00:48 melezhik_ joined 00:50 melezhik_ left 01:49 melezhik_ joined 01:51 melezhik_ left
[Coke] i have one job still running trying to test 'SparrowCI - super fun and flexible CI system with many programming languages support' 02:10
02:12 vrurg left
[Coke] which appears to have been a typo that was corrected. Can we remove that name from the ecosystem? 02:12
02:12 vrurg joined 02:50 melezhik_ joined 02:55 melezhik_ left 03:51 melezhik_ joined 03:56 melezhik_ left 04:45 melezhik_ joined 04:50 melezhik_ left 04:52 melezhik_ joined 04:56 melezhik_ left 05:54 melezhik joined
melezhik Yes. Need to remove it from eco system and add to skip list 05:54
The round took 2 hours and 25 minutes to certain 05:55
To be certain
06:55 melezhik_ joined 06:59 melezhik_ left 07:07 melezhik4 joined, melezhik4 left
melezhik github.com/melezhik/brownie/commit...cbfb1fb975 07:17
07:35 Geth joined 07:56 melezhik_ joined 07:58 melezhik_ left 08:53 evalable6 left 08:54 quotable6 left, committable6 left 08:57 evalable6 joined, melezhik_ joined, committable6 joined 08:59 quotable6 joined 09:01 melezhik_ left 09:16 librasteve_ joined 09:24 melezhik left
lizmat_ RakuIRCLogger 09:41
09:41 lizmat_ left 09:42 lizmat joined 09:52 melezhik joined 09:55 melezhik_ joined 09:57 melezhik_ left 10:03 melezhik_ joined 10:08 melezhik_ left, melezhik_ joined 10:20 melezhik_ left, melezhik_ joined
melezhik_ how can I remove distribution from eco system? Should I use fez for that? 10:25
I'd like ti remove 'SparrowCI - super fun and flexible CI system with many programming languages support' distribution 10:26
lizmat you can only remove "new" distributions (which have been uploaded less than 24 hours ago) 10:27
although that feature appears to be broken atm tonyo ugexe
10:32 melezhik_ left, melezhik_ joined 10:47 melezhik_ left 12:12 melezhik left 12:16 librasteve_ left
ab5tract gist.github.com/ab5tract/7f5410af7...1-raku-L10 13:59
it can't be correct that I had to add the `$first //= @i[$idx]`, can it? 14:00
ie, could there really be a genuine situation where max returns Any as the value but with a valid key? 14:01
14:02 melezhik joined
melezhik I see 14:02
[Coke]: could you please upgrade the agent to the latest version ? Thanks
ab5tract melezhik: I can add some to this run 14:03
melezhik ?
ab5tract I can run some agents
melezhik Ah, great 👍 14:07
Please upgrade to the latest version
Thanks
ab5tract sure thing :)
lizmat: ^^ (sorry for finding this) 14:08
*this bug
[Coke] melezhik: upgraded. 14:30
(sorry, I let the laptop sleep overnight) 14:31
melezhik [Coke]: thanks , there used to be still one stale job running , could you please terminate it ? 14:34
No worries 14:36
[Coke] melezhik: that's the "[SparrowCI - super fun and flexible CI system with many programming languages support]" distro
melezhik patrickb: if you are going to run your agent , please upgrade it too , thanks . Have made some changes , so want to make it sure all the agents has the same version 14:37
“5:36 PM <[Coke]> melezhik: that's the "[SparrowCI - super fun and flexible CI system with many programming languages support]" distro” yeah . I don’t know why it hangs , but I already added it to skip list , you can just terminate it 14:40
patrickb melezhik Do you think it would be technically possible to implement an auto update functionality? I.e. that the agent listens for a ping from your end, shuts down updates and boots up again? 14:42
melezhik It’s possible .
