IQSS logo

IRC log for #dataverse, 2018-09-19

Connect via chat.dataverse.org to discuss Dataverse (dataverse.org, an open source web application for sharing, citing, analyzing, and preserving research data) with users and developers.

| Channels | #dataverse index | Today | | Search | Google Search | Plain-Text | plain, newest first | summary

All times shown according to UTC.

Time S Nick Message
09:18 juancorr joined #dataverse
10:52 juancorr joined #dataverse
13:31 donsizemore joined #dataverse
13:55 pdurbin donsizemore: mornin
14:18 poikilotherm joined #dataverse
14:18 poikilotherm Morning guys... ;-)
14:18 poikilotherm Currently starting to write unit tests for S3AccessIO
14:19 poikilotherm And stumbled over the first thing while mocking S3: I need to have org.junit.ClassRule available, but this is only implemented in JUnit 4.9+
14:20 poikilotherm Within the Maven pom.xml, the JUnit4 version is fixed to 4.8.1
14:20 poikilotherm Did you ever try to update that to at least 4.9? Are there any third parties relying on <4.9 ?
14:21 poikilotherm There seems to be some upgrade pain (https://github.com/junit-team/junit4/issues/357), but this sound like "take care of it in 2018" as JUnit 5 is here to stay...
14:21 pdurbin poikilotherm: hi! If you want I can re-open https://github.com/IQSS/dataverse/issues/2940 and assign it to you.
14:22 poikilotherm Wow, this feels like digging deeeep into a rabbit hole anytime I look at the code :-)
14:22 pdurbin :)
14:23 pdurbin welcome to our world :)
14:23 poikilotherm Ok, then I will reopen, change the version and then let's see what happens. I don't have the docker-aio testing on place right now - would it be ok if you test the IT tests etc on phoenix?
14:23 poikilotherm I want to be absolutely sure that no tests break by updating that version
14:24 poikilotherm (I bet it will fail... I know it will... Bad karma...)
14:24 pdurbin Sorry, phoenix is only for *after* the pull request is merged. It tests the "develop" branch.
14:24 donsizemore @pdurbin covering my eyes and hitting return https://github.com/IQSS/dataverse-ansible/commit/42f9ef63118d99142ed5698bc206f838a547354c
14:24 poikilotherm Meh...
14:24 pameyer joined #dataverse
14:25 pdurbin donsizemore: thanks! Looks like it has the change to "repo" that I need. Please stand by.
14:26 poikilotherm Is Phoenix just using the run-test-suite.sh?
14:26 poikilotherm (In addition to the mvn test / mvn verify)
14:26 pameyer poikilotherm: I'm not sure if it's been switch over to use that file, but it should be using the same list / syntax
14:26 pdurbin Sadly, no. We copy and paste the list of tests in that script over to the Jenkins job as needed.
14:27 poikilotherm Ok, but then I know I got everything covered and don't have to fear about things exploding after merging, right?
14:27 pameyer my efforts to spread the word of jenkins declarative pipelines have been unconvincing
14:27 poikilotherm Jenkins... Brrr....
14:28 poikilotherm However: anything else to think of?
14:28 pdurbin poikilotherm: ideally, you'd use the docker-aio stuff to run the integration tests.
14:28 poikilotherm pdurbin: Sure, I'll do that. Just asking if there is anything else that might blow up
14:28 poikilotherm (Which is not covered by the docker-aio testing)
14:28 pameyer poikilotherm - it's possible, although probably more annoying than it needs to be, to measure code coverage with the integration tests if you're concerned about breaking things in a specific area
14:29 poikilotherm I don't think that is going to blow by upgrading the Junit version...
14:29 pameyer there's what feels like an unfortunately large fraction of stuff that's uncovered
14:29 pdurbin I hope I haven't lost track of what we're talking about. We're talking about bumping the version of JUnit we're using, right? I'm not too concerned that stuff will break but who knows. :)
14:29 pameyer yeah - upgrading junit is probably low probability of breaking production somewhere
14:29 poikilotherm Ok, that definitly won't be blown by using a newer version of Junit anyway :-D
14:31 poikilotherm pdurbin: thx for assigning ;-)
14:33 donsizemore @pdurbin the warfile still reflects 4.9.2 but that may be an artifact from dvinstall.zip. i was... not able to work on this in an interrupted fashion so please holler at me for anything i missed.
14:34 pdurbin donsizemore: no worries. Do you know about https://github.com/IQSS/dataverse/blob/v4.9.2/scripts/installer/custom-build-number ?
14:34 pameyer and from the last time I run the coverage measurements, S3AccessIO is at 0%
14:35 pameyer SwiftAccessIO is somehow at 7%, despite me not having access to a swift system.  but FileAccessIO is at 76%
14:36 pdurbin pameyer: we should put your report somewhere so we can see it
14:37 pameyer pdurbin: might not be a bad idea; but I also wouldn't be surprised if parts of it are out of date
