[Coke] We aren't really supporting the JVM backend at this point, correct? 00:09
It's basically listed as broken in every release announcement, and we don't run tests for it as part of the release.
We do have a half-hearted "It is also a good idea to look for JVM issues at this point." 00:10
Geth rakudo/coke/github_release: 632378cbe8 | (Will Coleda)++ | docs/release_guide.pod
Eliminate some markdown-isms.
 00:11
rakudo/coke/github_release: a69582ea70 | (Will Coleda)++ | docs/release_guide.pod
minor cleanups in formatting and phrasing
rakudo/coke/github_release: 74927a8dae | (Will Coleda)++ | 3 files
Rename .p6 script to avoid extension

cleanup release guide to track
00:12 librasteve_ left
timo2 now the gist has a raku port of the builtin NQP NFA optimizer 01:00
Geth rakudo/coke/github_release: c06cf642f2 | (Will Coleda)++ | tools/releasable/github-release
don't use Perl6 in useragent
 01:02
rakudo/coke/github_release: bf0488e244 | (Will Coleda)++ | tools/releasable/github-release
Specify headers from API docs
rakudo/coke/github_release: 147791df6e | (Will Coleda)++ | tools/releasable/github-release
Update auth header, add content-length header
rakudo: coke++ created pull request #5986:
Coke/GitHub release		 01:03
[Coke] Anyone has any suggestions on how to avoid the "bad content-length" error message there, greatly appreciated. 01:06
timo2 now my gist is not as slow any more 01:18
Geth rakudo/coke/github_release: 61429b0c46 | (Will Coleda)++ | tools/releasable/github-release
Pass in headers with correct named arg.
 01:25
rakudo/coke/github_release: 1741ccdf33 | (Will Coleda)++ | tools/releasable/github-release
Avoid keeping the old content length around
 01:30
[Coke] nevermind fixed it
Geth ¦ rakudo: coke self-assigned Remove or fix github_release.p6 github.com/rakudo/rakudo/issues/5862 01:33
rakudo/coke/github_release: 1e483e1f05 | (Will Coleda)++ | docs/release_guide.pod
Script now works

also, unify the release example number in one more place		 01:42
[Coke] Ah, finally.
That's only been bothering me since April 01:43
timo2 the raku port of the optimizer and the original optimizer have behaviour differences :( 02:29
i'm missing the crucial difference :| 04:06
05:12 kjp left 05:14 kjp joined 05:15 kjp left, kjp joined
patrickb The release binaries hit rakudo.org yesterday. 06:09
08:52 librasteve_ joined
ab5tract I think a JVM revisit would be a bit of fun, at least if it were in Kotlin 08:53
But I also have advocated for deprecating it from our outward facing materials, eg website, release announcements, etc 08:54
OTOH, my dispatcher patches were all ported to the base dispatcher, for the JVM's sake but also any future implementations too I guess 10:31
11:21 librasteve_ left
lizmat m: dd %%%%%%%%%%%%%%42 # hmmmm 12:16
camelia {}
lizmat bisectable6: old=2020.01 sdd 4.5 mod pi 12:26
bisectable6 lizmat, Bisecting by output (old=2020.01 new=cdde49e) because on both starting points the exit code is 1
lizmat grr
bisectable6: old=2020.01 dd 4.5 mod pi
bisectable6 lizmat, bisect log: gist.github.com/2cbd9e92f4615e60e7...772136c2ba 12:27
lizmat, (2024-11-22) github.com/rakudo/rakudo/commit/a9...b4dced35bc
lizmat, Bisecting by exit code (old=2020.01 new=cdde49e). Old exit code: 1
lizmat, bisect log: gist.github.com/a8ed7ab93bc7a93749...f284781e43 12:30
lizmat, (2024-02-04) github.com/rakudo/rakudo/commit/71...7429f88fb7
lizmat see also: github.com/Raku/problem-solving/issues/326 12:31
but that commit obliterated the difference between infix:<mod> and infix:<%>
librasteve ideas? 12:32
14:11 ds7823 joined
[Coke] what is lavm-perl6infra-1.atikon.io ? 14:18
(it's in the Akefile for the release)
Geth rakudo/coke/test-less: efba5598a9 | (Will Coleda)++ | tools/releasable/Akefile
Remove comments noting step numbers

