00:37 lucasb left 01:04 nwc10 left 02:53 vrurg joined 03:03 MasterDuke left 05:36 Altai-man joined
nine In fact my uncomitted fix makes it much worse. I guess that shows how far away from working interning of parametrics we were and that there's quite some more way to go 06:27
As much as I'd like to return to in-process-precompilation, I guess that smart thing is to use my new found familiarity with parametrics to see this through to the end 06:28
moritz why do I keep reading "paramedics"? :D 06:44
nine That's just your personal bias showing. You obviously value rescuing people's lives over fixing software. Weirdo... 07:24
nine rolls his eyes exaggeratedly 07:25
07:41 domidumont joined 08:03 jmerelo joined 08:04 Kaiepi left 08:06 Kaiepi joined 08:07 Kaiepi left 08:08 Kaiepi joined
lizmat Files=1336, Tests=113712, 221 wallclock secs (28.83 usr 8.49 sys + 3079.73 cusr 293.86 csys = 3410.91 CPU) 08:11
08:11 MasterDuke joined 08:14 Kaiepi left 08:16 Kaiepi joined
Geth rakudo: 964f6e48a5 | (Elizabeth Mattijsen)++ | src/core.e/hash_multislice.pm6
We must also reduce recursion level :-)
rakudo: 3ca1fede64 | (Elizabeth Mattijsen)++ | 3 files
Implement @[;] with all the bells and whistles

Since there are subtle differences with the current (incomplete) implementation, this is only available in v6.e.PREVIEW. Allows for use of * , and WhateverCode such as *-1, as well as iterables at any level. Optimised for the simple 2 and 3-dimensional cases, but due to checks for named parameters, is about 1% slower than the current ... (5 more lines)
MasterDuke timotimo: i'm looking a coverage report that says `--` and `++` are uncovered for native nums. but i can see tests in roast that do use them on natives. could moarvm and/or rakudo optimization be the problem? would it make sense to try and run the spectest with optimization and spesh disabled? 09:17
09:21 sena_kun joined 09:23 Altai-man left
timotimo MasterDuke: yes rakudo optimizes these to nqp::add_i and nqp::sub_i (or maybe inc and dec?) if it sees that it can 09:30
er, not int of course, num i guess 09:31
moritz and for natives, it always can, no?
timotimo should, yeah
MasterDuke i'm running a spectest now with MVM_SPESH_DISABLE=1 and `--optimize=off`, but there are tons of fails
timotimo try optimize=0 09:32
MasterDuke ha, and one TODO passed
timotimo the optimizer is responsible for a few checks that the language specifies must be done
like the "this call to that sub can never work"
MasterDuke yeah, i know jnthn has mentioned breaking the optimizer up into more specific pieces, which i'm all in favor of 09:33
timotimo rakuast has this lookup / resolve / something phase which would be able to do this particular thing
09:34 travis-ci joined
travis-ci Rakudo build failed. Elizabeth Mattijsen 'We must also reduce recursion level :-)' 09:34
travis-ci.org/rakudo/rakudo/builds/736997940 github.com/rakudo/rakudo/compare/1...4f6e48a5ca
09:34 travis-ci left
MasterDuke still a bunch of fails with --optimize=0, not sure if fewer yet 09:35
but i guess i can just merge all the coverage logs to get a superset report 09:36
it is fewer with '0' instead of 'off' 09:37
timotimo good 09:38
lizmat Travis mumbling appears to be caused by t/02-rakudo/15-gh_1202.t failing: which is a flapper type test 09:47
MasterDuke about same with 1, far fewer with 2 09:50
ah ha. there are more lines covered in my superset report. including those num multis! 10:00
from t/02-rakudo/15-gh_1202.t, `No such method 'cue' for invocant of type 'Awaiter::Blocking'` 10:15
10:54 frost-lab joined 12:09 jmerelo left
lizmat notable6: weekly 12:32
notable6 lizmat, 4 notes: gist.github.com/a42bdf6a5a28c8dfd4...15215ccb88
lizmat notable6: weekly reset
notable6 lizmat, Moved existing notes to “weekly_2020-10-19T12:32:43Z”
13:08 jmerelo joined 13:20 Altai-man joined 13:23 sena_kun left
Geth rakudo: d8db03b5da | (Patrick Böker)++ | tools/lib/NQP/Config/Rakudo.pm
Build: If NQP is stale and `--gen-nqp` is passed: rebuild

