Chapel project meeting: 2026-01-06 (agenda items, meeting link, notes)

[github-actions[bot]]

This discussion is for the weekly Chapel project meeting.

Date/Time: 2026-01-06 at 10am PT
Join the meeting using this Teams link

---

Prior to the meeting, propose any agenda items using comments below. After the meeting, rough notes will be added in comments.

Typically, these meetings consist of:

• News: Quick updates or announcements to keep people in the loop on key efforts or topics
• Demos: Have something cool to show off? Do it in a ~5-minute demo
• Short Topics: Have something that needs a quick discussion, description, or straw poll? Propose a ~5-minute discussion topic
• Longer Topics: If you have a topic needing more time, propose it with an estimated duration and we'll try to fit it in.

If you'd like to propose an item in any of these categories, please reply to this discussion.

In addition to the above sections, we have some standing agenda items, such as:

• Triage: What is the state of testing and are there issues that need more attention?
• User Issues: What new user issues have come up in the past week, and which should we prioritize?
• Q&A: Have something you want to ask, but not on the clock? We'll open the floor to Q&A as the official meeting wraps up.

[DanilaFe]

Copying from the previous thread, which stood for a cancelled meeting.

I'd like to talk about three language semantic questions, or however many we have time for.

• Do we allow x && y for x, y non-bool? (eg, x : owned C?, x : int?)
• Should we allow distinct child classes (eg different subclasses of Error) to be returned from a function with an inferred return type?
• Is the current interplay between forwarding and POI correct?

[jabraham17]

I'd like to discuss the Chapel project meeting at HPSFcon

[lydia-duncan]

Here are today's user issues slides

[bradcray]

Agenda / Notes

News/Announcements

• Happy 2026!

• Chapel 2.7 was released on Dec 18

  • thanks to all who contributed!
  • announcement: https://chapel-lang.org/blog/posts/announcing-chapel-2.7/
  • Spack formula is still waiting for a merge:
    • chapel: add v2.7 release candidate spack/spack-packages#2749

• HPSFCon: March 16-20

  • talk proposals are due Jan 11:
    • https://events.linuxfoundation.org/hpsf-conference/
  • we also submitted a half-day Chapel session, which was accepted for Mar 18
    • format TBD (see discussion later in this meeting)
    • guide to what we're planning due Jan 16

Triage Status

• Engin: end-of-rotation update
  • Anna agreed to set up new triage rotation
    • Daniel volunteers to do triage if needed
  • during break, turned testing off
  • some things still being investigated after shutdown:
    • Docker struggling to fetch repo (maybe Jenkins token issue?)
    • build-chapel-packages failing more frequently with OOM than before
      • Brad: could adding new configs into packages cause this?
        • Jade: more memory, but only mildly so...
      • Jade: thinks clearing the Docker cache may fix?
      • Anna: could also be due to OS upgrade on Mac
        • updated some memory setting, but maybe not enough? lost prev value
      • Anna and Jade will own
    • IB system missing some common directories after break (Engin)
  • not related to shutdown:
    • test that's sensitive to internal module line numbers (Jade/Daniel)
  • day13-error needs unfuturizing (Daniel)
  • core dump in linux64-clang in release announcement test (Ben)
  • longstanding sporadic nqueens (Engin)
    • planning to skipif when oversubscribed, which seems to be the factor
  • longstanding ofi timeout (Brad)
  • chapcs11 seems to be "down" in slurm / jobs not progressing (Engin)
    • Engin: chapelu should be able to put it into a drain state
  • CHAMPS false positive in Jenkins (Jade)
    • waiting for Apollo system to be in good state again
  • new sporadic in a debugger memory leak test (Jade)
    • planning to rewrite to make it less sporadic

User Issues

• Lydia: weekly report (see slide in GitHub discussion for links)
  • 6 new user issues:
    • fileWriter equality check (Nelson)
    • support for combining short-form arguments (e.g., -aux == -a -u -x)
    • clang warning, maybe Nix-specific (Tom)
    • question about removing libstdc++ dependency (Tom)
    • transmute issues w.r.t. docs, stabilizing (Lydia, on Damian's behalf)
  • 1 resolved user issue:
    • more information in user error message (Lydia)
  • ongoing:
    • jmag still working on isClose() for UnitTest
  • old issues
    • new comment on GPU support for Apple GPUs
      • Jade: Feels like we're lacking a reasonable approach
    • Thitrin's issue about perf -- split remaining item into own issue?

