01:32 guifa joined 01:33 [Coke] left
Geth rakudo: ugexe++ created pull request #6604:
RakuAST: optimize code compiled ahead of the unit
01:53
rakudo/main: 4 commits pushed by (Nick Logan)++ 02:49
nqp/main: 8b9fec97a8 | (Nick Logan)++ (committed using GitHub Web editor) | 2 files
Serialize precomputed grammar NFAs in their VM form (#862)

Previously a grammar's precomputed protoregex and alternation NFAs were serialized as QRegex::NFA objects still holding the full nested statelist they were built from, with each edge as three boxed ints in an NQPArray per state. The VM-level NFA object was only built on first use, and the statelist stayed reachable from the HOW cache forever. For a bare ... (9 more lines)
02:50
rakudo: ugexe++ created pull request #6605:
Bump NQP to get precomputed grammar NFAs serialized in their VM form
02:56
03:30 [Coke] joined
Geth rakudo: ugexe++ created pull request #6606:
RakuAST: resolve an operator again before folding or deciding its dispatch
03:39
06:00 finanalyst joined 06:39 finanalyst left, finanalyst joined
Geth rakudo/main: 37afd7796d | (Elizabeth Mattijsen)++ | tools/templates/NQP_REVISION
Bump NQP to get NFA serialization opt, ugexe++
08:24
09:20 lizmat left, Geth left, samebchase left, ilogger2 left, jdv left, jjatria left 09:21 Geth joined, jdv joined, jjatria joined, samebchase joined 09:22 lizmat joined 09:24 ilogger2 joined
lizmat . 09:59
Geth rakudo/main: df41eb1977 | (Elizabeth Mattijsen)++ | src/core.c/Deprecations.rakumod
Slightly change checking for deprecations in END processing

Instead of checking the env first, just try to create a report first. That fails quickly if there are no deprecations, and then doesn't need to check the environment variable.
Only disadvantage is that *if* there are deprecations *and* the ... (5 more lines)
10:40
10:49 lizmat left 10:50 lizmat joined
Geth rakudo/main: 60c19932d2 | (Elizabeth Mattijsen)++ | src/core.c/Raku.rakumod
Make Raku.legacy completely lazy

Which means one less call to &DYNAMIC at every startup
11:20
11:48 finanalyst left
Geth rakudo: ugexe++ created pull request #6607:
RakuAST: report worries once and render them readably during the setting build
15:10
15:28 djinni` left 15:32 djinni` joined
Geth rakudo/main: 5 commits pushed by (Nick Logan)++ 15:57
rakudo/main: 6727d6a97a | (Nick Logan)++ (committed using GitHub Web editor) | 5 files
RakuAST: resolve an operator again before folding or deciding its dispatch (#6606)

Previously the optimize walk trusted the resolution stored on an operator node when folding a constant expression or deciding a dispatch at compile time. That resolution is made when the operator is parsed, so a user operator declared later in the unit, which shadows the setting's, was never seen: `2 ** 3` folded to 8 with a user `**` ... (14 more lines)
rakudo/main: 8f994c6f6c | (Nick Logan)++ (committed using GitHub Web editor) | 0 files
Bump NQP to get precomputed grammar NFAs serialized in their VM form (#6605)
15:58
rakudo: ugexe++ created pull request #6608:
RakuAST: honour a user operator wherever the setting's was assumed
17:07
[Coke] There are a bunch of commits (e.g. 47db6ac5e0) that are in rakudo/main, but not showing in releaseable6:next OR the changelog draft. 17:47
Geth rakudo/main: 6 commits pushed by (Nick Logan)++, (Elizabeth Mattijsen)++ 18:16
ugexe [Coke]: does it look like it started happening after the nqp bump? i had a pr open to bump nqp last night, lizmat bumped it directly in main outside of that pr, and then i merged my pr this morning 18:35
i don't quite see how that would break anything, but its the only out-of-the-ordinary thing i notice 18:36
Geth rakudo/main: 5 commits pushed by (Nick Logan)++ 19:19
rakudo: ugexe++ created pull request #6609:
RakuAST: assign a native compound assignment's result without the metaop
19:59
rakudo/main: 6 commits pushed by (Nick Logan)++ 20:44
21:03 guifa left
ugexe m: my int $i = 1; say $i + ($i = 7) 21:32
camelia 8
ugexe m: my Int $i = 1; say $i + ($i = 7)
camelia 14
ugexe 😬 21:33
[Coke] releasable6: next 22:07
releasable6 [Coke], Next release in ≈27 days and ≈20 hours. There are no known blockers. 49 out of 49 commits logged
[Coke] ... looks like maybe it's just stuck, honestly. it's been saying 49 for hours
I'll try to restart the bots, maybe that'll help 22:08
22:09 quotable6 left, notable6 left, greppable6__ left, releasable6 left, tellable6 left, evalable6 left, sourceable6 left, committable6 left, bisectable6 left, huggable6 left, shareable6 left, linkable6 left, coverable6 left, bloatable6 left
[Coke] Didn't we have geth reporting on whateverable? 22:09
22:09 benchable6 left, nativecallable6 left, unicodable6 left
lizmat I may have cherrypicked the commit into main after I checked it out locally :-( felt like a good idea at the time 22:09
lizmat checks
nope, no reporting on whateverable... shall I? 22:10
webhook added 22:11
22:11 unicodable6 joined 22:12 huggable6 joined, coverable6 joined, tellable6 joined, shareable6 joined, greppable6 joined 22:13 bloatable6 joined, evalable6 joined, sourceable6 joined, benchable6 joined, committable6 joined, notable6 joined, bisectable6 joined, nativecallable6 joined, quotable6 joined, releasable6 joined 22:14 linkable6 joined