Time
S
Nick
Message
07:30 bencomp joined #dataverse
11:13 bencomp joined #dataverse
12:41 pdurbin
bencomp: did you notice I changed the default branch from "master" to "4.2"
12:41 pdurbin
I'm trying to decide if I should switch it back.
15:03 axfelix joined #dataverse
15:04 axfelix joined #dataverse
16:02 JonathanNeal joined #dataverse
16:25 jeffspies_ joined #dataverse
16:35 metamattj joined #dataverse
16:36 bencomp
pdurbin: I saw and think it's fine
17:08 pdurbin
bencomp: mostly I think it's a good thing. It lets contributors know which branch we're working on for the next release. But there's one downside, I think.
17:10 pdurbin
bencomp: when we delete the 4.2 branch (after it's merged to master) the link to the contributing doc here will be a 404: https://groups.google.com/d/msg/dataverse-community/qujUFJDYXG0/0BhECCp0HQAJ
17:32 pdurbin
not the end of the world, though :)
17:32 pdurbin
hopefully it will solve the problem of people opening pull requests against the "wrong" branch
17:32 pdurbin
we definitely don't want pull requests against master
18:44 bencomp joined #dataverse
18:45 bencomp
pdurbin: you can never delete the branch then :)
18:46 pdurbin
bencomp: 4.2 is not a persistent branch: http://guides.dataverse.org/en/latest/developers/branching-strategy.html :)
18:47 bencomp
btw, pdurbin, are you related to a Michael Durbin who works on Fedora 4?
18:47 pdurbin
not that I know of!
18:47 pdurbin
whenever I hear of Fedora I think of the operating system
18:48 bencomp
I meant the repository system :)
18:50 pdurbin
yeah, the 4 gave it away :)
18:55 bencomp
did you see https://wiki.duraspace.org/display/FF/Fedora+4.3.0+Release+Notes ?
18:55 pdurbin
uh. should I have?
18:55 * pdurbin
looks
18:55 bencomp
migration utils, audit service, reindex endpoint
18:56 pdurbin
ok
18:57 bencomp
I consider Fedora based repository software competitors for Dataversee
18:58 pdurbin
oh, sure. I didn't know they made a release recently
19:01 bencomp
looking at the version number, fcrepo is ahead ;)
19:02 pdurbin
heh
19:02 pdurbin
really we should have called it Dataverse 1.0 since it's a rewrite
19:04 bencomp
2015 is the year of 4
19:05 pdurbin
:)
19:05 bencomp
iRODS, Fedora Repo, Dataverse
19:05 pdurbin
:)
19:05 bencomp
Q (and don't cheat by looking what happens in the code)...
19:06 bencomp
should `ingestableAsTable(DataFile)` where the argument is null return false or throw a NPE?
19:09 pdurbin
hmm. we should probably use Checker to document if you can pass null to a method or not
19:09 pdurbin
searchbot: lucky java checker
19:09 searchbot
pdurbin: https://www.java.com/verify
19:09 pdurbin
bah
19:09 pdurbin
searchbot: lucky java checker framework
19:09 searchbot
pdurbin: http://types.cs.washington.edu/checker-framework/
19:09 pdurbin
that's the one. I went to a talk about it at JavaOne
19:10 bencomp
interesting
19:11 pdurbin
Preventing Errors Before They Happen. here are my notes: http://irclog.greptilian.com/sourcefu/2014-09-30#i_78554
19:11 bencomp
I generated some unit tests and went quick and dirty on the test implementations
19:12 bencomp
pdurbin: I'm slightly disappointed that javac reports unchecked conversions and raw type references in SearchIT.java
19:12 pdurbin
whoops
19:15 pdurbin
hmm, I guess SearchIT isn't compiled by mvn -DcompilerArgument=-Xlint:unchecked compile
19:15 pdurbin
not sure how to add it
19:17 bencomp
I see warnings when I `mvn test`
19:17 pdurbin
in SearchIT?
19:24 bencomp
yes
19:27 pdurbin
bencomp: can you please stick it on http://danceb.in ?