14:37 pdurbin sure
14:39 juancorr21 joined #dataverse
14:43 pdurbin pameyer: what we really want, of course, is to run the code coverage report regularly. Every night or whatever. I don't think we have a GitHub issue for this.
14:44 pameyer pdurbin: well, running ITs from cron (or equivalent) got bumped - it seems like coverage reports might be a step or two past that
14:45 pdurbin yeah, who knows
14:46 pdurbin donsizemore: hmm, 4.9.2 was deployed to AWS rather than the branch I specified, which was "4610-upload-dual-mode". Any ideas for me?
14:46 poikilotherm Ok, doesn't look that bad - only one test failed...
14:46 pdurbin poikilotherm: nice
14:47 poikilotherm Will dig into that... (IngestableDataCheckerTest using a @Rule)
14:47 pdurbin poikilotherm: sounds good. Thanks.
14:48 poikilotherm Will dig into that later or tomorrow - my kids are waiting for me ;-) (5pm over here...)
14:48 poikilotherm Cu
14:48 pdurbin o/
14:48 poikilotherm left #dataverse
14:56 donsizemore @pdurbin i'll retrace the build steps... and i'll check out custom-build-number, thank you
15:00 pdurbin donsizemore: thanks! Please take a look at the docs I just wrote at https://github.com/IQSS/dataverse/commit/0ae6dc1 when you get a chance.
15:02 pameyer pdurbin: are you seeing IT failures on develop?
15:02 pameyer dd4ba227c50507989ed011de7b1ef69432a6a96c
15:04 pameyer nm - https://build.hmdc.harvard.edu:8443/job/phoenix.dataverse.org-apitest-develop/
15:16 donsizemore @pdurbin fixed a stupid mistake i should've caught. i'm doing that XKCD "compiling!" thing and heading for lunch, then will push a fix once i've verified
15:30 pdurbin donsizemore: awesome that you found it so quickly. Please ping me once you push.
15:35 pameyer joined #dataverse
16:42 donsizemore @pdurbin testing one more time with 4610-upload-dual-mode
16:43 pdurbin donsizemore: great! Thanks!
17:03 donsizemore so, i think it's building the correct warfile, i just need to add your custom build number to Build.properties
17:18 pdurbin donsizemore: if it's easy. No rush on the custom build number thing.
17:19 donsizemore @pdurbin it's currently my top priority =)
17:20 pdurbin I love it. :)
17:29 pameyer joined #dataverse
17:30 pameyer @donsizemore there's a helper script for the custom build number
17:38 pdurbin pameyer: yep. He's on it.
17:39 pameyer great :)
17:58 poikilotherm joined #dataverse
18:02 pameyer joined #dataverse
18:05 poikilotherm Hey guys... Does anybody disagree on adapting JUnit 5 in parallel with JUnit 4? See https://github.com/IQSS/dataverse/issues/2940
18:06 poikilotherm That's quite some work and if that would be just wasted time because QA is not buying it, that really would be a pitty...
18:06 poikilotherm (Needs testing the tests, maybe some code fixing, etc)
18:16 donsizemore @pdurbin it's still picking up 4.9.2 from... somewhere, but: https://github.com/IQSS/dataverse-ansible/commit/d4dc7cbc213f1fbb0e20ad08c571dfc2f8c22e68
18:28 pameyer joined #dataverse
19:28 poikilotherm left #dataverse
19:44 donsizemore joined #dataverse
19:44 pdurbin_m joined #dataverse
19:45 pdurbin_m donsizemore: thanks. So, a new commit but it's still always getting 4.9.2 rather than the specified branch?
19:46 donsizemore @pdurbin_m it's building the specified branch and deploying the correct warfile, but the homepage returns "4.9.2 4610-upload-dual-mode" (which, in a way, is exactly right)
19:48 pdurbin_m oh! that sounds just fine
19:48 donsizemore i trust you've seen the most "Boston" Boston Globe video ever?
19:48 pdurbin_m looking forward to trying it but I'm in a meeting
19:49 pdurbin_m no, link please
19:50 donsizemore this had me cackling https://www.bostonglobe.com/metro/2018/09/19/red-sox-fans-find-division-title-banner-somerville-street/8GZmeXh5TVkEyyocd2vWiO/story.html
19:57 Thalia_UM joined #dataverse
20:10 pdurbin wow
20:13 pameyer joined #dataverse
20:20 pameyer pdurbin: any thoughts on the junit4 in parallel with junit5 question?  I probably don't have enough expertise to have a useful opinion on it
20:24 pdurbin Oh. Parallel? I thought it was one or the other.
20:24 pameyer check the ircbot
20:26 pameyer I tried to copy/paste the link, and it didn't work for me today :(
20:36 pdurbin pameyer: ah, thanks. I left a comment.
20:39 pdurbin Where's Don. It worked! Thanks!
20:39 pdurbin I guess I should spin this down since I'm leaving for the day. I think Amazon charges by the minute.
21:22 pameyer joined #dataverse
21:32 pameyer joined #dataverse
21:49 donsizeore joined #dataverse

| Channels | #dataverse index | Today | | Search | Google Search | Plain-Text | plain, newest first | summary

Connect via chat.dataverse.org to discuss Dataverse (dataverse.org, an open source web application for sharing, citing, analyzing, and preserving research data) with users and developers.