These are chat archives for coala/coala-bears

3rd
Jan 2017
euleram
@vijeth-aradhya
Jan 03 2017 14:38
@jayvdb Can you assign me to coala/documentation#379
mixih
@Mixih
Jan 03 2017 14:40
done, btw pinging for assignment can make it slower because other maintainers usually ignore tagged requests:)
euleram
@vijeth-aradhya
Jan 03 2017 14:42
sorry had gone out awhile @sils just clearing things out as to what is required, I should create a string representation for the correction result and compare it with the test results right ( probably using assert in a unit test )
@Mixih Oh :no_mouth: I should keep that in mind!!
Lasse Schuirmann
@sils
Jan 03 2017 14:45
@vijeth-aradhya use cobot assign!
euleram
@vijeth-aradhya
Jan 03 2017 14:47
Ahh right! I was going through the messages properly now :)
euleram
@vijeth-aradhya
Jan 03 2017 15:45
Guys reviews for coala/documentation#380 :)
Lasse Schuirmann
@sils
Jan 03 2017 15:47
cobot explain review
coala-bot
@coala-bot
Jan 03 2017 15:47

After creating your Pull Request, it is under the review process. This can be deduced from the process/pending review label. Now you have to wait for the reviewers to review your PR. You should not ask for reviews on our Gitter channel - we review those PRs continuously.

We're usually swamped with reviews, while you are waiting please review other people's PRs at coala.io/review: that helps you and will make your review happen faster as well. As a rule of thumb, for every review you receive, give at least one review to someone else!

For a good review, look at every commit on its own and place <sha> is ready or <sha> needs work comments on the pull request, be sure to remove other spacing like tabs. If you're done with a pull request, you can use cobot pull wip <pull URL> to mark it work in progress finally.

Lasse Schuirmann
@sils
Jan 03 2017 15:47
Hah! :P
mixih
@Mixih
Jan 03 2017 15:47
@sils cleared for release?
euleram
@vijeth-aradhya
Jan 03 2017 15:48
:+1: :sweat_smile:
Lasse Schuirmann
@sils
Jan 03 2017 15:48
@Mixih release all you want as long it's acked, there is a newline missing after the title underline in the rst but that hardly matters
mixih
@Mixih
Jan 03 2017 15:48
oh no
should I go repair it?
pypi won't like it
Lasse Schuirmann
@sils
Jan 03 2017 15:48
on master, not really needed for the release branch since it has no effect
pypi never gets to see that file
mixih
@Mixih
Jan 03 2017 15:49
weird...rstLintBear should have caught it....
euleram
@vijeth-aradhya
Jan 03 2017 16:49
@sils Can I please know once I again what you meant by not checking the testcases with any, and checking it with it's result? Should I write the error that occurs and compare it with whenever the test runs? Is that what oyu meant
Lasse Schuirmann
@sils
Jan 03 2017 16:52
what are you referring to?
euleram
@vijeth-aradhya
Jan 03 2017 16:52
#1223 GhcModBear remember? :)
Lasse Schuirmann
@sils
Jan 03 2017 17:10
we just had a discussion with @jayvdb that our test quality isn't sufficient and we have a lot of bears breaking
because we only test for the good and bad file with validate_bears
that's nice and well and it kind of gives you some basic assurances but you don't test if the results of the bear actually make sense
you just test if there are any results or not
so we need a test that actually checks the result for it's message and that tests that the source ranges are set properly
etc
Maximilian Scholz
@sims1253
Jan 03 2017 17:15
we have a result template the bears fill in right? Like giving a file line column severity diff and so on. Can we check if the given result matches what we expect?
lot of manual work
so we'd have to create a test file for each bear that ideally hold every case in a working and a broken way and we check if the results are right?
Lasse Schuirmann
@sils
Jan 03 2017 17:17
check existing tests, we have a few decorators that make the file for you and run the bear and then you can put asserts on the results
ideally we'd improve those testing mechanisms so they're as easy to use as verify_local_bear
euleram
@vijeth-aradhya
Jan 03 2017 17:28
@sils Hmm makes sense - I'll check mainly for the result message though - that's the important thing :+1:
But damn so many bears use only verify_local_bear
Lasse Schuirmann
@sils
Jan 03 2017 17:56
that's what we're trying to solve now
no merges with only verify local bear :)