://wiki.duraspace.org/display/FF/Fedora+3+%3D%3E+Fedora+4+migration12:31
in the other page.
<cbeer>"if you know the fcrepo3 model, read about the differences here.." sure
<cbeer>or, you could. if we had actually said something useful
<eddies>ajs6f: never tried cargo:package12:32
(sorry for the delay, on another call)
<ajs6f>First, a link. Then I'll enhance https://wiki.duraspace.org/display/FF/Fedora+3+%3D%3E+Fedora+4+migration
<ajs6f>eddies: I;ll give it a try. If it's really easy, could be nice to have on board.
<cbeer>eddies: on https://wiki.duraspace.org/display/REPONEXT/Fedora+4.0+Alpha+1+Release+Notes.. by announcing these features at the same level, are we obligating ourselves to do all of them equally well?12:33
i wonder if organizing it by core principles vs stuff we've tinkered with would help12:34
or, put another way, i don't want someone seeing fcrepo4 supports policy driven storage, realizes it's kinda incomplete, and thinks fcrepo4 in general is like that12:37
<ajs6f>Isn't it? :)12:39
<cbeer>we've done a good job at SOME stuff at least12:40
<ajs6f>He says, speaking for himself. I've just spent six months complaining that we're not using Scala.
<barmintor>cbeer: I like that angle on organizing things
<ajs6f>cbeer: Seriously, I agree.
Do we want to go as far as describing our list of priorities?12:41
<eddies>cbeer et al: how about keeping the functional grouping but having an indicator of half-baked work12:43
<ajs6f>Like a cake icon with runny batter oozing out of it?
<cbeer>'wow, this fcrepo4 group sure did a lot of half-baked work!'12:44
<eddies>as a manager, i want to see at a high level what the alpha's about and grouping via functional areas helps me assimilate
<ajs6f>"But if you take a little from each cake, you can still get dessert."
How about using color to distinguish?
<eddies>for folks who are concerned about specific line item, they can see if it's still work-in-progress
<cbeer>is this alpha actaully about those features, though? or the ability to actually make meaningful progress on figuring out what those features even are?12:45
<eddies>well those are items we're delivering in the alpha
<ajs6f>But not all in the same way...
cbeer: Isn't that your original point?
<cbeer>maybe. for me, at least, most of the alpha development was proving building those features out was possible12:46
proof-of-concept style
<ajs6f>eddies: Cargo works real good. Now I just have to figure out how to ZIP the app directory and make a new artifiact from it.
<cbeer>and i'd hate to lose that in "Here's the things you can do with fcrepo4"
<ajs6f>But there's a tension here:
<eddies>cbeer: that seems a bit too subtle for release notes12:47
<ajs6f>we want to stay loose and flexible, but we want people to start working with it, too
<eddies>or put another way, if you can wordsmith what you just said into release-note speak, we could try it12:48
i.e. patches welcome ;-)
<cbeer>that sounds like real work!
huh. i can't create pages in REPONEXT?
that's new
<ajs6f>Finally: http://qz.com/99413/first-ever-human-head-transplant-is-now-possible-says-neuroscientist/12:49
Now we can graft eddies' head onto awoods' body and vice versa. Much easier handover of tech leads.
<cbeer>maybe i'm off the project?
<eddies>ajs6f: i saw uva in the news today: http://www.dailyprogress.com/news/bottled-water-purchase-leads-to-night-in-jail-for-uva/article_b5ab5f62-df9b-11e2-81c4-0019bb30f31a.html
<cbeer>eddies: should i just draft some stuff in FF, or a google doc, or do you have the power to give me write access again?
<eddies>cbeer: i can have a look. i hate mucking w/ perms in confluence
<ajs6f>eddies: Yeah, that's a real cluster*&*). The town is pretty angry about it.
Lots of people are asking why the local police (who would not have done something that stupid) weren't involved.12:51
<eddies>cbeer: hrm. it looks like all the perms are set up using Crowd groups on that space12:52
and i think only duraspace staffers have access to crowd
i can add you as a one-off
<cbeer>ajs6f: do you still have write access there too?
<eddies>but i think jonathan markow just redid all the perms there
<cbeer>no sense drafting something if you have to re-add all of us
s/drafting something/drafting something there/12:53
<ajs6f>cbeer: dunno, but I can check. Which page are we talking about now?
<cbeer>ajs6f: anything in REPONEXT
<eddies>it looks like jonathan (i assume) created a reponext-committer group, which should have page add permissions
but i have no way of seeing who is a member of that group
<cbeer>i'm happy to draft publicly, unless there's a reason to keep it under wraps?12:54
<eddies>doesn't matter
oh, if/when you talk to jonathan about perms
it'd be nice to have permission to delete pages in FF
<eddies>i just started in reponext for the release notes cause i thought i might as well keep it under wraps until it was "ready"
you can't delete pages?
<eddies>i think these are things to queue up for andrew
he should be able to do all the crowd managment
<cbeer>ok. i can bug him after OR
duraspace.org really needs a UX designer12:56
and/or when your list of tags is longer than the blurb, you're doing it wrong.12:57
<eddies>ajs6f: la croix should use this as a local marketing opportunity. La Croix, it should be a crime to have bubbly this good.
