Geth roast/6.d-errata: 510d7aefbd | (Elizabeth Mattijsen)++ | S06-currying/named.t
Another set of incorrect .assuming signature tests skipped
07:05
08:04 hurufu joined
Geth roast/6.d-errata: dd832b0333 | (Elizabeth Mattijsen)++ | S03-operators/set_difference.t
Avoid test failures due to some dispatch issue

Don't test both (-) and ∖ using the same test data: this somehow confuses something in dispatch.
Since the chance that this would occur in real life is slim, this feels like the correct decision to unblock the 2026.05 release.
08:30
rakudo/main: c78c69da35 | (Elizabeth Mattijsen)++ | src/core.c/Rakudo/QuantHash.rakumod
Fix action-at-a-distance in infix (-) and ∖

Spotted really by Coke++ in the 6.d-errata spectest: the Great QuantHash Refactor introduced an action-at-a-distance with the (-) (and ∖ ) set operator, because it only cloned the outer IterationSet, and not the Pair objects in them.
... (5 more lines)
09:20
lizmat [Coke] ^^ that will have to make it into the release :-) 09:21
Geth roast: 85d9d853f9 | (Elizabeth Mattijsen)++ | S03-operators/set_difference.t
Enable all infix (-) ∖ tests again

Also add line number information, to make it easier to spot any errors in the future.
09:25
roast/6.d-errata: 59c5299294 | (Elizabeth Mattijsen)++ | S03-operators/set_difference.t
Revert "Avoid test failures due to some dispatch issue"

This reverts commit dd832b0333a2efc59984ae4b98499bbd8f909a3d.
It is no longer needed, thanks to:
   github.com/rakudo/rakudo/commit/c78c69da35
09:26
09:57 hurufu left
Geth roast/6.c-errata: afe5495aa4 | (Elizabeth Mattijsen)++ | 3 files
Make QuantHash immutability tests less specific

The Great QuantHash Refactor changed the error classes to be more in line with general QuantHash behaviour
09:59
roast/6.c-errata: 6d944224ba | (Elizabeth Mattijsen)++ | 4 files
Cat licensed test adaptations from 2026.05 6.d-errata fixes

Basically just replaced "use v6.d" with "use v6.c" and removed some tests that weren't in the 6.c-errata branch yet
10:11
10:32 hurufu joined
lizmat m: use v6.*; dd sprintf "%d", 0 | 1 12:00
evalable6 any("0", "1")
lizmat m: use v6.d; dd sprintf "%d", 0 | 1
evalable6 (exit code 1) Directive %d not applicable for value of type Junction ((any(0,
1),).Iterati...) in format '%d'
in block <unit> at /tmp/RxRNVN91H3 line 1
lizmat afk for a few hours& 13:48
Geth rakudo/release-2026.05: cd723f2261 | (Will Coleda)++ | 3 files
Update changelog + announcement

Deliberately not logged:
  [4acea051][eb798378][f6d89e38][a1f6cb4f][4f641591][0523e6b2][155528fe]
  [ed5ddf8c][300ee1d9][f2273bf1]
13:58
[Coke] I did a git rebase main on the release branch, and will just redo the Ake stuff with a clean checkout. 13:59
Geth rakudo/release-2026.05: 2ef1759a35 | (Will Coleda)++ | 3 files
Update changelog + announcement

