ab5tract timo2: that explanation made sense to me. I hope it's not too hard to try... a faster grammar would be an incredible Christmas presentr 07:17
10:50 tellable6 left, coverable6 left, notable6 left, bloatable6 left, bisectable6 left, benchable6 left, releasable6 left, linkable6 left, committable6 left, unicodable6 left, nativecallable6 left, evalable6 left, shareable6 left, quotable6 left, greppable6 left, sourceable6 left 10:54 bloatable6 joined, releasable6 joined, committable6 joined 10:55 bisectable6 joined, quotable6 joined, unicodable6 joined, linkable6 joined, notable6 joined 10:56 coverable6 joined, benchable6 joined, sourceable6 joined, greppable6 joined 10:57 evalable6 joined, shareable6 joined, nativecallable6 joined, tellable6 joined 12:30 librasteve_ left 14:23 librasteve_ joined 15:21 ds7823 joined
[Coke] yay, back from the thing. 15:33
Geth rakudo/coke/release: bb515e4a1a | (Will Coleda)++ | tools/create-release-announcement.raku
Minor cleanups to release announcement

Mention the discord, don't repeat ourselves.
16:04
rakudo: coke++ created pull request #5983:
Minor cleanups to release announcement
rakudo/release-2025.10: 8545a055bc | (Will Coleda)++ | 3 files
Update changelog + announcement

Deliberately not logged:
  [d370e051][596e8394][89765fe8][86d216d3][9c5fc915]
16:09
[Coke] ok I added a note about not using the trailing zero in the moarvm branches, but I see it was used on origin/2025.080 16:20
Guess I'll try to use it and see how it breaks for me again.
ah. my other other note of "--Coke - note that jdv does not have these errors when doing a release." is also still there. 16:23
Someday we need to sort out how that is working for you. 16:26
moarvm release guide suggests running git tag -n9 to see previous release managers. The output isn't showing release managers but 2025.08's line looks Very weird compared to other releases. 16:29
er, release messages, not managers, misread. but the last entry is weird. we also have a 9999 tag in there for some reason 16:31
yay, I have moarvm.org access these days! 16:37
... fuck 16:38
lizmat que?
[Coke] origingit@github.com:coke/moarvm.org.git (fetch)
It's fine, I was just prematurely happy. explains why it looked like the website hadn't been updated since I did it last! :)
and maybe why the push worked. 16:39
ah, good, push to actual repo also worked. switched over so I don't get confused next time 16:44
MoarVM release done, tagged, files uploaded. 16:55
If possible, someone please verify it looks OK, small break before restarting rakudo release. 16:56
17:23 librasteve_ left 17:28 ds7823 left
[Coke] (running spectests...) 17:53
Thanks again to jdv and all the release guide and tooling authors for making this mostly painless.
gist.github.com/coke/cd4f905bfc2ce...8017a2a4f5 - some todo pass tests 17:55
lizmat yes, that is correct 17:59
because these tests actually depend on static optmization
they are marked todo as they fail in RakuAST (which doesn't have any static optimization) 18:00
[Coke] OK
lizmat and all spectests *should* pass without static optimization
[Coke] semi-regular ping: anyone interested in doing a release? Sign up here! 18:14
... I should print like a release manager challenge coin or something. :)
lizmat :-) 18:24
Geth nqp/main: 913b9dffaf | (Will Coleda)++ | tools/templates/MOAR_REVISION
[release] Bump MoarVM revision to 2025.10
18:36
nqp/main: b242f7e692 | (Will Coleda)++ | VERSION
[release] Bump VERSION to 2025.10
rakudo/release-2025.10: cd7d3af0bc | (Will Coleda)++ | tools/templates/NQP_REVISION
[release] Bump NQP revision to 2025.10
rakudo/release-2025.10: 3f8de27bcb | (Will Coleda)++ | VERSION
[release] Bump VERSION to 2025.10
[Coke] rakudo.org@trinity.raku.org: Permission denied (publickey,keyboard-interactive). 18:43
scp: Connection closed
coleman: any pointers? 18:44
This worked the last time I did a release. I am using a key for will@coleda.com 18:45
ah, nevermind. my ssh config has an entry for "trinity", and I was trying to ssh to the full hostname - so it was ignoring the special config. 18:50
Geth rakudo: coke++ created pull request #5984:
Release 2025.10
18:52
rakudo/main: 4 commits pushed by (Will Coleda)++ 18:53
[Coke] release is done. 19:09
patrickb, el_che, you can work on the binary releases now, thanks in advance. 19:10
updating rakudo.org...
all done - moarvm.org/ isn't showing as updated. 19:16
... my bad
pushed - should be updated shortly. 19:17
(binaries are already on the moarvm.org site, html update shortly) 19:18
Geth rakudo/coke/release: 6fe6c5f0c6 | (Will Coleda)++ | tools/create-release-announcement.raku
whitespace
19:21
19:22 MasterDuke joined 19:24 librasteve_ joined
timo2 ab5tract: i don't have measurements yet for whether it's actually noticeably faster, and if yes by how much 19:43
ab5tract Fair points :) 20:36
20:59 ds7823 joined
timo2 i'm now writing a raku script that allows you to grab the existing NFAs out of a Grammar (for example Perl6::Grammar) and manipulate them, so you can write optimizer passes in raku instead of having to change nqp and recompile nqp and rakudo 21:06
MasterDuke oh, very interesting
21:24 MasterDuke left
timo2 not sure how to override the optimize method on QRegex::NFA in the running process, which would be necessary to allow actually overwriting the entire optimizer, otherwise we can only operate on the already-optimized NFA 22:00
and nqp routines don't support being wrapped 22:09
22:19 ds7823 left