<ajs6f>Use what as a l. m. o.?12:58
<eddies>if only you or robin were involved we could have tied it into some fedora marketing. the bottle water arrest :P
<ajs6f>Oh. {sigh}
<eddies>see. you didn't know the brand of water. a missed opportunity
<ajs6f>Fedora: You've never seen access control this tight.12:59
<barmintor>Fedora: You did it because there was a gun pointed at you.13:00
<ajs6f>Well dog my cats!13:10
eddies: cargo:package works.
I can build prepacked Tomcats with fcrepo-webapp in them, ready to run.13:11
Would Jetty be better?
<cbeer>can you?
<ajs6f>I'll try.13:12
<barmintor>ajs6f: uh oh. Don't go undermining the use-case for my FCR3 fork that makes FEDORA_HOME the FileSystem impl of a config/IO provider
<ajs6f>barmintor: I already went through this Cargo crap with fcrepo3 for the triplegenerator module I'm postering at OR. I'm not doing it again for love or money.13:13
<barmintor>ajs6f: well, I'm going to be writing FCR3 JAXRS resources for probably another year, so I need to work out some kind of testing framework13:15
<ajs6f>barmintor: Cargo is the best thing you have for system tests. I think eddies would agree.13:16
See this pom.xml:
Or look at MediaShelf's client.
<cbeer>eddies: ok, i've gotten tired of typing and need to head to the office, but here's the gist of my thinking: https://wiki.duraspace.org/pages/viewpage.action?pageId=3465634613:17
obviously the features actually need to say what they are, probably taking your copy
(and needs to be polished, lose the snark, etc)
<barmintor>ajs6f: sure, I just can't bear the system test approach to development. *start fedora* did it work? *stop fedora, make changes, copy new jar* did it work? (repeat)13:18
<ajs6f>barmintor: That's what Cargo fixes.
It requires some maven craftiness at first, but then tests just run.
It's the best band aid I've found for fcrepo3 monolithic quality.
<cbeer>heading to the office. back in a bit.13:21
* bljenkins leaves
* bljenkins joins13:39
* bljenkins leaves14:21
* bljenkins joins14:27
<ajs6f>Can anyone tell me why we have a security realm enabled in fcrepo-webapp?14:29
It's giving me fits trying to get this packaging thing working with jetty.
And it's not clear to me that we're actually using it for anything.
afk bbi514:31
* ajs6f leaves
* ajs6f joins14:34
* github-ff joins14:38
[fcrepo4] ajs6f created RemoveSecRealm (+1 new commit): http://git.io/NR0ARA
fcrepo4/RemoveSecRealm c24b45e ajs6f: Remove unused security apparatus from web.xml
* github-ff leaves
* github-ff joins14:39
[fcrepo4] ajs6f opened pull request #91: Remove unused security apparatus from web.xml (master...RemoveSecRealm) http://git.io/TBbyig
* github-ff leaves
<ajs6f>cbeer: If you have a chance (because I know you have nothing else to do) can you pull that? It just removes the unused security stuff from fcrepo-webapp's web.xml.
<cbeer>ajs6f: go for it. i'm trying to record audio for the first-run screencast.14:40
<barmintor>that looks uncontroversial, I'll hit the button.14:41
<ajs6f>barmintor to the rescue!
* github-ff joins
[fcrepo4] barmintor pushed 1 new commit to master: http://git.io/9IhrUg
fcrepo4/master 7a5c02d Benjamin Armintor: Merge pull request #91 from futures/RemoveSecRealm...
* github-ff leaves
<ajs6f>Woo-hoo! That fixed the stupid problems with getting Cargo to play with Jetty. Back on the road to a "download and run" artifact set.
* travis-ci joins14:58
[travis-ci] futures/fcrepo4#835 (master - 7a5c02d : Benjamin Armintor): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/8e5f9b558a05...7a5c02d3e716
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/8668326
* travis-ci leaves
* ajs6f1 joins15:01
* ajs6f leaves15:02
<cbeer>nice: https://github.com/blog/1547-release-your-software15:20
<ajs6f1>"binary assets"?
Sounds like they're bringing back downloads in a better-curated form.15:21
<barmintor>cbeer: just bookmarked that for later, looks interesting
ajs6f1: Ain't nothing wrong with helping me provide usable downloads
<ajs6f1>I was using them. Then they went away. I was sad.
<barmintor>Fedora > Core Test > #104 failed "No one is responsible for this build."15:50
<ajs6f1>That's been Fedora's problem for years now.15:51
afk bbi1015:53
* ajs6f1 leaves
* ajs6f joins15:56
<barmintor>hooray, that failed test passes in al configs on my desktop15:58
<ajs6f>Fedora 4: It ran on Ben's machine?
<barmintor>ajs6f: FCR316:00
F4 tests never fail16:01
it's from THE FUTURE
<ajs6f>And in the future, it works. :)16:32
* ajs6f leaves16:44
* ajs6f joins16:46
* bljenkins leaves16:55
* bljenkins joins16:56
<ajs6f>Out for the evening. See y'all soon.16:57
* ajs6f leaves
* barmintor leaves17:02
* kaarefc leaves17:06
* ksclarke leaves18:13
* jonathangee_ joins21:00
* jonathangee leaves
* ksclarke joins21:12
* ksclarke leaves23:48

Generated by Sualtam