These are not in sync, and the unicode numbering was cute but hard to maintain		 14:20
rakudo/coke/test-less: 393a34adf3 | (Will Coleda)++ | tools/releasable/Akefile
Don't run spectest on master twice during release
 14:34
14:38 librasteve_ joined
Geth rakudo: coke self-assigned TEST_JOBS ignored in akefile github.com/rakudo/rakudo/issues/5987
coke++ created pull request #5988: Coke/test less
 15:04
rakudo/coke/respect-threads: b1eca4bc40 | (Will Coleda)++ | tools/releasable/Akefile
Unify threads/test_jobs

Default to num cores -1 (previously hardcoded to 7). Now release managers can use TEST_JOBS=x to set both parallel testing and compiling with a single variable.
Resolves #5987		 16:20
rakudo: coke++ created pull request #5989:
Unify threads/test_jobs		 16:22
timo2 phew, the raku port of the nfa optimizer is now 1:1 with the nqp version 16:28
[Coke] neat 16:33
jdv: got a few PRs for you to sign off on when you can. I think they're good but will wait for your OK 16:42
16:48 librasteve_ left
Geth nqp/main: ea336280d1 | (Elizabeth Mattijsen)++ | tools/templates/MOAR_REVISION
Bump MoarVM to get rapidhash v3, MasterDuke++
 16:49
17:10 rakkable left, rakkable joined
jdv [Coke]: i think i did 17:15
Geth rakudo/main: 74bbe19029 | (Elizabeth Mattijsen)++ | tools/templates/NQP_REVISION
Bump NQP to get rapidhash v3, MasterDuke++
 17:34
17:48 bisectable6 left 17:51 bisectable6 joined
timo2 m: class Foo { has $.one; has $.two }; Foo.new(one => 1, two => 2;); 17:56
camelia Default constructor for 'Foo' only takes named arguments
in block <unit> at <tmp> line 1
timo2 funny situation i just found myself in due to copypasting some code and not throwing the ; out
18:15 librasteve_ joined
librasteve_ ab5tract: I think that both (working on it and removing it from public facing stuff) can coexist … 18:20
ugexe wasn't ; in argument lists supposed to do something in the design docs? maybe i'm thinking of a different language though 18:34
something like marking the point at which other arguments didn't affect multi dispatch or some such
timo2 i think it's just the regular list-of-lists syntax
there is also what you mean, though 18:35
i don't think i've ever had a need for it. could have been `::` for the syntax, or maybe it was `;`, but it's supposed to go in signatures, not in calls
ugexe ah, yeah
timo2 so i'm writing a raku version of the nfa runner with more introspection capabilities and a way to easily go backwards and forwards and such 18:37
i'm hoping that will help others with understanding how NFAs work, and also help me figure out possible improvements for the optimizer
Geth rakudo/main: efba5598a9 | (Will Coleda)++ | tools/releasable/Akefile
Remove comments noting step numbers

These are not in sync, and the unicode numbering was cute but hard to maintain		 19:14
rakudo/main: 393a34adf3 | (Will Coleda)++ | tools/releasable/Akefile
Don't run spectest on master twice during release
rakudo/main: e90f70e384 | (Will Coleda)++ (committed using GitHub Web editor) | tools/releasable/Akefile
Merge pull request #5988 from rakudo/coke/test-less

Skip an entire duplicate spec test run during release
rakudo/main: b1eca4bc40 | (Will Coleda)++ | tools/releasable/Akefile
Unify threads/test_jobs

Default to num cores -1 (previously hardcoded to 7). Now release managers can use TEST_JOBS=x to set both parallel testing and compiling with a single variable.
Resolves #5987		 19:16
rakudo/main: 4a043ba510 | (Will Coleda)++ (committed using GitHub Web editor) | tools/releasable/Akefile
Merge pull request #5989 from rakudo/coke/respect-threads

Unify threads/test_jobs
21:07 sjn left 21:08 sjn joined 21:19 coleman left 21:21 coleman joined 21:30 ds7823 left
timo2 gist.github.com/timo/ca341da345d0c...05cecf2d74 - run the script with a single string argument to filter the names of methods, then you can pick one interactively and enter "a" to automatically explore the NFA's state space with random-ish strings. it doesn't handle fates or longlit yet 22:26
lizmat timo2++ 22:28
timo2 cool, i wasn't getting codepoint edges right and i fixed that now so i do get to cover a lot more of the NFA 22:41
wow, it takes forever to go through the termish NFA automatically. i wonder if i got it wrong somehow? 22:51
22:52 ds7823 joined
timo2 it's also really ram-hungry ... 22:56
gist.github.com/timo/9ab3625e52b31...a7afe2057a 23:09
23:26 ds7823 left
timo2 this "little script" has long outgrown the state where it would be good to organize stuff a bit ... 23:34
23:52 ds7823 joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!