Miking Meeting Notes 2025

Meeting Notes 2025-02-20

Note-taker: Viktor Palmkvist

Participants: Tim Virgoulay, Lars Hummelgren, David Broman, Gizem Çaylak, Viktor Palmkvist, John Wikman, Oscar Eriksson, Didrik Munther, William Muth

Pull Requests

• William: https://github.com/miking-lang/miking/pull/905
  • We like. Also change the same file in miking-lang/miking-dppl.
• Marten: https://github.com/miking-lang/miking/pull/882
  • Now updated for the new build system, we like.
• Didrik: https://github.com/miking-lang/miking/pull/906
  • Make readBytesBuffered tail-recursive, then we like. (DONE)

Discussion

New script for testing multiple repositories at once seems promising, make it output the PR-groups to the log, then experiment with automatically committing and pushing to a repository when the run has completed.

Meeting Notes 2025-02-13

Note-taker: Oscar Eriksson

Participants: Tim Virgoulay, Lars Hummelgren, Viktor Palmkvist, John Wikman, Oscar Eriksson, David Broman, Gizem Çaylak

Pull Requests

• Gizem:
  • https://github.com/miking-lang/miking/pull/903
    • We like, merge.

Old PRs left to merge

• Miking: 876
• Miking-DPPL: 189

Discussion

We discussed microbenchmarks for external arrays. We should switch from Tensor to external array-based vectors and matrices, although this does not solve all performance problems. We need to look into how we generate OCaml code to be on par with OCaml. We will start by implementing a Bigarrays-based matrix library and move to using that library instead of tensors in coreppl/treeppl. We should also remove tensors from the documentation.

Meeting Notes 2025-02-06

Note-taker: John Wikman

Participants: Tim Virgoulay, Lars Hummelgren, Viktor Palmkvist, John Wikman, Oscar Eriksson, Anders Ågren Thuné, David Broman

Pull Requests

• Viktor:
  • https://github.com/miking-lang/miking/pull/897
    • We like
  • https://github.com/miking-lang/miking/pull/898
    • We like
  • https://github.com/miking-lang/miking/pull/899
    • We like
  • https://github.com/miking-lang/miking/pull/900
    • Add a named variable for the magic constant 4. Then we like.
  • https://github.com/miking-lang/miking/pull/902
    • We like, but there is a discussion to be had how we should treat Ref and Symbol compared to Int, Float, etc. Should we have a TySymbol AST node or should all be created as TyCon? Add it as a separate issue.
  • https://github.com/miking-lang/miking/pull/901
    • (Includes all the PRs above)
    • Now there is a generate-utest.mc, in addition to utest-generate.mc. Should aim to remove utest-generate.mc in the future when we can.
    • On a separate note, would also be good to have a test that checks that utest are actually running.
    • Viktor need to remove some debugging things, then we like.
  • https://github.com/miking-lang/miking-dppl/pull/192
    • We discussed this PR, but there are still things to be done. Will present again at a later date when it is ready.

Meeting Notes 2025-01-23

Note-taker: Viktor Palmkvist

Participants: Lars Hummelgren, Oscar Eriksson, Viktor Palmkvist, Tim Virgoulay, John Wikman, Gizem Çaylak, David Broman, Anders Ågren Thuné

Pull Requests

• Tim: https://github.com/miking-lang/miking-dppl/pull/191
  • Implement fallback to prior distributino if no explicit drift-kernel is supported. Add a test that is fast and included in normal test suite, but is much less precise. Also fix Oscar's comment. Then we like.
• Oscar: https://github.com/miking-lang/miking/pull/896
  • We like

Discussions Proposal: remove tensors as a built-in part of the language. Refactor or remove dependent code, primarily shifting to use externals that implement vectors and matrixes directly. Oscar can write that library.

Meeting Notes 2025-01-16

Note-taker: Lars Hummelgren

Participants: Lars Hummelgren, Oscar Eriksson, Viktor Palmkvist, Anders Ågren Thuné, Tim Virgoulay, John Wikman, Gizem Çaylak

Pull Requests

• Viktor
  • https://github.com/miking-lang/miking/pull/889
    • Wait until we know how it conflicts with PR 882
  • https://github.com/miking-lang/miking/pull/891
    • Add at least one test for avlIntersectWithKey, then we like
  • https://github.com/miking-lang/miking/pull/894
    • Looks good, but we would like to know about the performance diff
  • https://github.com/miking-lang/miking/pull/888
  • https://github.com/miking-lang/miking/pull/890
  • https://github.com/miking-lang/miking/pull/892
  • https://github.com/miking-lang/miking/pull/893
  • https://github.com/miking-lang/miking/pull/895
    • We like, merge

Discussions

• https://github.com/miking-lang/miking/pull/876 and the corresponding one in DPPL https://github.com/miking-lang/miking-dppl/pull/189 are ready to be merged (they were presented in November, though the DPPL one is not in the notes).