[Coke] release tomorrow still has a blocker. the mcp failure on rakudo head has a moarvm unwind error 00:15
releasable6: next
releasable6 [Coke], Next release in ≈18 hours. 1 blocker. 102 out of 179 commits logged
[Coke], Details: gist.github.com/76735cdf238398915a...78893a55b1
[Coke] linked blocker ticket in there. 00:16
git llog 00:23
oops
can someone test Text::Markdown::Discount with latest rakudo? 00:38
I am having trouble installing the markdown lib on my personal box. 00:42
ugexe mcp might just be flakey though 00:43
i installed it fine
[Coke] ok.
ugexe i assume Text::Markdown::Discount is also flakey 00:49
timo can't repro the mcp one after running it a few times in a row
ugexe the tests fail for me regardless of raku version im using, seemingly because the lib markdown library im using is returning e.g. < and > but the test expects &lt; 00:50
timo and i don't want to keep looking at this slop 00:53
03:27 donaldh left 04:31 ShimmerFairy left 04:40 ShimmerFairy joined
[Coke] firsty pass changelog done 04:42
cleanup, reorg welcome
releasable6: next
releasable6 [Coke], Next release in ≈14 hours. 1 blocker. 177 out of 179 commits logged
[Coke], Details: gist.github.com/29f977a463014a0fdf...751caa7c43
[Coke] releasable6: next 04:43
releasable6 [Coke], Next release in ≈14 hours. 1 blocker. 179 out of 179 commits logged
[Coke], Details: gist.github.com/ffaf901bf8c0d1d19b...af18cd3f20
Geth rakudo: ugexe++ created pull request #6329:
RakuAST: Take the value of a standalone smiley-keyed colonpair		 05:07
05:26 hurufu joined 05:35 donaldh joined 13:08 librasteve_ joined 13:34 guifa_ joined
Geth rakudo: ugexe++ created pull request #6330:
RakuAST: Allow adverbs on match-capture variables		 15:25
15:53 guifa_ left
Geth rakudo: ugexe++ created pull request #6331:
RakuAST: Accept the macron negative sign in a superscript power exponent		 15:55
[Coke] firing up the release vm... 15:56
15:59 guifa_ joined
[Coke] releasable6: next 16:03
releasable6 [Coke], Next release in ≈2 hours. There are no known blockers. 179 out of 179 commits logged
Geth rakudo/release-2026.06: 2d532627ab | (Will Coleda)++ | 3 files
Update changelog + announcement

Deliberately not logged:
  [e88c5f6e][fb9cf62d][bd750d4f][b9fa3b98][1c586eb2][ee30c34f]		 16:12
16:15 hurufu left
Geth rakudo: ugexe++ created pull request #6332:
RakuAST: Resolve an indirect name with a non-literal expression at runtime		 16:46
[Coke] MoarVM release done. 17:08
timo [Coke]++ 17:17
improving the moar gdb plugin right now 17:20
17:32 librasteve_ left
[Coke] rakudo stress test fails. 17:44
t/spec/S02-types/bag.t (Wstat: 256 (exited 1) Tests: 257 Failed: 1) Failed test: 115 17:45
Will check 6c and 6d
6.c: 17:51
committable6 [Coke], I cannot recognize this command. See wiki for some examples: github.com/Raku/whateverable/wiki/Committable
[Coke] t/spec/S32-array/shift.t (Wstat: 256 (exited 1) Tests: 33 Failed: 1) Failed test: 32 Non-zero exit status: 1
that failure on 6.d also. Also also: 17:58
t/spec/S32-exceptions/misc.rakudo.moar (Wstat: 256 (exited 1) Tests: 37 Failed: 0)
ugexe "Fix Array.pop, .shift on holes in arrays" - maybe this is related to the shift.t failure 18:06
hard to say since none of these tests fail if i just run them
do you happen to have output from the failures? 18:07
[Coke] no, the ake process just shows the summary 18:10
yah, works fine on my mac also (master) 18:11
Might be a flapper. will retry on the release box
18:16 ShimmerFairy left 18:18 ShimmerFairy joined
[Coke] ok, flapper. One of the things I did was remove flapper protection on the release process. Opening a ticket for that test. 18:21
again, tap rant: finding "test 115" in a file with 300 tests is a PITA. 18:26
github.com/Raku/roast/issues/892 18:32
timo yes, it really is 18:33
18:35 guifa_ left
[Coke] I picked a random one, made it fail, got that number, and then tried to bisect it. :) 18:35
18:36 librasteve_ joined, guifa_ joined
ugexe sometimes you get lucky and you can just run the test, copy the test description from the numbered test output you want, then grep for that 18:38
[Coke] Have to add "run the spec tests" as a blin-time activity 18:39
timo good idea 18:40
though ideally CI would make that not necessary ... :| 18:41
18:41 guifa_ left
[Coke] Yes, please. 18:43
18:46 ShimmerFairy left
Geth roast/6.c-errata: 3a439ff89b | (Elizabeth Mattijsen)++ (committed by Will Coleda) | S32-array/shift.t
Fix some .shift tests

