timo Failure is explicitly allowed through in the dispatchers.nqp line i refered to 00:00
japhb timo: I missed that in the scrollback. What line? 00:01
timo 4430 in src/vm/moar/dispatchers.nqp in rakudo 00:05
japhb Huh. I wonder why that is? 00:14
lizmat m: my $a := Failure.new # because 00:22
camelia ( no output )
lizmat m: my Int $a := Failure.new # more correct
camelia Earlier failure:
Failed

Final error:
Type check failed in binding; expected Int but got Failure (Failure.new(exceptio...)
in block <unit> at <tmp> line 1
lizmat hmmm
m: my Int() $a := Failure.new # more correct 00:23
camelia ( no output )
lizmat ok, I guess we haz a bug in the coercion logic
timo we compile the coercing code into binding into a register, which doesn't cause type checks 00:30
that's inside Actions.nqp, look for "raku-coerce" or something being used in a dispatch op qastnode
01:59 librasteve_ left 04:56 [Coke]_ joined 04:59 lizmat_ joined 05:01 [Coke]__ joined 05:04 [Tux] left, [Coke] left 05:05 unicodable6 left, sourceable6 left, lizmat left, releasable6 left, [Coke]_ left, committable6 left, committable6 joined, unicodable6 joined, notable6 left, quotable6 left, quotable6 joined, notable6 joined, releasable6 joined, sourceable6 joined, [Tux] joined 05:47 quotable6 left, committable6 left, bloatable6 left, bloatable6 joined, releasable6 left, notable6 left, committable6 joined, notable6 joined, quotable6 joined, releasable6 joined
releasable6 Next release in ≈2 days and ≈11 hours. There are no known blockers. Please log your changes in the ChangeLog: github.com/rakudo/rakudo/wiki/ChangeLog-Draft 07:00
07:37 melezhik joined 09:15 lizmat_ left, lizmat joined
patrickb [Coke]__: It seems you managed to get the nqp CI working by recreating the pipeline. I have now deleted the broken one. We are done, right? 09:23
Geth rakudo: patrickbkr++ created pull request #6017:
Add a missing error check in runner.c
09:35
09:38 vrurg joined
ab5tract Oh, interestingly the coercion stuff may overlap with R#9317 09:38
Sorry R#5317 09:39
linkable6 R#5317 [open]: github.com/rakudo/rakudo/issues/5317 [roles][Will be addressed in RakuAST] role-constrained variable does not accept derived role defined inline
ab5tract At least the location of it 09:40
09:41 vrurg_ left
Geth rakudo/main: 2785bfe909 | (Patrick Böker)++ | src/runner/runner.c
Add a missing error check in runner.c

Found by Snyk.
10:28
rakudo/main: a58018a24e | (Patrick Böker)++ | src/runner/runner.c
Explain what src/runner/runner.c is doing
rakudo/main: e2c5571785 | (Patrick Böker)++ (committed using GitHub Web editor) | src/runner/runner.c
Merge pull request #6017 from patrickbkr/fix-missing-check-in-runner