Deliberately not logged:
  [4acea051][eb798378][f6d89e38][a1f6cb4f][4f641591][0523e6b2][155528fe]
  [ed5ddf8c][300ee1d9][f2273bf1]
14:05
[Coke] ... and one more to change the release date to today.
lizmat++ getting that fixed! 14:07
that==the tests & the bug 14:08
back up to stress test on master 14:22
starting 6c stress 14:30
6c failure: t/spec/S09-hashes/objecthash.rakudo.moar (Wstat: 256 (exited 1) Tests: 36 Failed: 1) Failed test: 30 Non-zero exit status: 1 14:48
trying 6d... 14:49
looks like that particular one passes in 6.d 14:52
6.d is clean - 6.c has that one failure. 14:55
14:56 liztormato joined
liztormato Coke I think the 6.c one is a flapper 14:56
If not. I’ll look into it when I get home 14:58
15:04 finanalyst joined
[Coke] I'll rerun 15:05
pretty sure the ake stuff in the release is trying everything 2x
(I just opened a ticket to turn off flap protection, we shouldn't need it this many years in. :(
failed again 15:07
(saw it go past, not done with the run yet)
liztormato Ack. Will look in about an hour from now 15:08
15:08 liztormato left
[Coke] it's a flapper 15:12
lizmat: it's a flapper, don't stress out. it just happened to fail 3 times in a row. :)
15:14 finanalyst left
[Coke] github.com/Raku/roast/issues/890 15:14
15:15 finanalyst joined
nine_ A few weeks back I upgraded the OS on camelia's VM and didn't check if camelia was actually responsive afterwards. Service claimed to start successfully, but logs showed missing Perl modules. Anyway I can't seem to get Camelia to run again due to: rt.cpan.org/Public/Bug/Display.html?id=104493 15:15
This prevents installation of POE::Component::SSLify which is required for ssl support 15:16
15:17 camelia joined
nine_ Ah, looking at that patch in the bug report, it may only be an issue in the tests. Good old force install might be all that's needed 15:18
m: say "I'ma back?!"
camelia I'ma back?!
nine_ Given that it will only get harder to get these Perl modules installed, I wonder if someone feels motivated to vibe code a replacement for the evalbot in Raku? 15:19
Geth nqp/main: 38f8db4391 | (Will Coleda)++ | tools/templates/MOAR_REVISION
[release] Bump MoarVM revision to 2026.05
15:36
nqp/main: 92f2909605 | (Will Coleda)++ | VERSION
[release] Bump VERSION to 2026.05
rakudo/release-2026.05: 87eab18a34 | (Will Coleda)++ | tools/templates/NQP_REVISION
[release] Bump NQP revision to 2026.05
rakudo/release-2026.05: 2f9a672e30 | (Will Coleda)++ | VERSION
[release] Bump VERSION to 2026.05
15:38 finanalyst left
Geth rakudo: coke++ created pull request #6206:
Release 2026.05
15:41
[Coke] hurm. the release branch doesn't seem to have the correct release date, despite my efforts. 15:42
This makes me wonder that includes lizmat's most recent fix. :| 15:44
.. ah, ok, I think I managed to not commit that change from may 22 to may 23. I'm going to leave it as is. Seems to include liz's fix, so that's fine. 15:45
anyone asks, that's the day we start the release, not finish it. :)
Geth rakudo/main: 4 commits pushed by (Will Coleda)++ 15:46
[Coke] (also I can just fix it post release)
.tell patrickb 2026.05 is ready for binary releases. 15:58
tellable6 [Coke], I'll pass your message to patrickb
[Coke] .tell el_che 2026.05 is ready for binary releases.
tellable6 [Coke], I'll pass your message to El_Che
16:04 finanalyst joined
[Coke] ... I realize only now that I didn't changelog the last minute fix. 16:19
OK. release is done
this one had a minor hiccup, so please let me know if anything looks weird.
(changelog) Wonder what that'll do to releasable. might have to add that one to the ignore bucket on next release.
Geth roast/6.c-errata: c3affe4ff6 | (Elizabeth Mattijsen)++ | S09-hashes/objecthash.t
Remove flappiness from test

Test depended on hash order, as observed in #890
16:29
roast/6.d-errata: 7505fda2d4 | (Elizabeth Mattijsen)++ | S09-hashes/objecthash.t
Remove flappiness from test

Test depended on hash order, as observed in #890
16:31
lizmat oh, and BTW: [Coke]++ for yet another release! 16:34
16:36 finanalyst left
[Coke] Thanks to you and ugexe and everyone for doing all the hard work. 16:46
I do like doing the moarvm release the day before, may keep doing that.
lizmat sounds like a plan to me! 16:50
Geth rakudo/lizmat-6: abf246b73d | (Elizabeth Mattijsen)++ | src/core.e/Formatter.rakumod
Properly support %g in 6.e sprintf

The previous implementation had several errors *and* erroneously assumed that the precision indication was for number of digits after the period. The "standard" requires that the precision is to be interpreted as "number of significant digits", a subtle difference!
17:01
rakudo: lizmat++ created pull request #6207:
Properly support %g in 6.e sprintf
rakudo/main: fb9cf62d31 | (Will Coleda)++ | docs/release_guide.pod
Correct #193 release date, add #194

_We finished a day after we started 193._
17:09
rakudo/main: 216c855e31 | (Will Coleda)++ | tools/releasable/Akefile
Remove antiflap protection from the release.

We should ideally fix any failing tests so they cannot flap, this will help us catch any of the remaining ones sooner.
Closes #6205
17:13
rakudo/coke/pre-compile: a199d5d01e | (Will Coleda)++ | 5 files
Convert pre-compiled to precompiled

The exception text is not specifically checked for in t or roast.
17:23
rakudo/coke/pre-compile: 7442556533 | (Will Coleda)++ | 5 files
Convert pre-compiled to precompiled

The exception text is not specifically checked for in t or roast.
Closes #6204
rakudo: coke++ created pull request #6208:
Convert pre-compiled to precompiled
17:24
lizmat I was just thinking: would now be a good time to rename the .t and .pm6 files in roast to .rakutest / .rakumod ? 17:43
or should we wait until after 6.e is released ? 17:44
[Coke] I think having it done by the time 6.e is released would be good. Doesn't have to be exactly same time. 18:02
lizmat I guess you could technically argue that .rakumod and .rakutest are 6.e things 18:07
[Coke] they are the current recommended extensions, though. 18:39
18:55 rakkable left 18:56 rakkable joined 21:27 hurufu left
ugexe github.com/Raku/roast/blob/85d9d85...ay16.t#L59 21:46
the author's intent seems to be "return the string if $full-of-vowels matches the regex." But EXPR when COND is EXPR if $_ ~~ COND i.e. it always smartmatches against $_, not whatever expression is inside COND 21:47
that is what is causing github.com/rakudo/rakudo/pull/6203 to fail its spectests 21:48
in other words, i think it should be `is ("My God, it's full of vowels!" if $full-of-vowels ~~ /^ <[aeiou]>+ $/), ...` 21:50
21:51 librasteve_ left
ugexe i guess we could revision gate on a better behavior and add use v6.c to that test as well 22:04