1. github.com/rakudo/rakudo/pull/6242 changed the behaviour
   of .shift on a hole to returning the default value of the array
   rather then Nil (also what one would get if one would do @a[0]
   on a hole)
2. Activate test of shifting off an infinite array: that apparently
   works nowadays :-)		 18:50
18:51 ShimmerFairy joined
Geth roast/6.d-errata: 10be28a38f | (Elizabeth Mattijsen)++ (committed by Will Coleda) | S32-array/shift.t
Fix some .shift tests

1. github.com/rakudo/rakudo/pull/6242 changed the behaviour
   of .shift on a hole to returning the default value of the array
   rather then Nil (also what one would get if one would do @a[0]
   on a hole)
2. Activate test of shifting off an infinite array: that apparently
   works nowadays :-)		 18:51
[Coke] the exception test in 6.d is a flapper. 19:01
moving on 19:06
19:07 MasterDuke joined
Geth nqp/main: cbb14c423c | (Will Coleda)++ | tools/templates/MOAR_REVISION
[release] Bump MoarVM revision to 2026.06
 19:24
nqp/main: f2092f96c8 | (Will Coleda)++ | VERSION
[release] Bump VERSION to 2026.06
rakudo/release-2026.06: 75c66c5422 | (Will Coleda)++ | tools/templates/NQP_REVISION
[release] Bump NQP revision to 2026.06
rakudo/release-2026.06: 878e8b8d17 | (Will Coleda)++ | VERSION
[release] Bump VERSION to 2026.06
rakudo: coke++ created pull request #6333:
Release 2026.06		 19:28
rakudo/main: 4 commits pushed by (Will Coleda)++ 19:32
lizmat whee! 19:33
[Coke]++
19:36 guifa_ joined
[Coke] .tell patrickb the source release for 2026.06 is done. 19:37
tellable6 [Coke], I'll pass your message to patrickb
[Coke] .tell el_che the source release for 2026.06 is done.
tellable6 [Coke], I'll pass your message to El_Che
[Coke] I *think* I'm done 19:48
Let's do another blin-ast run next weekend. 19:50
(sooner is fine)
ugexe i'd suggest doing it whenever we get most of the existing ~18 rakuast prs merged. i've been trying to find more stuff to fix from the list, but its so stale at this point that its hard to triage much further 19:52
[Coke] OK. I can kick it off asap, just let me know and we'll get the next round done. 19:53
Geth rakudo: ugexe++ created pull request #6334:
RakuAST: Remove the grammar sanity check		 20:07
21:06 guifa_ left
Geth rakudo/main: 2dba534b41 | (Nick Logan)++ (committed using GitHub Web editor) | 3 files
RakuAST: Remove the grammar sanity check

A compile-time check (added in 7a61872f7) rejected a grammar containing a
  `<reference>` to a rule it could not find. In practice it has too many false
positives: it does not account for rules reachable through inheritance, imported tokens, proto/multi regexes, runtime mixins, or references in branches that are never taken, so it rejects grammars that compile and run fine on the ... (8 more lines)		 21:09
21:12 guifa_ joined 21:25 MasterDuke left
patrickb binaries are out. 21:47
tellable6 2026-06-27T19:37:02Z #raku-dev <[Coke]> patrickb the source release for 2026.06 is done.
21:52 guifa_ left
Geth rakudo/main: 3357843820 | (Nick Logan)++ (committed using GitHub Web editor) | 2 files
RakuAST: Compile a package-qualified compile-time variable (#6292)

  `$?CALLER::PACKAGE` and other `?`-twigil variables with a `::`-qualified
name threw "Variable '...' is not declared" at compile time. The action treated the whole qualified name as a flat compiler-variable lookup rather than a `$?<name>` access through the package, so the lexical resolve failed. This broke installing Green.
Route a qualified `?`-twigil variable through Var::Package with the twigil folded into the sigil, so it resolves `$?<name>` through the package, the same as the explicit CALLER::<$?PACKAGE> form.		 21:56
21:57 guifa_ joined
Geth rakudo/main: 2dcce2d97b | (Nick Logan)++ | 2 files
RakuAST: let a role shadow a setting type of the same name