Add a missing error check in runner.c
11:54 librasteve_ joined
ab5tract lizmat: do these Nqp nodes in the --target=ast output look familiar to you? gist.githubusercontent.com/ab5trac...tfile1.txt 12:05
lizmat i usually look at them like this: 12:13
gist.github.com/lizmat/37e05d33f5c...170ac8d35e
you mean the "Declaration::ResolvedConstant" in your gist ? 12:14
[Coke]__ patrickb: (done) I think so 12:39
12:40 [Coke]__ is now known as [Coke]
[Coke] two failures in the 2nd blin run, just cutting and pasting here, haven't diagnosed: 12:40
* [ ] [Acme::Test::Module::Zef](raku.land/zef:skaji/Acme::Test::Module::Zef) – Fail, Bisected: [7106759](github.com/rakudo/rakudo/commit/71...f662f5fa6)
* [ ] [LWP::Simple](raku.land/zef:dwarring/LWP::Simple) – Fail, Bisected: [c1d8805](github.com/rakudo/rakudo/commit/c1...5409643f0)
but both of those commits are the brand new commits. :| 12:41
... who merged something to main? 12:45
I am VERY surprised to find 3 new commits to main since I started the followup blin run.
(yes, that one seems fine) 12:48
gist.github.com/coke/0a5f70703ba4f...fd8de8332d - latest blin files 12:52
lizmat FWIW, LWP::Simple installs without issues for me 12:53
as does Acme::Test::Module::Zef
[Coke] ... just added the output files from the two failures. 12:54
in the output files, acme says it can't see a META6.json file, and LWP says [LWP::Simple] Failed to get passing tests, but continuing with --force-test 12:55
but there is no instance of 'not ok' or "failed test" in that output
lizmat: thanks for double checking. It seems it's been a few releases since blin has found a *real* issue. 12:56
this run took 14 hours - I tried to overuse CPUs a bit instead of slightly underuse, and apparently the machine didn't like that. Will switch it back to using 87.5% of cores 12:57
(previous one was 12)
(ah, the limit I was at with this VM setup is that I can't easily increase the # of CPUs past 8; memory I can go up to 672G if I need 13:04
Setting up a VM with more CPUS would be a good exercise for creating an azure ARM template to run blin/setup a machine to do a release on/test raku.land, etc. 13:05
(since this one was done ad hoc). Less of a need for blin if brownie ends up working out. (will dig into that more post-release, melezhik)
There are 688 modules tested in the last run that are "alwaysfail", which we don't consider exceptional. 13:07
I'll start a list of them, and we can open tickets with the modules and/or recommend to blin/brownie that we exclude these going forward from testing. 13:09
lizmat could you key them on auth ?
[Coke] blin isn't tracking that, but we could look it up when generating the table. 13:16
(blin has it when testing, just not writing it out when there's a failure)
I imagine a lot of these are "found in REA" which is basically an insta-skip for me at this point. 13:17
e.g. Slang::Nogil:ver<0.09> 13:18
13:20 melezhik_ joined
melezhik_ . 13:20
lizmat yeah, that's not on raku.land 13:28
rakkable: eco-provides Slang::Nogil
rakkable lizmat, Running: eco-provides Slang::Nogil, please be patient!
lizmat, Found 6 lines in 2 files (1 distributions): 13:29
lizmat, gist.github.com/6d74631c74fe94c619...5e06bcd555
13:31 melezhik_ left, melezhik_ joined
lizmat github.com/tinmarino/nogil/issues/3 13:37
guess I'll fork that one of these days
afk& 13:38
13:42 melezhik_ left, melezhik_ joined
melezhik [Coke]: ++ 13:45
13:56 melezhik_ left, melezhik_ joined 14:05 melezhik_ left 14:06 melezhik_ joined
[Coke] If someone wants to write up some raku, could use a script to take that data.json, parse it, find anything that isn't "OK", make a markdown table with name, version, source-url, status, space for comments, space for a ticket link if we have one open in the module about the failure. if the source-url is in REA, the comment can be something like "REA: skip" (or maybe have an ecosystem column) - will have 14:09
to likely run 'zef info' on the each row that isn't OK.
(if someone picks up that task, tell me so I don't duplicate effort when I swing back around to it. :) 14:15
also I have about 15 lines of code thrown at it so far if someone wants those
14:17 melezhik_ left, melezhik_ joined 14:24 melezhik_ left, melezhik_ joined 14:31 melezhik_ left, melezhik_ joined 14:38 melezhik_ left, melezhik_ joined 14:48 melezhik_ left, melezhik_ joined 14:55 melezhik_ left, melezhik_ joined 15:02 melezhik_ left, melezhik_ joined 15:04 melezhik_ left, melezhik_ joined 15:13 melezhik_ left, melezhik_ joined 15:20 melezhik_ left, melezhik_ joined
[Coke] I have a volunteer thing early saturday morning, so I will probably start the release Friday night and finish closer to noon (Eastern US) saturday 15:26
15:32 melezhik_ left, melezhik_ joined 15:39 melezhik_ left, melezhik_ joined
Geth rakudo/main: 5438dea2fa | (Patrick Böker)++ | 4 files
Win: Tell users they might need to do a reboot post install
15:45
rakudo/main: 5ddbdb9855 | (Patrick Böker)++ | 11 files
Improve the Windows first time experience

  - Reword the READMEs
  - Improve the `set-env` scripts to also change the shell to UTF-8
  - Explain the Windows unicode mess in the README
  - Update the BuildTools installer wizard to the latest BuildTools version
   and make it fully automated
rakudo/main: 3d21d069d7 | (Geoffrey Broadwell)++ (committed using GitHub Web editor) | 11 files
Merge pull request #6016 from patrickbkr/install-improvements

