[01:25] <Geth> ¦ rakudo/main: bd5bab3cfe | (Nick Logan)++ (committed using GitHub Web editor) | 4 files

[01:25] <Geth> ¦ rakudo/main: Clean up all generated build artifacts on make clean (#6575)

[01:25] <Geth> ¦ rakudo/main: 

[01:25] <Geth> ¦ rakudo/main: Previously make clean removed the bytecode under blib/Perl6 but left

[01:25] <Geth> ¦ rakudo/main: blib/Raku in place, so Grammar and Actions survived a clean and could

[01:25] <Geth> ¦ rakudo/main: be picked up by a later build. CLEANUPS_ALL now covers BLIB_RAKUDO_R

[01:25] <Geth> ¦ rakudo/main: alongside BLIB_RAKUDO, and the js backend cleanup list gains the

[01:25] <Geth> ¦ rakudo/main: matching source map entry.

[01:26] <Geth> ¦ rakudo/main: <…commit message has 5 more lines…>

[01:26] <Geth> ¦ rakudo/main: review: https://github.com/rakudo/rakudo/commit/bd5bab3cfe

[01:26] <Geth> ¦ rakudo/main: db40e8ca9d | (Nick Logan)++ (committed using GitHub Web editor) | 2 files

[01:26] <Geth> ¦ rakudo/main: Share the revision gate mixin across gated routines (#6573)

[01:26] <Geth> ¦ rakudo/main: 

[01:26] <Geth> ¦ rakudo/main: Previously every `is revision-gated` application parameterized the

[01:26] <Geth> ¦ rakudo/main: is-revision-gated role with a freshly computed Int. Parameterizations

[01:26] <Geth> ¦ rakudo/main: intern by argument identity, and during the setting build each site's

[01:26] <Geth> ¦ rakudo/main: revision Int is a distinct object, so every gated routine minted its own

[01:26] <Geth> ¦ rakudo/main: curried role, mixin type, and mixin cache type, all serialized into the

[01:26] <Geth> ¦ rakudo/main: <…commit message has 6 more lines…>

[01:26] <Geth> ¦ rakudo/main: review: https://github.com/rakudo/rakudo/commit/db40e8ca9d

[01:26] <Geth> ¦ rakudo/main: 7070da8300 | (Nick Logan)++ (committed using GitHub Web editor) | 4 files

[01:26] <Geth> ¦ rakudo/main: RakuAST: generate the POPULATE method from the build plan (#6559)

[01:26] <Geth> ¦ rakudo/main: 

[01:26] <Geth> ¦ rakudo/main: Previously only the legacy frontend implemented the metamodel's

[01:26] <Geth> ¦ rakudo/main: generate_buildplan_executor hook, so every class it compiled got a

[01:26] <Geth> ¦ rakudo/main: POPULATE submethod compiled from its build plan while a class the

[01:26] <Geth> ¦ rakudo/main: RakuAST frontend compiled fell back to Mu.POPULATE, which interprets

[01:26] <Geth> ¦ rakudo/main: the plan on every object construction. This implements the hook on

[01:26] <Geth> ¦ rakudo/main: <…commit message has 17 more lines…>

[01:26] <Geth> ¦ rakudo/main: review: https://github.com/rakudo/rakudo/commit/7070da8300

[01:27] <Geth> ¦ rakudo/main: d88b94ac58 | (Nick Logan)++ (committed using GitHub Web editor) | 2 files

[01:27] <Geth> ¦ rakudo/main: Cache interpolated regex compilations by pattern source (#6568)

[01:27] <Geth> ¦ rakudo/main: 

[01:27] <Geth> ¦ rakudo/main: Previously a string interpolated into a regex was compiled with EVAL

[01:27] <Geth> ¦ rakudo/main: and the resulting regex was cached by mixing a role into the string

[01:27] <Geth> ¦ rakudo/main: object itself. Code such as <{ %h.keys.join("|") }> produces a new

[01:27] <Geth> ¦ rakudo/main: string object on every match attempt, so every attempt compiled the