Declaring a role whose name matches a type from the setting died or misbehaved. The Hash::Ordered module's `role Hash::Ordered` (CORE has a
  `Hash::Ordered` class) died "Redeclaration of symbol 'Hash::Ordered'",
and `role Numeric` (CORE has a `Numeric` role) was added as a candidate to CORE's role group rather than replacing it. Installing a role looked ... (11 more lines)		 22:00
rakudo/main: 51435e9c38 | (Elizabeth Mattijsen)++ (committed using GitHub Web editor) | src/Raku/ast/package.rakumod
Refactor IMPL-SYMBOL-FROM-SETTING method logic

A ternary looks more appropriate than needing to have a return statement
rakudo/main: f5ab9b09c7 | (Nick Logan)++ (committed using GitHub Web editor) | 2 files
Merge pull request #6305 from ugexe/ugexe/rakuast-role-shadows-setting

RakuAST: let a role shadow a setting type of the same name
rakudo/main: 2bd7510489 | (Nick Logan)++ | 2 files
RakuAST: Allow a pseudo-package nested in a declared name

A pseudo-package such as CORE or MY is only illegal as the leading component of a declared name. contains-pseudo-package-illegal-for-declaration checked every part, so `Rakudo::CORE::META` was rejected with "Cannot use pseudo package CORE in package name" even though CORE there is an ordinary nested name. Check only the first part, matching the legacy frontend.		 22:01
rakudo/main: a142cee849 | (Nick Logan)++ (committed using GitHub Web editor) | 2 files
Merge pull request #6315 from ugexe/ugexe/rakuast-pseudo-package-nested

RakuAST: Allow a pseudo-package nested in a declared name
rakudo/main: d13257d2ed | (Nick Logan)++ | 2 files
RakuAST: Defer an unresolved qualified-method-call qualifier to runtime

A qualified method call such as `self.Some::Type::method` took the qualifier's type object from the implicit lookup's resolution at code-gen and died when that type was not resolvable in the current scope. A type that is only brought in by the consuming program was rejected at compile time even though the legacy frontend defers it. `Terminal::Widgets` hits ... (6 more lines)		 22:08
rakudo/main: b1cbbda2f4 | (Nick Logan)++ (committed using GitHub Web editor) | 2 files
Merge pull request #6319 from ugexe/ugexe/rakuast-qualified-method-call-unresolved

RakuAST: Defer an unresolved qualified-method-call qualifier to runtime
22:13 librasteve_ left
Geth rakudo/main: abc006c7f7 | (Nick Logan)++ | 2 files
RakuAST: Vivify an is-typed container before a BEGIN-time STORE

  `my %h is SomeType = ...` declares the container with a type-object default
and a separate vivification bound at block entry. When the declaration is evaluated in a BEGIN, that vivification has not run yet, so the initializing STORE was called on the type object. An immutable Associative such as Map or Set, whose STORE requires a defined invocant, then failed with ... (7 more lines)		 22:20
rakudo/main: 3b82dbb0ce | (Nick Logan)++ | 2 files
RakuAST: Identify the bare-create QuantHashes by type, not name

The vivification helper chose nqp::create over .new for an `is`-typed container by comparing the type's name string to Set/Bag/Mix, so a user type that merely shares one of those names was wrongly bare-created and its .new never ran. Resolve the setting's Set/Bag/Mix and compare by identity instead, matching the legacy frontend. ... (6 more lines)
rakudo/main: 0f5c806d6f | (Nick Logan)++ (committed using GitHub Web editor) | 3 files
Merge pull request #6320 from ugexe/ugexe/rakuast-begin-is-container-store

RakuAST: Vivify an is-typed container before a BEGIN-time STORE
rakudo/main: a565a02606 | (Nick Logan)++ (committed using GitHub Web editor) | 4 files
RakuAST: Treat :$foo in a use/import/no argument as an import selector (#6321)

  `use Foo :$bar` (and `import Foo :$bar`, `no Foo :$bar`) failed to compile
under RakuAST with "Variable '$bar' is not declared", because the `:$bar` colonpair was parsed as the live pair `bar => $bar` and the lexical `$bar` was demanded. In such an argument list `:$bar` names an import to select, the same as the tag `:bar`, so there is no lexical to read. ... (5 more lines)		 22:36
rakudo/main: 024861d95a | (Nick Logan)++ (committed using GitHub Web editor) | 4 files
RakuAST: Add the leading package of a qualified lookup to the SC (#6322)

A qualified term like `@Foo::Bar::baz` resolves its leading package and embeds it as a compile-time value. When that package is one vivified for the compiling unit's own name (`Foo` for a `unit class Foo::Bar`), it had no serialization context, so a BEGIN-time use of the term (for example in a `constant` initializer) died at code generation with "Object of type ... (5 more lines)		 22:48
rakudo/main: 95386937be | (Nick Logan)++ | 3 files
RakuAST: Allow anon subset and install no symbol for it

