https://pantsbuild.org/ logo
Powered by
#development
  • w

    witty-crayon-22786

    12/03/2019, 3:44 AM
    Hm. Yea, 
    union_member.function_name
    is a bit of a hack... at rule registration time we don't know what function implements a particular Get.
  • w

    witty-crayon-22786

    12/03/2019, 3:44 AM
    Neat though: thanks!
  • a

    aloof-angle-91616

    12/03/2019, 3:44 AM
    the 
    function_name
    part is just for parsing the 
    --query
    strings, not trying to do metaprogramming
  • w

    witty-crayon-22786

    12/03/2019, 3:45 AM
    Oh, I see.
  • a

    aloof-angle-91616

    12/03/2019, 3:46 AM
    i have it much much cleaner and will push the rebase soon
  • a

    aloof-angle-91616

    12/03/2019, 3:46 AM
    "soon" = this week is reasonable, i'm most of the way done and just have a rule registration error i brought upon myself
  • w

    witty-crayon-22786

    12/03/2019, 3:47 AM
    Hm, yea... I guess that you can just do this with unions... Neat!
  • a

    aloof-angle-91616

    12/03/2019, 3:47 AM
    ^_^! i thought it was super cool
  • w

    witty-crayon-22786

    12/03/2019, 3:47 AM
    One thing is that for something like formatting you might care about ordering of the members
  • a

    aloof-angle-91616

    12/03/2019, 3:47 AM
    ah, i hadn't considered that
  • a

    aloof-angle-91616

    12/03/2019, 3:47 AM
    hence more like filter chains like you said
  • w

    witty-crayon-22786

    12/03/2019, 3:48 AM
    Yea. But probably mostly fine to ignore...?
  • w

    witty-crayon-22786

    12/03/2019, 3:48 AM
    As long as it was stable, it would be something like "plugin registration order"
    👌 1
  • a

    aloof-angle-91616

    12/03/2019, 3:48 AM
    yes!
  • h

    hundreds-breakfast-49010

    12/03/2019, 10:25 PM
    I just went down a minor rabbit hole becuase I was looking at the docs for the 0.3.1 version of rust futures
  • h

    hundreds-breakfast-49010

    12/03/2019, 10:26 PM
    without realizing that we were still on 0.1.27
  • h

    hundreds-breakfast-49010

    12/03/2019, 10:26 PM
    which seems like kind of an old version of futures to be on, and I wonder if there's any reason we'd want to avoid updating to the latest version - also doesn't using async/await in rust require this?
    👍 1
    e
    w
    • 3
    • 9
  • h

    hundreds-breakfast-49010

    12/03/2019, 10:30 PM
    naively updating the crate in our Cargo.toml to 0.3.1 causes 13 errors 🙂
  • h

    hundreds-breakfast-49010

    12/03/2019, 10:30 PM
    not sure how hard they are to deal with
  • h

    hundreds-breakfast-49010

    12/06/2019, 1:16 AM
    does anyone know what the v2 UI is using this specific termion commit: https://github.com/pantsbuild/pants/blob/master/src/rust/engine/ui/Cargo.toml#L16 ? @red-balloon-89377
    r
    w
    • 3
    • 3
  • a

    aloof-angle-91616

    12/06/2019, 8:29 PM
    i made a google doc regarding a proposal for an interface to fetch specific cache entries from task output with `./pants fetch`: https://docs.google.com/document/d/1F-sqqYX77m5iB6116bXxYugOiPbOGPGRwd75wSZYwBY/edit?usp=sharing
    pants fetch proposal
  • a

    aloof-angle-91616

    12/06/2019, 9:04 PM
    ^one amendment after stu's suggestion: this could and probably should be using the LMDB store and the engine rust code to retrieve and materialize cache entries, instead of interacting with the v1 artifact cache. it's not 100% clear, but i was vaguely thinking an adaptor from v1 cache output to LMDB would be a necessary intermediate step, which should then make "porting to v2" a matter of removing all v1 tasks.
  • h

    hundreds-breakfast-49010

    12/06/2019, 11:50 PM
    is 
    pants_distdir
    set from config in the pants repo somehow?
  • h

    hundreds-breakfast-49010

    12/06/2019, 11:51 PM
    if I call 
    OptionValueContainer.get_rank("pants_distdir")
    in pants in the pants repo, I get the 
    CONFIG
    rank
  • h

    hundreds-breakfast-49010

    12/06/2019, 11:52 PM
    but it doesn't look like 
    pants.ini
    sets it
  • w

    witty-crayon-22786

    12/07/2019, 12:15 AM
    @hundreds-breakfast-49010: it will be in tests
  • h

    hundreds-breakfast-49010

    12/07/2019, 3:38 AM
    @witty-crayon-22786 this was outside of a test though, I was just printing what 
    get_rank
    got me in the pants repo
  • w

    witty-crayon-22786

    12/07/2019, 3:40 AM
    Hm! Yea, weird.
  • h

    hundreds-breakfast-49010

    12/07/2019, 3:42 AM
    it would be convenient for a thing I'm working on if I could detect whether pants-distdir was set explicitly by a user
    a
    • 2
    • 3
  • h

    hundreds-breakfast-49010

    12/07/2019, 3:42 AM
    I can work around it, but now I'm kinda curious why this isn't working how I expect it to
1...679680681...906Latest