| [Coke] | can someone bump to get the mimalloc changes for next blin? | 00:02 | |
| oh, that was just a PR. | 01:35 | ||
| Geth | nqp/main: ed3e57ba17 | (Will Coleda)++ | tools/templates/MOAR_REVISION Bump MoarVM - Bump mimalloc to v2.4.1 |
01:50 | |
| rakudo/main: 73a74b5f86 | (Will Coleda)++ | tools/templates/NQP_REVISION Bump NQP - Bump mimalloc to v2.4.1 |
01:52 | ||
| rakudo: ugexe++ created pull request #6454: RakuAST: emit callstatic for named sub calls with bound-once callees |
03:44 | ||
| rakudo: ugexe++ created pull request #6455: RakuAST: reduce a smartmatch against a type object to a type check |
04:35 | ||
|
05:23
[Coke] left
05:28
[Coke] joined
05:37
hurufu joined
06:26
finanalyst joined
08:14
finanalyst left
08:23
finanalyst joined
10:07
finanalyst left
|
|||
| ugexe | anyone wanna make a new release of github.com/raku-community-modules/...pers-Blob/ ? the rakuast fix PR was merged. i'd try to do it but i'm not familiar with the .META/SOURCE.cdn.json stuff | 17:31 | |
| [Coke] | failed its windows test? | 18:02 | |
| japhb | [Coke]: Has been failing its windows test the whole time. :-( | 18:46 | |
| lizmat | I'dd do a release | 18:51 | |
| *'ll | |||
| ugexe | lizmat++ | 18:53 | |
| Geth | rakudo/main: 9e2652bf15 | (Nick Logan)++ (committed using GitHub Web editor) | 3 files RakuAST: reduce a smartmatch against a type object to a type check (#6455) Previously a smartmatch always compiled to a call of the setting's infix:<~~>, which dispatches ACCEPTS and boolifies its result. When the matcher is a compile-time known type object the legacy optimizer instead emits a plain nqp::istype of the topic, which answers identically: istype runs a subset's refinement and honors definite and ... (16 more lines) |
18:55 | |
| rakudo/main: 0587f61be6 | (Nick Logan)++ (committed using GitHub Web editor) | 2 files RakuAST: literalize an unhandled node to the sentinel, not a die (#6452) Previously the catch-all literalize candidate died with "literalize on <type> NYI". That die is not a CannotLiteralize, so the proto's CATCH could not turn it into the sentinel and it leaked out. Building an indirect package name from a runtime expression then reported "literalize on RakuAST::Ternary NYI" instead of that the name is not compile-time ... (6 more lines) |
|||
| rakudo/main: 54bd488246 | (Nick Logan)++ (committed using GitHub Web editor) | 2 files RakuAST: dispatch native-typed constants as native (#6450) Previously a reference to a native-typed constant compiled to a lexical lookup of the boxed value, so `my int constant PICK = 0; foo PICK` passed an Int and could never select a `foo(int)` candidate. A native multi candidate is only chosen when the argument is intrinsically native at the call site, the way a `my int $x` read is; a boxed lookup never qualifies. ... (9 more lines) |
|||
| lizmat | ARGH, looks like the cert on 42.zef.pm has expired | ||
| lizmat | ugexe ^^ | ||
| rakudo/main: 2e49395508 | (Nick Logan)++ | 2 files Make Format.new record its directives reliably Previously Format.new read the @*DIRECTIVES dynamic after calling Formatter.new, but that dynamic only ever existed inside Formatter.AST, so the attribute always received a Failure. That broke Format.directives, and constructing a Format during a compilation made the instance unserializable: the Failure backtrace references ... (7 more lines) |
18:57 | ||
| rakudo/main: 2ed2d94c17 | (Nick Logan)++ | 3 files RakuAST: build constant format string literals at compile time A constant q:o literal previously emitted a runtime Format.new call, paying the Formatter parse and EVAL on first use in every process, with a comment disabling the compile-time construction because the resulting objects misbehaved. Two problems stood in its way: the Failure that Format.new bound as its directives made the instance ... (9 more lines) |
|||
| rakudo/main: 87da335a40 | (Nick Logan)++ (committed using GitHub Web editor) | 5 files Merge pull request #6447 from ugexe/ugexe/rakuast-format-literal-compile-time RakuAST: build constant format string literals at compile time |
|||
| ugexe | tony will have to fix the cert :( | 18:59 | |
| lizmat | pinged him on #raku-infra | ||
| Geth | rakudo: ugexe++ created pull request #6456: RakuAST: give a stash reference its stash as compile-time value |
19:17 | |
| [Coke] | sounds like it was fixed, based on infra chat | 19:18 | |
|
19:34
japhb left
19:35
japhb joined
|
|||
| lizmat | yeah: raku.land/zef:raku-community-modul...s?v=0.1.13 | 20:36 | |
| Geth | rakudo: ugexe++ created pull request #6457: RakuAST: rewrite a square by the power operator to a multiply |
21:32 | |
| rakudo/main: 52331171d9 | (Nick Logan)++ (committed using GitHub Web editor) | 2 files RakuAST: give a stash reference its stash as compile-time value (#6456) Previously a name with a trailing :: claimed the resolved package as its compile-time value, where evaluating the name produces the .WHO of that package. A consumer of the compile-time value then saw the wrong object: my constant S = Foo::Bar:: bound S to the package where the legacy frontend binds the stash, and a compile-time consumer like the ... (5 more lines) |
21:39 | ||
|
22:27
hurufu left
|
|||
| [Coke] | Having a announcement that just dumps the changelog seems less useful each release. | 22:28 | |
| esp. since we have 202 commits so far for this release. | 22:29 | ||
| japhb | "A lot of stuff happened. Whoo-boy, it sure was a lot." :-P | 22:33 | |
|
22:35
apogee_ntv left
|
|||
| ugexe | yeah i think something like "RakuAST: various bugfixes and optimizations towards parity with the legacy frontend" would cover most of it | 22:40 | |
| the specific bugfixes probably aren't relevant to most people since generally speaking no one is using rakuast yet | 22:41 | ||
| [Coke] | I will do that. I don't think the release announcement needs even the shas | ||
| ugexe | but people are probably interested in progress being made in rakuast, so a few line items outlining that is still good | ||
| [Coke] | The other sections we can call out the individual items. | ||
| ugexe | yeah | 22:42 | |
| [Coke] | Yup. | ||
| Cool. | |||
| Geth | rakudo/main: 228e029c96 | (Will Coleda)++ | docs/release_guide.pod Add note about cleaning up release announcement |
22:46 | |
| [Coke] | releasable6: next | ||
| releasable6 | [Coke], Next release in ≈7 days and ≈20 hours. There are no known blockers. 122 out of 202 commits logged | ||
| [Coke], Details: gist.github.com/5deac54b80126d2efd...04e2cabe28 | |||