19:27 pdurbin
(or your favorite pastebin)
19:29 bencomp
pdurbin: http://danceb.in/nZqPvE25RGkh6OxDMIysQ/raw
19:30 bencomp
I could add that I'm on the 4.0.2 branch with my small changes to DataFileServiceBean
19:30 bencomp
and a DataFileServiceBeanTest
19:31 pdurbin
ah. for me maven-compiler-plugin:3.1:testCompile shows "Nothing to compile - all classes are up to date"
19:32 bencomp
that means: add more tests
19:32 bencomp
:)
19:32 pdurbin
no, I think it means to run `mvn clean` :)
19:33 pdurbin
SearchIT.java:[255,39] unchecked conversion
19:33 pdurbin
there is it
19:33 pdurbin
on my uncommitted code
19:39 dataverse-user joined #dataverse
19:57 pdurbin
bencomp: hmm, actually that was quite helpful
20:02 bencomp
good :)
20:02 pdurbin
bencomp: ok, SearchIT should be clean as of https://github.com/IQSS/dataverse/commit/fdfd818
20:05 bencomp
it is, but you're teasing me by putting the warnings in a different file... http://danceb.in/RH1TRfY25RGUuaKxDMIysQ/raw
20:06 pdurbin
uh. which file?
20:07 bencomp
ShapefileHandlerTest.java
20:07 bencomp
surely it's not yours
20:07 pdurbin
that's https://github.com/raprasad :)
20:10 bencomp
hmm, it's quite an old file
20:11 pdurbin
I'm not too worried about the warnings in test files to be honest
20:11 pdurbin
let's not discourage people from writing tests :)
21:20 garnett joined #dataverse
22:27 bencomp
so I couldn't help myself again: https://github.com/IQSS/dataverse/pull/2422
22:31 pdurbin
bencomp: does it pass integration tests that exercise all of that?
22:32 bencomp
this is really unit test level code
22:33 pdurbin
well, the em.createQuery stuff isn't. is it?
22:33 pdurbin
since it touches the database
22:34 bencomp
hmm, you're probably right
22:34 bencomp
but I really worked to lower the number of compiler warnings
22:35 bencomp
i.e. my changes are on sub-unit-test level
22:35 pdurbin
you know, I'm sad to no longer have my apitest builds running anymore. it showed the number of compiler warnings
22:35 bencomp
what happened?
22:36 pdurbin
well, now that we are in production, https://apitest.dataverse.org runs the same version as https://dataverse.harvard.edu
22:36 pdurbin
(ok, ok, I need to upgrade it to 4.1)
22:36 pdurbin
whereas before apitest would run the latest code
22:36 bencomp
where did Jenkins go?
22:37 pdurbin
whenever i felt like doing a build, basically
22:37 pdurbin
jenkins didn't go anywhere :)
22:37 bencomp
why did it stop continuous delivery?
22:38 pdurbin
because I can't just upgrade apitest to a commit I made today. I have to wait for a release
22:38 pdurbin
basically, I need a new server that means "code we worked on today"
22:39 bencomp
vagrant up?
22:40 pdurbin
yeah, I guess
22:41 pdurbin
but it would be good to see the compiler warnings going down
22:41 pdurbin
on jenkins
22:42 bencomp
I found out today that Netbeans can talk to Jenkins
22:42 bencomp
or Hudson, which is the original Jenkins I thikn
22:42 pdurbin
hmm. ok
22:43 pdurbin
bencomp: speaking of Netbeans, is that how you reformatted the code?
22:43 bencomp
just Source > Remove Trailing Spaces
22:44 bencomp
other "reformatting" was manual
22:46 pdurbin
but with netbeans, ok
22:46 pdurbin
honestly, I'd rather have a command line tool that does the reformatting
22:47 pdurbin
they we could have it fail the build ;)
22:48 bencomp
don't get distracted now
22:49 bencomp
but isn't there a git option that makes git remove trailing spaces (in certain files, I suppose) before doing the actual commit?
22:50 pdurbin
hmm. dunno
22:51 pdurbin
but my looking at http://docs.travis-ci.com/user/languages/java/ I bet we could have Travis run mvn -DcompilerArgument=-Xlint:unchecked test
22:51 pdurbin
I think it goes under `script` according to http://docs.travis-ci.com/user/customizing-the-build/
22:51 pdurbin
bencomp: you could set up travis for your account and try it
22:52 bencomp
maybe some other time
22:52 pdurbin
and if it works... the next step would be to collect the number of compiler warnings somewhere
22:52 bencomp
Netbeans tells me it has 1116 warnings
22:53 bencomp
and 623 TODOs
22:56 bencomp
and nothing about missing integration tests
22:56 bencomp
:O
22:57 pdurbin
:)
22:57 pdurbin
ok, I'm trying it in a branch: https://travis-ci.org/IQSS/dataverse/builds/73469710
22:57 bencomp
I'll see what happened tomorrow
22:58 bencomp
it's 1 am so time for bed...
23:00 pdurbin
hmm, I think it worked but maybe it should go under "install" rather that "script"
23:30 bencomp
I'm really good at creating stubs of repositories :) https://github.com/bencomp/dataverse-docker
23:30 bencomp
good night