Demos

• Brad: fileWriter/Reader equality bug
  • [Bug]: fileWriter argument checks false against stdout, when it should be true #28247

Discussion Topics

• Daniel: Should we support x && y for non-bool x/y expressions?
  • issue: [Bug]: logical operators do not work on nilable classes #25382
  • e.g., if x is owned C? or int
  • DavidL noticed that this didn't work for classes even though 'if x' works
    • Daniel: similarly for other non-bool expressions
  • initial Slack discussion was "maybe this was just an oversight
  • historically, may have intentionally not allowed this to avoid confusion
    with bitwise & for integers, say
  • Engin: What would x be in 'var x = y && z;'?
    • Daniel: bool
  • Engin: Can user define how to bool-ize their own type, as with class?
    • Daniel: Yes, but not yet user-facing
  • Brad: Generally likes this, particularly if there's precedent; believes
    we got where we are out of lots of conservatism earlier in the project
    and doesn't think we should feel beholden to it.
    • Daniel: Was stating Cons more based on own feelings than history
  • Jade: Pretty in favor of changing -- we currently feel weirdly asymmetrical
  • Engin: Neutral because he doesn't think we should support 'if 1 { }' either
    • is this about "types being coercible to bools?" or different?
    • e.g., if I have 'foo(x: bool)', can I call it with 'foo(myInt)'?
    • Daniel: thinks there is precedent (in C++, e.g.) thinks there's a
      distinct cast/coercion vs. conditional check...
  • Brad: Clarifying, does C++ use the same, or different mechanism to
    coerce ints to bools in general vs. in conditional expressions?
    • Daniel, thought different, but not confident; will check
    • offline Q: Seeming problem: If && takes bools and we rely on coercion
      to send non-bools to it...
    • offline A:
  • Anna: Doesn't believe we necessarily ought to allow coercion of
    arbitrary types to bools in arbitrary places just because we
    permit non-bool types to be treated as bools in conditionals.
    E.g., supporting 'if x' or 'if y' doesn't mean we must support 'if
    x && y' nor coercions of x and y to bool in other locations.
    • Brad: Realizes that if we think of operator &&(x: bool, y:bool) then how do we support it without also permitting
      coercions to bools in arbitrary places? Seems like a slippery
      slope.
    • Daniel (post-meeting): If we consider && and || not to be
      operators taking arguments but special expressions (which
      arguably they already are due to the short-circuiting "don't
      evaluate RHS until/unless needed" behavior), then this seems
      like less of an issue — like conditionals, they can be treated
      specially.

• Daniel/Anna: Brief discussion about support for the pattern
  if x && x!.foo() then ...
• Action: Daniel will do a bit more checking into what C++ and
  potentially other key languages do and capture notes, and
  maybe a proposal for what Chapel should do, on the issue.

• Jade: HPSFCon and Chapel project meeting

  • submitted project meeting that was accepted
  • now they want to know what content we'll have in our meeting
    • specifically, talks? Or something more unstructured?
  • Brad: in HPE outreach/community meeting, discussed K.I.S.S. approach:
    • give a scheduled intro talk of 30-45 minutes for people who want 101
    • make the rest of the session demos, office hours, drop by, etc.
    • encourage Chapel talks to the main conference for bigger audience and
      to avoid coopting ChapelCon
  • Jade: Sounds good but maybe different than what HPSFCon organizers expect
    • Brad: TAC members signed off on this idea, so suspect that
      there's just a disconnect between HPSF TAC (who matter) and
      Linux Foundation staff who are used to setting up conferences
      with multiple talk tracks. Can help with that messaging to
      HPSF/LF if useful.

• Daniel: Should an inferred-return-type proc be able to return distinct
  child classes?

  • e.g.,

    class C1: Parent {}
    class C2: Parent {}
    
    proc foo() {
      if coinflip then return new C1(); else return new C2;
    }
    

  • people generally in favor of proceeding
  • will need to determine what to do in cases like:
    • one branch shared, the other owned
    • one branch nilable, the other not
    • both of the above
    • classes have no parent class (other than the built-in root class of all classes)

Didn't get to this, will take it up in the future

• Daniel: Is the current interplay between forwarding and POI correct?

Post-meeting

• Jade/Daniel/Brad discussed 'x?.bla()' a bit more
  • what it means in other languages, why it may not be tractable in Chapel
    • at least, without adding union types, nilable types, result types, etc.

Office Hours / Q&A

TODOs / Open Issues

• Brad check on travel vs. submission auth w.r.t. HPSFCon