00:12 maggotbrain joined 00:25 leont left 00:32 MasterDuke left 04:43 Kaiepi left 05:10 Altai-man joined 05:42 sena_kun joined 05:44 Altai-man left 06:56 domidumont joined 07:30 MasterDuke joined 08:21 leont joined
sena_kun releasable6, status 08:35
releasable6 sena_kun, Next release in ≈12 days and ≈10 hours. There are no known blockers. 0 out of 36 commits logged
sena_kun, Details: gist.github.com/ad5b5eed94f53d08f8...76cb49cb78
09:41 Altai-man joined 09:44 sena_kun left 10:03 lizmat_ joined 10:05 lizmat left 10:06 lizmat_ is now known as lizmat 10:07 sena_kun joined 10:09 Altai-man left 10:14 AlexDaniel joined, AlexDaniel left, AlexDaniel joined
Geth rakudo/prefix-pipe-pipe: 77b80bd36a | (Elizabeth Mattijsen)++ | src/core.c/Hash.pm6
Make Hash.DELETE-KEY on a non-existing key 40% faster

And much more memory churn friendly: there is no point creating a scalar with the right descriptor, having it deconted on the return anyway. So just return the default value of the descriptor instead.
Mind you, I think returning Nil would be more correct: but that breaks several spectests.
10:56
lizmat argh, on the wrong branch :-( 10:57
Geth rakudo: 3131ce25ac | (Elizabeth Mattijsen)++ | src/core.c/Hash.pm6
Make Hash.DELETE-KEY on a non-existing key 40% faster

And much more memory churn friendly: there is no point creating a scalar with the right descriptor, having it deconted on the return anyway. So just return the default value of the descriptor instead.
Mind you, I think returning Nil would be more correct: but that breaks several spectests.
10:58
rakudo: fcabad4470 | (Elizabeth Mattijsen)++ | src/core.c/Hash.pm6
Eh, the *default*, not the *of* :-(

In reference to github.com/rakudo/rakudo/commit/3131ce25ac
11:29
rakudo/prefix-pipe-pipe: 8dcba57368 | (Elizabeth Mattijsen)++ | src/Perl6/Actions.nqp
Move version test later

Not sure what is the cheaper test, but it possibly expresses the intent better. AlexDaniel++
12:09
rakudo/prefix-pipe-pipe: 0f552af543 | (Elizabeth Mattijsen)++ | src/core.c/Hash.pm6
Revert "Make Hash.DELETE-KEY on a non-existing key 40% faster"

This reverts commit 77b80bd36ad3636237eaeb816c6e9fac7c29a63d.
12:11
MasterDuke lizmat: when on the prefix-pipe-pipe branch, you should be able to do something like `git rebase -i HEAD~4` and just delete that revert+its original commit, force push, and then only the relevant commits will be left 12:13
Geth rakudo/prefix-pipe-pipe: 48b1b20fa3 | (Elizabeth Mattijsen)++ | src/Perl6/Actions.nqp
Move version test later

Not sure what is the cheaper test, but it possibly expresses the intent better. AlexDaniel++
12:14
lizmat MasterDuke++ 12:15
MasterDuke my bash history is littered with `git rebase -i`s
AlexDaniel` mine isn't, but then it's littered with `add -p` :) 12:18
nwc10 mine has both. and cherry-pick onto branches 12:25
I'm up to `redo5` in one place
lizmat fwiw, it looks like all MacOS tests on CircleCI are current borked 12:27
fatal: Could not parse object '8dcba5736809bd645927458fc264b267a9cb67f3'.
:-(
timotimo that's a git error message? 12:28
12:30 Geth left, Geth joined
lizmat looks like... 12:34
Geth rakudo: e0e24a8dbf | (Elizabeth Mattijsen)++ | src/core.c/Main.pm6
Produce a better error message for my &MAIN = { }

Blocks don't have a .cando method.
Spotted at github.com/rakudo/rakudo/pull/3947...-707081211
12:38
13:41 Altai-man joined 13:44 sena_kun left
Geth rakudo/Hash.DELETE-KEY-Nil: e65f5f3161 | (Elizabeth Mattijsen)++ | src/core.c/Hash.pm6
Make Hash.DELETE-KEY return Nil on non-existent keys

Currently, it returns the default value. Which basically does not allow you to differentiate between:
   my %h; dd %h.DELETE-KEY("a"); # Any
... (21 more lines)
15:06
rakudo: lizmat++ created pull request #3954:
Make Hash.DELETE-KEY return Nil on non-existent keys
15:07
15:12 domidumont left 15:31 [Tux] left 15:36 [Tux] joined
Geth rakudo/setops-prec-non: 07af60670e | (Elizabeth Mattijsen)++ | 2 files
Use additive / multiplicative precedence

As AlexDaniel++ pointed out, the precedence setting in the grammar was forgotten to be updated as well. Changing that accordingly, brought up spectest failures.
Change the precedence to "additive" or "multiplicative" (for (.) ). This also fixes the issue that was at the base of this PR, and feels more natural for the type of operator we're talking about here, as you could see each of these operators as a type of addition (or multiplication).
16:26
AlexDaniel well, I can appreciate the humour xD 16:35
Geth rakudo/setops-prec-non: dc1f9758cb | (Elizabeth Mattijsen)++ | 3 files
Restore precedences, change metaop "list" associativity handling

So, no changes in precedence, but change in handling of operators with "list" associativity when creating a meta-mapper for a given operator.
Is also spectest clean, and makes the Z(&) tests pass as well.
17:04
17:42 sena_kun joined 17:44 Altai-man left 18:47 vrurg left 18:48 vrurg joined
lizmat and yet another Rakudo Weekly hits the Net: rakudoweekly.blog/2020/10/12/2020-...irst-year/ 19:13
19:16 unclechu left, AlexDaniel` left 19:17 sena_kun left 19:23 AlexDaniel` joined 19:33 unclechu joined 20:00 HarmtH joined 20:19 melezhik joined
[Tux] Rakudo version 2020.09-39-ge0e24a8db - MoarVM version 2020.09-8-g60070970c
csv-ip5xs0.786 - 0.851
csv-ip5xs-207.852 - 7.890
csv-parser24.368 - 25.429
csv-test-xs-200.389 - 0.392
test7.398 - 7.492
test-t1.921 - 1.963
test-t --race0.829 - 0.855
test-t-2031.065 - 31.203
test-t-20 --race8.778 - 9.428
20:22
20:28 melezhik left 20:31 Geth left
lizmat afk& 20:52