I think I can do something in the next version 14:43
[Coke] I think it would be helpful if we had a way to delete items in the fez (any) ecosystem that were a) obvious mistakes b) malware 14:44
I would be fine if this required something manual like opening a ticket and having a small set of admins who can approve/reject the request. 14:45
melezhik Agree 14:46
Run another round of 2000 modules against Rakudo 2025.11 14:47
Added some meta information for every module which is tested
For example 14:49
add module 6pm ... OK, status: FAIL, agent: cokebot-27928036, check_method: direct, parent: not set
check method is how module has been checked
It’s one of two 1) direct ( zef install Foo ) - direct installation 2) testfail/testok - indirect , as a dependency of Foo 14:51
Parent - is not null if module has been tested as a dependency for “parent” module 14:52
Also there is “testfail” flag ( not shown in this example ) that if it’s False means the reason of failure of module test is not passing unit tests , which is probably most interesting case as we look for situation when/if Rakudo breaks module unit tests 14:54
[Coke] might be nice to have a timer check - if the tests take longer than... 30m? kill them, mark that as a new status. 14:59
melezhik Yeah , that’s possible as well. From other hand this now reveals such a problematic modules . I think termination on deadline and sending back to o10r timeout signal for further analytics 15:03
Another example of meta information “add module MIME::Base64 ... OK, status: OK, agent: distracted-hopper-96388688, check_method: testok, parent: Cro::WebApp::Evaluate” 15:04
ab5tract has to remember that kv forms no longer have ((k1,v1),...) semantics 15:58
it's only been a decade ;)
timo time to Parcel these memories away 15:59
melezhik [Coke]: could you please upgrade agent agents? Hit weird bug with caching file content in Rakudo 16:26
Agents -> again
[Coke] Done 16:29
lizmat ab5tract: can you repro a case where .max produces Any ? 16:32
I can't seem to 16:33
melezhik [Coke]: thanks
ab5tract lizmat yeah sorry about that, I was just clarfying a moment ago that I was holding it wrong 16:40
lizmat okidoki :-) 16:41
ab5tract I was calling .max(:kv).first
expecting ($k,$v), but only getting ($k)
lizmat ah, the Parcel confusion 16:42
:-)
Geth nqp/main: 31b1fa10eb | (Elizabeth Mattijsen)++ | tools/templates/MOAR_REVISION
Bump MoarVM for Unicode 17 support, Shimmerfairy++
 16:43
rakudo/main: 4d62268aa6 | Faye++ (committed using GitHub Web editor) | 7 files
Update Rakudo for Unicode 17.0.0 (#6034)

This includes updating the version detection, updating the list of Unicode properties in unicodey.rakumod, and updating rakudo's internal tests for the new version. Note that the number of allowed failures for the DerivedGeneralCategory test was bumped up by one, which seemed OK to me.
The makeUNIPROP.raku script was updated to have the latest property aliases, and a minor fix so that it stopped calling itself a .p6 script.		 16:47
rakudo/main: 1e66a50cb3 | (Elizabeth Mattijsen)++ | tools/templates/NQP_REVISION
Bump NQP to get Unicode 17 support, Shimmerfairy++
 16:51
[Coke] c: 1e66a50cb3 3.put 17:08
committable6 [Coke], ¦1e66a50: «3␤»
[Coke] Shimmerfairy++ 17:09
[Coke] kicks off a blin run 17:10
ShimmerFairy Right as I was going to sleep, I'll merge the roast updates too unless someone's already doing that. 17:17
lizmat ShimmerFairy: I didn't see a PR for roast ?
ShimmerFairy sorry, I didn't submit a PR, I just pushed a branch. My fault for not making it clearer. 17:18
lizmat ok, leaving it up to you then :-) 17:20
Geth roast/master: 5 commits pushed by Faye++ 17:21
ShimmerFairy There, now people can run the new tests on the new code. Again, apologies for making that a bit more hidden, I'm just glad I happened to check the chat one final time before zzz. 17:22
[Coke] Do those tests need to be updated on 6.c/6.d ? 17:23
(don't want any surprises december release)
ShimmerFairy Not sure what that means, I just had to update the tests that check MoarVM for its Unicode support. 17:24
[Coke] There are branche specifically for 6.c and 6.d - master is for "latest" 17:25
so if the 6.c or d roast tests will now fail on rakudo master, that's an issue.
ShimmerFairy Yeah, that'll be an issue. Whatever normally has to be done when things are updated. (I'm guessing the 6.c/6.d branches didn't exist when the docs/unicode-generated-test.asciidoc was written?) 17:26
[Coke] docs.raku.org/language/faq#What_is...ification? 17:28
it's not specific to unicode, it's just a generic roast issue.
no worries, worst case will need a cherry pick probably 17:29
ShimmerFairy Luckily Unicode likes its stability, so it shouldn't be too big a disruption. The grapheme rule updates ought to be backwards-compatible (though they may not be, especially if they tested any Indic script graphemes), and over the past few versions there were some changes made to DUCET that could possibly cause collation test failures, not sure though. 17:31
japhb ShimmerFairy++ # Thanks for driving this to completion! 17:57
18:52 melezhik left 18:59 melezhik joined
melezhik Test round for 2025.11 in progress , everything looks good so far , now reports have essentially meta data 19:00
Essential
||| TESTS STAT: time: 002h | tests total: 2000 | finished tests: 1286 | sent to queue: dist=0 / redist=10 | agents cnt: 2 19:01
[Coke] there are only 2389 modules (per blin) btw - might as well run them all? 19:06
melezhik Ok, rerun round for 2025.11 with 3000 limit 19:25
Had to fix some issues . Now it’s running 19:41
||| TESTS STAT: time: <1m | tests total: 2389 | finished tests: 1563 | sent to queue: dist=20 / redist=0 | agents cnt: 2
It’s exactly 2389 btw )))
github.com/melezhik/brownie/blob/3...wfile#L136 19:42
I hire blin uses the same source
hire -> guess
Typos …
19:52 Pixi left, Pixi joined
melezhik Actually run the same for 2025.10 , put this round in o10r queue 19:54
And for 2025.08 19:55
Hopefully agents will do their job till tomorrow
22:42 melezhik left
japhb nine: Inline::Perl5:ver<0.61>:auth<github:niner> failing t/v6.t on Rakudo HEAD 22:59
23:23 librasteve_ 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!