https://pantsbuild.org/ logo
Powered by
#general
  • h

    happy-kitchen-89482

    01/29/2016, 10:45 PM
    They succeed by not crashing or something?
  • r

    rough-minister-58256

    01/29/2016, 10:47 PM
    presumably the top level code gets exec’d and can explode (or succeed) w/o concrete tests being actually defined -> pytest. seems like a n00b mistake tho.
  • h

    happy-kitchen-89482

    01/29/2016, 10:48 PM
    how many of these do you have? I’d prefer if you could wrap them in test functions rather than change pants to accommodate pathologies like that
  • w

    witty-crayon-22786

    01/29/2016, 10:50 PM
    heads up! executing the release. please avoid merging anything for the next 30 or so
  • h

    happy-kitchen-89482

    01/29/2016, 11:02 PM
    @limited-terabyte-34878: suggested a combined RE, and also a better way of accessing the matched group
  • h

    happy-kitchen-89482

    01/29/2016, 11:02 PM
    But as always with regexes, testing is everything
  • w

    witty-crayon-22786

    01/29/2016, 11:12 PM
    ok, master is open for commits
  • w

    witty-crayon-22786

    01/29/2016, 11:12 PM
    thanks everyone
  • w

    witty-crayon-22786

    01/30/2016, 1:27 AM
    oof, went way over my annual dose of Makefile... https://github.com/tkruse/build-bench/pull/6
  • b

    bored-art-40741

    01/30/2016, 2:04 AM
    @witty-crayon-22786: Why does releasing have anything to do with pushing?
  • w

    witty-crayon-22786

    01/30/2016, 4:35 AM
    guess it doesn't matter. forgot that it was a tag and not a commit
  • b

    bored-art-40741

    01/30/2016, 5:52 AM
    Yeah thought so as well. Was worried that I had been doing it wrong
  • m

    many-engine-77019

    01/30/2016, 11:37 PM
  • b

    bored-art-40741

    01/30/2016, 11:38 PM
    argh why does it do that
  • b

    bored-art-40741

    01/30/2016, 11:38 PM
    I thought I fixed that
  • b

    bored-art-40741

    01/30/2016, 11:39 PM
    Anyway, for now just think of it as a helpful reminder to run CI!
    😆 1
  • m

    many-engine-77019

    01/30/2016, 11:39 PM
    simple smile
  • l

    limited-terabyte-34878

    02/01/2016, 4:57 PM
    thanks @happy-kitchen-89482 — looks like we need the non-greedy otherwise it just captures the whole thing in the first group
  • l

    limited-terabyte-34878

    02/01/2016, 4:57 PM
    but i was able to use your single regex to get this to work with just a slight mod
  • l

    limited-terabyte-34878

    02/01/2016, 10:18 PM
    hm looks like junit runner sometimes doesn’t drop the failed target into the exception either.
  • w

    witty-crayon-22786

    02/01/2016, 10:30 PM
    @happy-kitchen-89482, @ancient-stone-50795 : hey... i'd like to make one last plea for a decorator based approach to the public/private APIs
  • w

    witty-crayon-22786

    02/01/2016, 10:30 PM
    i don't think all of the churn we'd be introducing to APIs names would be worth it at this juncture
  • w

    witty-crayon-22786

    02/01/2016, 10:31 PM
    and i don't think a blacklist based approach (renaming things that are private) is as maintainable as a whitelist approach (marking public things public)
  • r

    rough-minister-58256

    02/01/2016, 10:32 PM
    whats the alternative to a decorator based approach?
  • m

    mammoth-pharmacist-54215

    02/01/2016, 10:33 PM
    a lot of refactoring
  • w

    witty-crayon-22786

    02/01/2016, 10:33 PM
  • m

    mammoth-pharmacist-54215

    02/01/2016, 10:33 PM
    renaming things with either private_<method> or _<method>
  • r

    rough-minister-58256

    02/01/2016, 10:34 PM
    the only issue w/ decorators is that it adds at least one extra function execution to every method call
  • w

    witty-crayon-22786

    02/01/2016, 10:34 PM
    blargh
  • m

    mammoth-pharmacist-54215

    02/01/2016, 10:34 PM
    I think what stu is talking about is basically a no-op decorator that is just used as an annotation
1...939940941...2420Latest