Install improvements
15:47 melezhik left
[Coke] HEY 15:49
PLEASE STOP MERGING THINGS TO MAIN
WITHOUT CHECKIING WITH THE RELEASE MANAGER
THIS CLOSE TO A RELEASE
japhb [Coke]: Very sorry! 15:52
[Coke] At least it wasn't the same person who did it twice. :|
If we can show that this is likely safe, we can leave it in, but it seems like it's changing things that are hard to test. 15:53
japhb Nah, that was pure screwup on my part. Had gotten so used to reviewing PRs in other repos, my brain didn't click that I should stay hands off of this one
It's trying to give a better install experience for Windows folk.
I had done a first-timer UX walkthrough with a friend, and discovered that the Windows experience can be confuzzling 15:54
So patrickb++ and I had been working through making that better.
[Coke] OK. seems like a nice thing to try to include in a release. Was this tested by someone other than the author?
japhb Not that I know of; I don't have a Windows build env. 15:55
patrickb: Could you create a test MSI for my friend to try?
[Coke] (and if you do, do so off main. :) 15:56
japhb Heh
16:00 melezhik_ left, melezhik_ joined 16:07 melezhik_ left, melezhik_ joined 16:09 melezhik joined 16:19 melezhik_ left, melezhik_ joined
patrickb japhb: Can send you an MSI this evening. 16:22
[Coke]: Those changes will break the binary release pipeline should they be flawed. So the source release will be unaffected and the issue will show up on my end after the release. 16:23
Fixing those are possible to do without point releases.
Ugly, but possible.
16:28 melezhik_ left, melezhik_ joined
japhb patrickb: Great, thank you 16:32
16:37 melezhik_ left, melezhik_ joined 16:51 melezhik_ left, melezhik_ joined 16:58 melezhik_ left, melezhik_ joined 17:10 melezhik_ left, melezhik_ joined
[Coke] patrickb: sounds good. Thanks for the clarification 17:13
17:19 melezhik_ left
[Coke] can I get feedback on github.com/rakudo/rakudo/pulls?q=i...l%3ARELENG ? 17:19
17:19 melezhik_ joined
[Coke] FYI, someone who I thought was blocked from interacting with our repositories reacted to the first comment on #6000 17:20
lizmat [Coke]: reviewed 17:22
Geth rakudo/main: f612611165 | (Will Coleda)++ | tools/create-release-announcement.raku
Remove mention of JVM in release announcement

We can make a big announcement when it returns
17:23
rakudo/main: 0949e869a4 | (Will Coleda)++ (committed using GitHub Web editor) | tools/create-release-announcement.raku
Merge pull request #6000 from rakudo/coke/quiet

Remove mention of JVM in release announcement
rakudo/main: bb515e4a1a | (Will Coleda)++ | tools/create-release-announcement.raku
Minor cleanups to release announcement

Mention the discord, don't repeat ourselves.
rakudo/main: 6fe6c5f0c6 | (Will Coleda)++ | tools/create-release-announcement.raku
whitespace
rakudo/main: 40b4e11b53 | (Will Coleda)++ (committed using GitHub Web editor) | tools/create-release-announcement.raku
Merge pull request #5983 from rakudo/coke/release

Minor cleanups to release announcement
[Coke] lizmat++
17:31 melezhik_ left, melezhik_ joined 17:38 melezhik_ left, melezhik_ joined 17:43 coleman_ joined 17:45 melezhik_ left, melezhik_ joined 17:50 coleman left, coleman_ is now known as coleman 17:57 melezhik_ left, melezhik_ joined 18:04 melezhik_ left, melezhik_ joined 18:11 melezhik_ left, melezhik_ joined 18:18 melezhik_ left, melezhik_ joined 18:25 melezhik_ left, melezhik_ joined
lizmat A working raku.land/zef:lizmat/Slang::Nogil has hit the ecosystem 18:36
melezhik_ .
[Coke] ... that was not what I was trying to do. :) 19:22
but that does hide the dead one as far as blin is concerned, so lizmat++ for that part too 19:23
lizmat yeah, thought there was at least that upside
also: it turns out that the grammar has improved in such a way since the original Slang::Sigil was released, that creating a new one was basically a cat license from raku.land/zef:lizmat/Slang::Emoji 19:24
19:27 librasteve_ left 19:29 melezhik_ left 19:30 melezhik_ joined 20:26 melezhik_ left, melezhik_ joined
[Coke] Lovely, since I authored the foundation name change, the URL has stopped working. 20:43
... or maybe I tested the wrong one. wtf. 20:45
lizmat ah... 20:46
perhaps I can make a raku.foundation URL do the correct redirect for now? 20:47
[Coke] I have one that works. It's rakufoundation.org/donate, not donate.rakufoundation.org
(that takes you TPF's site as I originally intended) 20:48
lizmat ok, then I don't have to do anything now :)
[Coke] Just frustrated at myself 20:51
21:23 melezhik_ left
Geth rakudo: patrickbkr++ created pull request #6018:
Fix a strange char in build-msi.ps1
22:02
patrickb [Coke] I'd like to merge the above before the release. (there's a Unicode dash in a Powershell script) It's a one char cleanup. It seems to have worked like that before, but I wonder how. OK? 22:04
[Coke] yup! 22:14
already app. the pr
apologies for the all caps before. :( 22:23
22:27 melezhik left
Geth rakudo/main: 221cf2c899 | (Patrick Böker)++ | tools/build/binary-release/msi/build-msi.ps1
Fix a strange char in build-msi.ps1

No idea how this managed to work on the CI in the past
22:39
rakudo/main: ac97ef2089 | (Patrick Böker)++ (committed using GitHub Web editor) | tools/build/binary-release/msi/build-msi.ps1
Merge pull request #6018 from patrickbkr/fix-build-msi-ps1

Fix a strange char in build-msi.ps1