If `--gen-nqp` is passed don't fail with "nqp version XXX is outdated" errors, rebuild in that case.
Fixes #3964
rakudo: f0d91a50a6 | (Elizabeth Mattijsen)++ (committed using GitHub Web editor) | tools/lib/NQP/Config/Rakudo.pm
Merge pull request #3967 from patrickbkr/fix-3964

Build: If NQP is stale and `--gen-nqp` is passed: rebuild
linkable6 RAKUDO#3964 [closed]: github.com/rakudo/rakudo/issues/3964 ===SORRY!=== nqp-m outdated +build crash
nqp: 6f3d917fc7 | (Elizabeth Mattijsen)++ | tools/templates/MOAR_REVISION
Bump MoarVM to get memory leak fix, nine++
rakudo: 2d461268f6 | (Elizabeth Mattijsen)++ | tools/templates/NQP_REVISION
Bump NQP to get memory leak fixes, nine++

Also a double check if github.com/rakudo/rakudo/pull/3967 had the desired effect :-). It did.
rakudo: patrickbkr++ created pull request #3968:
Don't die when using stdout with `rakuw.exe`
lizmat and another Rakudo Weekly News hits the Net: rakudoweekly.blog/2020/10/19/2020-42-recipes/ 14:01
14:06 patrickb joined 14:09 frost-lab left
patrickb .tell rba Can do. Though tyil should be on boat as well, as he currently creates star. Maybe it's automated already? 14:11
tellable6 patrickb, I'll pass your message to rba
tyil o/ 14:12
patrickb: what'd I miss that I should be aware of?
MasterDuke lizmat: you mention jnthn's report on the newdisp branch, but link to the rakuast report
lizmat artgh
patrickb tyil: A few days ago: #raku-dev <rba> patrickb: Can we discuss automation of star builds based on Azure Pipelines? As I understand all the building blocks are around for binary builds for Win/Mac. What do you think? 14:13
tyil ah
lizmat MasterDuke++ # fixed
MasterDuke lizmat++ 14:14
tyil I'm not using Azure myself due to a strong dislike for Microsoft, but I'm not against it being used, and if any help is needed that doesn't require me to use a microsoft account I'm open to it
patrickb .tell rba Also my next bigger raku project (moving our CI over to a custom tool that controls OBS and Azure) is still in the works, so I'd like to not get sidetracked too much...
tellable6 patrickb, I'll pass your message to rba
tyil I have binary rakudo star builds on Sourcehut right now, though those are all for personal testing purposes right now 14:15
(also only for GNU+Linux systems)
14:19 patrickb left
timotimo hell yeah to "window" applications %) 14:31
Geth rakudo: 54eb431482 | (Elizabeth Mattijsen)++ | src/core.e/hash_multislice.pm6
Add one more recursion fix, missed in 964f6e48a5

And update some comments to bring them more in line with array multislice handling.
roast: b2a0fe6ffc | (Elizabeth Mattijsen)++ | S32-hash/multislice-6e.t
Add some hash multislice recursion tests
15:02 Altai-man left
MasterDuke re stackoverflow.com/questions/644196...ex-in-raku is the naive solution just NYI, or not actually correct? 16:05
[Tux] Rakudo version 2020.09-71-g54eb43148 - MoarVM version 2020.09-12-g8c41f6692
csv-ip5xs0.836 - 0.865
csv-ip5xs-208.010 - 8.245
csv-parser25.856 - 28.440
csv-test-xs-200.384 - 0.397
test7.701 - 8.100
test-t2.106 - 2.148
test-t --race0.867 - 0.884
test-t-2033.100 - 33.434
test-t-20 --race9.784 - 9.967
Hmm. I'll kill teams and run again
lizmat yeah, that seems appropriate :-) 16:06
MasterDuke ?
lizmat referring to [Tux] 's remark
MasterDuke ah 16:07
lizmat i've been looking at INTERPOLATE with regards to the rakuast branch
it's... interesting
MasterDuke yes, i can imagine 16:08
timotimo oh yes 16:10
16:10 jmerelo left
lizmat hmmm.... interesting: on the one hand we have spectest todo'd for binding to hash slices 16:11
on the other hand we have at least one active spectest checking whether binding a hash slice dies
yeah, actually 2 16:12
Geth rakudo: d713e451af | (Elizabeth Mattijsen)++ | src/core.e/array_multislice.pm6
BIND value itself should also be raw!
[Coke] Teams is a *pig* on my work machine, I have to close it when I'm not actively using it. 16:15
(last I checked was eating nearly a gig on startup)
Geth rakudo/hash-slice-binding: b1e198dc30 | (Elizabeth Mattijsen)++ | src/core.c/hash_slice.pm6
Implement binding of hash-slices

