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

    witty-crayon-22786

    11/15/2022, 9:02 PM
    For the cases we care about there are 800k files... I'm not worried about it triggering often enough: I'm worried about it triggering too often
  • b

    bitter-ability-32190

    11/15/2022, 9:02 PM
    Meet in the (logarithmic) middle?
  • w

    witty-crayon-22786

    11/15/2022, 9:03 PM
    If you're comfortable changing it as we go, I don't think we need to get it right when it first lands.
  • w

    witty-crayon-22786

    11/15/2022, 9:03 PM
    @ancient-vegetable-10556 will want to confirm that it does the right thing consistently for yarn/npm projects: that's what matters most right now
  • a

    ancient-vegetable-10556

    11/15/2022, 9:07 PM
    let me know when there’s a commit to check and I’ll take a look 🙂
  • b

    bitter-ability-32190

    11/15/2022, 9:18 PM
    I would 
    ./pants_from_sources
    but I wiped my machine and so I gota upgrade my plugins all over again 😂
  • b

    bitter-ability-32190

    11/15/2022, 9:20 PM
    @ancient-vegetable-10556 https://github.com/pantsbuild/pants/pull/17520/commits/02e1d0d742005303e94b5a5cc4a2def85bf1251e from https://github.com/pantsbuild/pants/pull/17520 I need to test it myself. There aren't any real unit tests for immut paths today 😕
  • a

    ancient-vegetable-10556

    11/15/2022, 9:20 PM
    OK, one moment
  • a

    ancient-vegetable-10556

    11/15/2022, 9:24 PM
    @bitter-ability-32190 probably worth merging 
    main
    in given there’s a pertinent caching fix in there
  • a

    ancient-vegetable-10556

    11/15/2022, 9:25 PM
    (0a3d096c5e6c2915274826e96b4076bc3a6dfab7) is the relevant commit on main
  • a

    ancient-vegetable-10556

    11/15/2022, 9:25 PM
    My numbers will be based on a commit that merges main into your branch
  • b

    bitter-ability-32190

    11/15/2022, 9:26 PM
    merged and pushed
  • a

    ancient-vegetable-10556

    11/15/2022, 9:27 PM
    ooook
  • a

    ancient-vegetable-10556

    11/15/2022, 9:29 PM
    Copy code
    # Run 1:
ENABLE_PANTSD=True ./pants_from_sources --pantsd-max-memory-usage=4GiB run   0.50s user 0.16s system 0% cpu 1:31.38 total

# Run 2:
ENABLE_PANTSD=True ./pants_from_sources --pantsd-max-memory-usage=4GiB run   0.48s user 0.21s system 3% cpu 17.487 total

# Run 3:
  • a

    ancient-vegetable-10556

    11/15/2022, 9:29 PM
    Copy code
    ENABLE_PANTSD=True ./pants_from_sources --pantsd-max-memory-usage=4GiB run   0.47s user 0.16s system 67% cpu 0.950 total
  • a

    ancient-vegetable-10556

    11/15/2022, 9:30 PM
    very well done!
  • w

    witty-crayon-22786

    11/15/2022, 9:36 PM
    ehh? under a second on the last run? that’s surprising
  • w

    witty-crayon-22786

    11/15/2022, 9:36 PM
    i’m generally surprised about the difference between the 2nd and 3rd runs: i think that the 2nd run should be taking advantage of immutable inputs from the 1st run
  • a

    ancient-vegetable-10556

    11/15/2022, 9:37 PM
    @witty-crayon-22786 this is vaguely what it was looking like when I put immutable inputs in manually
  • w

    witty-crayon-22786

    11/15/2022, 9:38 PM
    yea, weird. for followup i suppose.
  • a

    ancient-vegetable-10556

    11/15/2022, 9:38 PM
    are we marking things as immutable when we save an output? That seems surprising, but OK
  • w

    witty-crayon-22786

    11/15/2022, 9:38 PM
    no, outputs aren’t affected: only inputs
  • a

    ancient-vegetable-10556

    11/15/2022, 9:39 PM
    Right, so run 1 is running the process; run 2 is making the immutable thing; run 3 is just using the immutable thing
  • a

    ancient-vegetable-10556

    11/15/2022, 9:40 PM
    this is exactly what I’d expect from that
  • w

    witty-crayon-22786

    11/15/2022, 9:40 PM
    i’ll start a new thread
  • a

    ancient-vegetable-10556

    11/15/2022, 9:40 PM
    k
  • w

    witty-crayon-22786

    11/15/2022, 9:40 PM
    @ancient-vegetable-10556: re: 
    experimental_shell_command
    a
    b
    • 3
    • 27
  • c

    curved-television-6568

    11/16/2022, 2:07 PM
    Q: why is the reported terminal size off when running pytest with Pants? (bug or missing feature?)
    Copy code
    import shutil
print(shutil.get_terminal_size())
>> os.terminal_size(columns=80, lines=24)
    expected:
    Copy code
    In [2]: shutil.get_terminal_size()
Out[2]: os.terminal_size(columns=245, lines=63)
    ✅ 1
    • 1
    • 4
  • b

    bitter-ability-32190

    11/16/2022, 4:33 PM
    Oh btw @hundreds-father-404 re updating build files: It occurred to me since we really are in just normal rule code for 
    fix
    , its not hard to request 3rdparty libs and run a process to write. I'm thinking 
    libcst
    . which likely makes the override field a lot easier to handle
    ➕ 1
    h
    • 2
    • 3
  • b

    bitter-ability-32190

    11/16/2022, 7:06 PM
    I'm late to my own roadmap here, but: Progress: • Symlinks are in Pants now! There's a couple of friction points, but the ground is laid and they can be catptured/re-materialized • immutably symlinking "large" files/dirs is almost landed Plan: • https://github.com/pantsbuild/pants/issues/17372 • https://github.com/pantsbuild/pants/issues/17540 • I might finally try and make a generic copyright formatter
    🚀 1
1...877878879...906Latest