  `anon subset NumericDay of Int where * <= 31` died with "Cannot use
  'anon' with subset declaration". RakuAST::Type::Subset listed only `my`
and `our` as allowed scopes, while RakuAST::Type::Enum already allowed
  `anon`. Add `anon` to the subset's allowed scopes.
... (7 more lines)		 23:03
rakudo/main: 21977c0022 | (Nick Logan)++ | 4 files
RakuAST: Treat a pair with an anon-subset value as a hash

A `{ }` containing a single pair is a hash, unless the block declares a symbol. The block-or-hash disambiguation walked the pair for any declaration and treated every one as symbol-declaring, so a pair whose value was an `anon` declaration that installs nothing, such as
  `{ day => (anon subset Foo of Int where * <= 31) }`, was compiled as a
... (6 more lines)
rakudo/main: efd004cbd7 | (Nick Logan)++ (committed using GitHub Web editor) | 5 files
Merge pull request #6323 from ugexe/ugexe/rakuast-anon-subset

RakuAST: Support anon subset declarations
rakudo/main: 4f041a4c5a | (Nick Logan)++ (committed using GitHub Web editor) | 2 files
RakuAST: Give a non-constant lexical its BEGIN value in a constant (#6325)

A `constant` initializer is evaluated at BEGIN and compiled as a standalone block, so finalizing it inlines the compile-time values of the outer lexicals it references. An outer lexical with no compile-time value, such as a sigilless `my \x` whose binding runs later at mainline time, died with "Symbol 'x' does not have a compile-time value". ... (12 more lines)		 23:15
rakudo/main: 24b322c9f6 | (Nick Logan)++ (committed using GitHub Web editor) | 2 files
RakuAST: Accept the macron negative sign in a superscript power exponent (#6331)

A superscript power exponent may be written with the high minus `⁻` (U+207B) or the macron `¯` (U+00AF), and the grammar's super-sign token accepts both. But RakuAST::Postfix::Power.from-superscripts translated only `⁻` to `-`, so a macron-signed exponent like `10¯²⁴` reached `val("¯24")`, whose `.Numeric` is a Failure. That Failure became the power node's literal value, and serializing it ... (6 more lines)		 23:20
rakudo/main: 334f31c3aa | (Nick Logan)++ (committed using GitHub Web editor) | 3 files
RakuAST: Hyper the postfixes the legacy frontend hypers (#6326)

  `>>` on a postfix rejected several postfixes with "Cannot hyper this
postfix" that the legacy frontend accepts. Brought the set in line with legacy:
  - A meta-method call (`@a>>.^name`, `(Range)>>.^compose`) had no hyper
... (13 more lines)		 23:33
rakudo/main: eb9725dd1f | (Nick Logan)++ (committed using GitHub Web editor) | 2 files
RakuAST: Take the value of a standalone smiley-keyed colonpair (#6329)

A standalone colonpair whose key is a smiley dropped its value, so `:D(1)` parsed as the True-valued pair `:D` followed by a postfix call `(1)` on it, dying with "No such method 'CALL-ME' for invocant of type 'Pair'".
A smiley only means definedness after a type name, where the `(...)` is the ... (7 more lines)		 23:37
rakudo/main: 80ace78d5e | (Nick Logan)++ (committed using GitHub Web editor) | 4 files
RakuAST: Resolve forward references in a role body before composing it (#6327)

A role body is code-generated when the role is composed, which happens during the enclosing package declaration ahead of the compilation unit's check pass. A forward reference inside the body, such as a parameter or attribute default that names a `sub` declared later, resolves to nothing at parse time and relies on that later check pass to re-resolve it. ... (22 more lines)		 23:46
23:47 guifa_ left 23:51 guifa_ joined
Geth rakudo/main: f7635853f7 | (Nick Logan)++ (committed using GitHub Web editor) | 2 files
RakuAST: Allow adverbs on match-capture variables (#6330)

  `$0:exists`, `$<foo>:exists` (and `:delete`, `:k`, `:kv`, etc.) failed to
compile under RakuAST with "You can't adverb $0". A match-capture shorthand parses to RakuAST::Var::PositionalCapture / RakuAST::Var::NamedCapture, which lower to a `postcircumfix:<[ ]>` / `postcircumfix:<{ }>` call on `$/` but had no way to carry an adverb, so the postfix handler rejected the colonpair.
Give both nodes `add-colonpair` and push each adverb as a named argument on that call, the same as `$/[0]:exists` / `$/<foo>:exists` already does. So
  `$0:exists` becomes `postcircumfix:<[ ]>($/, 0, :exists)`.		 23:57

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!