AKA %h<a b c> := 1,2,3 . For some reason, this was not implemented before, perhaps because of pre-GLR peculiarities making this difficult. Implementing turned out not to be such a big thing, so here goes.
This breaks 2 spectests that specifically test for throwing an ... (8 more lines)
rakudo: lizmat++ created pull request #3969:
Implement binding of hash-slices
[Tux] Rakudo version 2020.09-71-g54eb43148 - MoarVM version 2020.09-12-g8c41f6692
csv-ip5xs0.828 - 0.844
csv-ip5xs-207.987 - 8.220
csv-parser25.139 - 25.791
csv-test-xs-200.381 - 0.392
test7.952 - 7.965
test-t2.007 - 2.198
test-t --race0.851 - 0.853
test-t-2032.762 - 33.450
test-t-20 --race9.478 - 9.582
16:35 maggotbrain joined 17:10 jmerelo joined 17:16 domidumont left 17:17 domidumont joined 17:18 domidumont left
Geth rakudo/array-slice-binding: a7069182d5 | (Elizabeth Mattijsen)++ | src/core.c/array_slice.pm6
Implement binding of array-slices

AKA @a[1,2,3] := <a b c> . For some reason, this was not implemented before, perhaps because of pre-GLR peculiarities making this difficult. Implementing turned out not to be such a big thing, so here goes.
This makes one spectest fail: S32-exceptions/misc2.t, 46 which tests for specific X::Bind::Slice throwage.
This does not make any spectest pass, suggesting we need more tests.
rakudo: lizmat++ created pull request #3970:
Implement binding of array-slices
lizmat afk for a few hours& 17:40
18:00 Kaeipi joined, Kaiepi left
tbrowder hi, 18:07
i've been testing file copy and found an issue i can't find as open in rakudo: 18:08
trying to copy a dir to a file throws as expected, but it destroys the file somewhere during the process! 18:09
i'm looking for an existing test for that now. i'm also looking at rakudo/t ... 18:11
18:18 Kaeipi left
tbrowder i cannot find a test for this case, i'm filing a rakudo issue 18:22
18:33 jmerelo left
releasable6 Next release in ≈4 days and ≈23 hours. 1 blocker. Please log your changes in the ChangeLog: github.com/rakudo/rakudo/wiki/ChangeLog-Draft 19:00
Geth rakudo: MasterDuke17++ created pull request #3972:
Add .gitattributes file & specify file extensions
20:18 rypervenche left 20:19 Kaiepi joined 20:23 rypervenche joined
lizmat .ask sena_kun could you run a Blin on the "Hash.DELETE-KEY-Nil" branch ? 20:39
tellable6 lizmat, I'll pass your message to sena_kun
21:40 vrurg_ joined 21:42 vrurg left
tbrowder i have a fix almost ready to submit. imho it should go in the next release 22:04
test added also 22:05
Geth rakudo: 42e0434d1a | (Daniel Green)++ | .gitattributes
Add .gitattributes file & specify file extensions

This means if you add something like:
  [diff "raku"]
   xfuncname = "^\\s*(((my|our) )?((multi|proto) )?(sub|method|submethod) [^{]+)"
... (5 more lines)
rakudo: 4097370624 | MasterDuke17++ (committed using GitHub Web editor) | .gitattributes
Merge pull request #3972 from MasterDuke17/add_a_gitattributes_file

Add .gitattributes file & specify file extensions
rakudo: tbrowder++ created pull request #3973:
Add fix for issue #3971
linkable6 RAKUDO#3971 [open]: github.com/rakudo/rakudo/issues/3971 Attempting to copy a directory to an existing file throws but destroys the file
Geth roast: tbrowder++ created pull request #690:
Add test for issue #3971
tbrowder fyi, i added BLOCKER label on the issue and pr 22:31
23:39 HarmtH left 23:40 HarmtH joined