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

    witty-crayon-22786

    07/01/2020, 9:31 PM
    so this is only for “non-fatal” errors
  • w

    witty-crayon-22786

    07/01/2020, 9:31 PM
    a test failing this way fails as a 
    FalliableProcessResult
    , which is not an error according to the engine.
  • a

    aloof-angle-91616

    07/01/2020, 9:31 PM
    yes
  • a

    aloof-angle-91616

    07/01/2020, 9:49 PM
    is there a reason we decided 
    def global_subsystems()
    was v1-only? if i define a context-aware object factory in my v2 ruleset which pulls in a subsystem using 
    .global_instance()
    , i need that subsystem to be available. i was able to overcome this by just registering it as a v1 plugin as well, but unless we plan to literally deprecate 
    .global_instance()
    for v2 (which it's not clear that we are) it seems like we should allow 
    global_subsystems
    to be in v2 plugins as well
    h
    w
    • 3
    • 21
  • h

    hundreds-father-404

    07/02/2020, 4:21 AM
    seen in CI
    Untitled
  • a

    aloof-angle-91616

    07/02/2020, 5:26 AM
    i'm getting this error when attempting to 
    ./pants list
    with literally any target spec on pants 2.0.0dev0, and it's only reproing on linux, not osx (i tried two separate linux environments). are both of these failures likely to be from running out of memory (possibly running out of shared memory in LMDB)? this occurs without pantsd as well.
    Untitled
    • 1
    • 9
  • w

    wonderful-iron-54019

    07/02/2020, 1:09 PM
    [Tests V1 formatting thread] <<wrong channel please ignore>>
    • 1
    • 9
  • h

    hundreds-father-404

    07/02/2020, 4:01 PM
    Hey there’s a lot of noise in CI now from things like 
    Starting: store_bytes
    . Could we change those to DEBUG level?
    w
    f
    h
    • 4
    • 21
  • h

    hundreds-breakfast-49010

    07/02/2020, 11:21 PM
    is there anything we want to keep from the v1 
    reporting
    infrastructure?
    h
    • 2
    • 1
  • h

    hundreds-breakfast-49010

    07/02/2020, 11:21 PM
    re: log colors, I discovered that there's this mapping in some of that code:
    Copy code
    _COLOR_BY_LEVEL = {
        Report.ERROR: red,
        Report.WARN: yellow,
        <http://Report.INFO|Report.INFO>: green,
        Report.DEBUG: cyan,
    }
  • h

    hundreds-breakfast-49010

    07/02/2020, 11:21 PM
    which we could re-use
  • h

    hundreds-breakfast-49010

    07/02/2020, 11:21 PM
    in principle
  • h

    hundreds-breakfast-49010

    07/02/2020, 11:22 PM
    or at least adapting the currently-used pants colors of magenta, cyan, green, and red to this schema. although I don't think we want DEBUG to have a color
  • h

    hundreds-breakfast-49010

    07/02/2020, 11:22 PM
    or INFO for that matter
  • h

    hundreds-breakfast-49010

    07/02/2020, 11:23 PM
    well, hm, maybe we do? maybe there's value in distinguishing INFO and DEBUG with colors
  • h

    hundreds-father-404

    07/02/2020, 11:23 PM
    Is this just for the 
    [INFO]
    header, rather than the whole message? If so, that’s helpful. Otherwise, it’s super distracting
  • h

    hundreds-breakfast-49010

    07/02/2020, 11:26 PM
    I guess if we made only the header have color, then we can use more colors without it being distracting
    👍 1
  • h

    hundreds-breakfast-49010

    07/02/2020, 11:26 PM
    i..e every level header could in principle have a different color, but the text would be the ordinary terminal text color in all cases
    👍 1
  • h

    hundreds-breakfast-49010

    07/02/2020, 11:27 PM
    btw @hundreds-father-404 is there any logic anywhere in patns to try to detect the background terminal color and just text color based on that, or did you just try to select colors that look good with either light or dark backgrounds?
    h
    • 2
    • 1
  • h

    hundreds-father-404

    07/02/2020, 11:28 PM
    I like that. But still I think keep <=4 colors, in line with web design principles about using a limited color palette
    👍 1
    h
    • 2
    • 4
  • h

    hundreds-breakfast-49010

    07/02/2020, 11:30 PM
    but in any case, there's a bunch of infratructure in 
    reporting
    that I think we can safely delete for 2.0?
    h
    • 2
    • 1
  • h

    hundreds-father-404

    07/02/2020, 11:51 PM
    I’m pretty consistently getting the 
    [INFO] Completed
    messages printed twice when running 
    ./pants --changed-since=HEAD fmt
    . I noticed this since this morning
    h
    • 2
    • 1
  • h

    hundreds-breakfast-49010

    07/03/2020, 12:01 AM
    I just realized - the 
    help
    goal doesn't respect the 
    --no-colors
    flag
    👍 1
    h
    • 2
    • 1
  • h

    hundreds-father-404

    07/03/2020, 1:03 AM
    Was there a performance regression to fingerprinting?
    Copy code
    v42.11s Fingerprinting: src/python/toolchain/prod/builders/build_es_curator.py
    w
    • 2
    • 2
  • f

    fast-nail-55400

    07/06/2020, 4:40 PM
    https://github.com/pantsbuild/pants/pull/10260 switches the remote execution client to the new “streaming” client which is able to use the 
    WaitExecution
    method of the REv2 API. Given Scoot does not support 
    WaitExecution
    , should we keep the original “polling” client or remove it (and lose Scoot support)?
  • a

    aloof-angle-91616

    07/06/2020, 4:42 PM
    it sounds like scoot should be made to support WaitExecution
  • f

    fast-nail-55400

    07/06/2020, 4:56 PM
    that is my thought too given 
    WaitExecution
    is a required part of the REv2 API
  • a

    aloof-angle-91616

    07/06/2020, 4:59 PM
    it’s unlikely that scoot will be developing this itself given that twitter is not making use of remote execution right now but there were concerns previously that it was using the pants fs_util tool to implement some parts of the api, which wasn’t very efficient. i created this library to help fix that https://github.com/cosmicexplorer/upc
  • a

    aloof-angle-91616

    07/06/2020, 5:00 PM
    i’d definitely like to see scoot implementing the rest of the remexec v2 api
  • f

    fast-nail-55400

    07/06/2020, 5:06 PM
    ah so they are not even using scoot then?
1...749750751...906Latest