29 Pull requests merged by 11 people

• Remove incorrect debug-only accounting code (and rename a function)

  #3676 merged Mar 11, 2025

• Fix: compare peek and poke kinds

  #3674 merged Mar 11, 2025

• Restore the -no-mach-ir command-line switch

  #3675 merged Mar 11, 2025

• Keep track of relevant axes when descending in normalize

  #3670 merged Mar 10, 2025

• Remember to think about optimizing jkind_subst

  #3673 merged Mar 10, 2025

• Drop jkind_of_first_type

  #3659 merged Mar 10, 2025

• Infer with-bounds for tuples

  #3664 merged Mar 10, 2025

• Add some CRs for with-bounds (layouts v2.8) things we want to do later

  #3660 merged Mar 10, 2025

• Easy compare script

  #3571 merged Mar 10, 2025

• Remove the Unnecessary_allow_any_kind warning

  #3576 merged Mar 10, 2025

• Delete Mach

  #3287 merged Mar 10, 2025

• Automatic reordering of record fields to satisfy the mixed block restriction

  #3581 merged Mar 7, 2025

• Shape UIDs for parameter modules

  #3668 merged Mar 7, 2025

• Avoid polymorphic comparison in backend/arm64

  #3645 merged Mar 6, 2025

• Specialize comparisons in presence of immediate_or_null correctly.

  #3653 merged Mar 6, 2025

• Immediate_or_null and with-bounds for or_null

  #3655 merged Mar 6, 2025

• Bump magic numbers for 5.2.0minus-8

  #3666 merged Mar 6, 2025

• Locals x Effects II

  #3648 merged Mar 6, 2025

• Avoid polymorphic comparison in backend/{asm_targets,debug,internal_assembler}

  #3646 merged Mar 6, 2025

• Print the offending mod-bounds on mismatch for unsafe crossing

  #3661 merged Mar 6, 2025

• Move unboxed product arrays to stable

  #3665 merged Mar 6, 2025

• Linked unboxed-boxed versions of types via # paths

  #3526 merged Mar 5, 2025

• Enable more warnings under backend/regalloc

  #3658 merged Mar 5, 2025

• Refactor cmm functions to be more generic over operand width

  #3404 merged Mar 5, 2025

• Optimize and improve debugging for with-kinds

  #3630 merged Mar 5, 2025

• Remove -infer-with-bounds flag, defaulting to true

  #3657 merged Mar 5, 2025

• Ensure root safety in caml_register_named_value.

  #3656 merged Mar 5, 2025

• Avoid polymorphic comparison in backend/amd64

  #3643 merged Mar 5, 2025

• Remove dependency of trie.ml and leapfrog.ml on Named_ref.

  #3650 merged Mar 5, 2025

11 Pull requests opened by 5 people

• Remove as many magics and unsafe crossings as possible

  #3662 opened Mar 5, 2025

• Fix untypeast/pprintast bug (backport upstream 13845)

  #3663 opened Mar 5, 2025

• Improve formatting in Asmgen

  #3667 opened Mar 6, 2025

• optimize cmm shifts and tags

  #3669 opened Mar 7, 2025

• Fix mode crossing inference for `[@@unboxed]` types

  #3671 opened Mar 7, 2025

• Resurrect GC parameter caml_major_heap_increment

  #3672 opened Mar 10, 2025

• Add runtime events to record GC pacing computations.

  #3677 opened Mar 11, 2025

• Remove obsolete uses of dom_st->allocated_words_direct.

  #3678 opened Mar 11, 2025

• Ignore some common files

  #3679 opened Mar 11, 2025

• optimized bounds-checks to eliminate some shifts

  #3680 opened Mar 11, 2025

• Extend the Xgc_pacing_policy tweak to allow a "full 2025" option.

  #3681 opened Mar 11, 2025

22 Unresolved conversations

Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.

• Added Cmm_helpers.Scalar_type

  #3423 commented on Mar 10, 2025 • 31 new comments

• Implicit unboxed records

  #3527 commented on Mar 11, 2025 • 13 new comments

• Array reinterpret operations

  #3366 commented on Mar 7, 2025 • 10 new comments

• flambda2-types: New n-way join algorithm

  #3538 commented on Mar 11, 2025 • 10 new comments

• Avoid polymorphic comparison in `backend`

  #3649 commented on Mar 6, 2025 • 8 new comments

• Unique continuations

  #3542 commented on Mar 6, 2025 • 8 new comments

• Portable continuations

  #3541 commented on Mar 10, 2025 • 6 new comments

• Add `Mode.Crossing`

  #3639 commented on Mar 7, 2025 • 2 new comments

• `local` implies `yielding`

  #3532 commented on Mar 5, 2025 • 1 new comment

• [DRAFT] Continuation specialization, aka match-in-match

  #3501 commented on Mar 5, 2025 • 0 new comments

• Always use a macro when checking for a promoted header

  #3493 commented on Mar 11, 2025 • 0 new comments

• Add primitive types for unboxed small integers

  #3370 commented on Mar 10, 2025 • 0 new comments

• Async exns x effects

  #2455 commented on Mar 7, 2025 • 0 new comments

• Improved compilation of match-in-match

  #926 commented on Mar 5, 2025 • 0 new comments

• Reduce default space overhead from 160 to 120

  #3543 commented on Mar 11, 2025 • 0 new comments

• Runtime events should never be conditional on debug log level

  #3585 commented on Mar 11, 2025 • 0 new comments

• Generally better runtime logging

  #3587 commented on Mar 11, 2025 • 0 new comments

• More `or_null` tests

  #3633 commented on Mar 5, 2025 • 0 new comments

• Some `value_or_null` support in Stdlib

  #3636 commented on Mar 10, 2025 • 0 new comments

• [Ignore] Linked unboxed-boxed rebased on 5.2.0minus-6

  #3637 commented on Mar 5, 2025 • 0 new comments

• Merge atomic counter inc/dec functions and use them consistently

  #3644 commented on Mar 11, 2025 • 0 new comments

• Add a peephole optimization engine in Flambda2

  #2188 commented on Mar 5, 2025 • 0 new comments