[01:27] <Geth> ¦ rakudo/main: pattern again. MoarVM never deallocates bytecode, so long running

[01:27] <Geth> ¦ rakudo/main: <…commit message has 18 more lines…>

[01:27] <Geth> ¦ rakudo/main: review: https://github.com/rakudo/rakudo/commit/d88b94ac58

[01:29] <Geth> ¦ rakudo/main: 3b6a4052a3 | (Nick Logan)++ | 2 files

[01:29] <Geth> ¦ rakudo/main: Keep Range integer bounds from leaking a Failure

[01:29] <Geth> ¦ rakudo/main: 

[01:29] <Geth> ¦ rakudo/main: Previously int-bounds decided whether the endpoints fit integer bounds by

[01:29] <Geth> ¦ rakudo/main: asking each of them for .Int and checking the result was an Int. An endpoint

[01:29] <Geth> ¦ rakudo/main: with no integer value fails there, and nothing marked that Failure handled,

[01:29] <Geth> ¦ rakudo/main: so the garbage collector reported it as an unhandled Failure in DESTROY.

[01:29] <Geth> ¦ rakudo/main: Summing an endless Range was enough to produce one.

[01:29] <Geth> ¦ rakudo/main: <…commit message has 10 more lines…>

[01:29] <Geth> ¦ rakudo/main: review: https://github.com/rakudo/rakudo/commit/3b6a4052a3

[01:29] <Geth> ¦ rakudo/main: ac2ce1c1ce | (Nick Logan)++ | 2 files

[01:29] <Geth> ¦ rakudo/main: Answer reductions that do not need to reach the end

[01:29] <Geth> ¦ rakudo/main: 

[01:29] <Geth> ¦ rakudo/main: Previously every reduction folded its values one at a time until they ran

[01:29] <Geth> ¦ rakudo/main: out, so a reduction over values that never run out never returned. Some of

[01:29] <Geth> ¦ rakudo/main: those reductions have their answer well before the end.

[01:29] <Geth> ¦ rakudo/main: 

[01:29] <Geth> ¦ rakudo/main: A short-circuit operator settles on a value that no further operand can

[01:29] <Geth> ¦ rakudo/main: <…commit message has 33 more lines…>

[01:30] <Geth> ¦ rakudo/main: review: https://github.com/rakudo/rakudo/commit/ac2ce1c1ce

[01:30] <Geth> ¦ rakudo/main: 9bbff6bde2 | (Nick Logan)++ (committed using GitHub Web editor) | 4 files

[01:30] <Geth> ¦ rakudo/main: Merge pull request #6576 from ugexe/ugexe/reduce-short-circuit-infinite

[01:30] <Geth> ¦ rakudo/main: 

[01:30] <Geth> ¦ rakudo/main: Answer reductions over endless sequences

[01:30] <Geth> ¦ rakudo/main: review: https://github.com/rakudo/rakudo/commit/9bbff6bde2

[01:30] <Geth> ¦ rakudo/main: ec663ef241 | (Nick Logan)++ (committed using GitHub Web editor) | 2 files

[01:30] <Geth> ¦ rakudo/main: Merge pull request #6572 from ugexe/ugexe/rakuast-name-stub-through-container

[01:30] <Geth> ¦ rakudo/main: 

[01:30] <Geth> ¦ rakudo/main: RakuAST: name stub packages without their container

[01:30] <Geth> ¦ rakudo/main: review: https://github.com/rakudo/rakudo/commit/ec663ef241

[01:59] *** guifa left
[02:09] *** guifa joined
[02:59] *** guifa left
[03:11] *** guifa joined
[03:22] *** apogee_ntv left
[03:23] *** apogee_ntv joined
[04:53] <Geth> ¦ rakudo: ugexe++ created pull request #6578: RakuAST: re-form BEGIN-compiled code with the optimize walk's marks

[04:53] <Geth> ¦ rakudo: review: https://github.com/rakudo/rakudo/pull/6578