00:02 sena_kun joined 00:03 Altai-man_ left 00:07 lichtkind left 02:01 Altai-man_ joined 02:03 sena_kun left 02:09 softmoth joined 02:50 [Tux] left 04:02 sena_kun joined 04:03 Altai-man_ left 04:33 camelCaser left, ccamel joined 05:29 Kaiepi joined 05:38 Kaiepi left 05:41 Kaiepi joined 06:01 Altai-man_ joined 06:03 sena_kun left 07:05 softmoth left 07:13 JJMerelo joined 07:59 [Tux] joined 08:01 sena_kun joined 08:03 Altai-man_ left 08:11 [Tux] left 08:43 [Tux] joined 09:32 JJMerelo left 10:01 Altai-man_ joined 10:03 JJMerelo joined, sena_kun left 10:35 lichtkind joined
Altai-man_ lizmat, gist.github.com/Altai-man/491c9144...6fe7d3db0b <- latest blin 10:56
lizmat Altai-man_++
gfldex raku: sub check { True }; sub s($a where &check) { say $a }; s 'OK'; 11:58
evalable6 OK
gfldex using a &subname in a where clause seams not to be in roast. Should it? 11:59
jnthn m: sub check { False }; sub s($a where &check) { say $a }; s 'OK' 12:00
camelia Constraint type check failed in binding to parameter '$a'; expected anonymous constraint to be met but got Str ("OK")
in sub s at <tmp> line 1
in block <unit> at <tmp> line 1
jnthn (just checking ;))
Yes, though I guess it falls out of smart-match. But these interactions are still worth covering.
12:02 sena_kun joined
gfldex raku: my &block = { False }; sub s($a where &block) { say $a }; s 'OK'; 12:02
evalable6 (exit code 1) Constraint type check failed in binding to parameter '$a'; expected anonymous constraint to be met but got Str ("OK")
in sub s at /tmp/tDC2OVgXz7 line 1
in block <unit> at /tmp/tDC2OVgXz7 line 1
12:02 sena_kun left 12:03 sena_kun joined, Altai-man_ left
gfldex filed as #650 12:05
12:41 JJMerelo left 12:51 lucasb joined 14:00 Altai-man_ joined 14:03 sena_kun left 14:04 statisfiable6 left, reportable6 left 14:07 statisfiable6 joined, reportable6 joined
lizmat finally found one of the match-refactor issues: 14:57
m: dd "a-z".subst(/. <( "-" )> ./, "..", :g) # comes out as ".." after the refactor
camelia "a..z"
jnthn ah, that'll be from/to handling issue 14:59
lizmat yeah...
indeed. the from falls back to the NQP one, which assumes the $!from has been updated by .MATCH 15:01
but HLL Match cannot have a .from, as that would interfere with any rule called "from"
which it does in one of the modules 15:02
that's what you get when you're conflating Cursor with Match and Grammar 15:04
moritz aren't there lotsa tests for Match.from in roast?
moritz confused 15:05
lizmat apparently not in conjunction with <( and )>
timotimo i wonder if we want to put from, and to, and such as ^from and ^to and such 15:08
moritz capture-from, capture-to
lizmat moritz++ 15:09
timotimo prematch, postmatch as well
lizmat timotimo: those imply capture-from/capture-to, no ? 15:16
timotimo well, they return something else 15:19
15:20 JJMerelo joined 15:21 softmoth joined 15:31 lichtkind_ joined 15:32 lichtkind left
Geth rakudo/match-refactor: 2a76754af2 | (Elizabeth Mattijsen)++ | src/core.c/Match.pm6
Change Match!from to Match.capture-from

Because the NQP .from still refers to the raw $!from attribute, and we needed a smarter .from method that would introspect <( and )> captures *without* updating the upper $!from attribute. However, this clashes with grammars that use a rule called "from", so the method was initially made private. ... (9 more lines)
rakudo/match-refactor: 4bc9918c0e | (Elizabeth Mattijsen)++ | src/core.c/Str.pm6
Make Str.subst use capture.from/capture-to
lizmat this fixes File::Ignore, but does not fix Inline::Perl5 :-*(
moritz m: say 'abc' ~~ / b <( c /; say $/.prematch 15:44
camelia 「c」
16:01 sena_kun joined 16:03 Altai-man_ left
lizmat moritz: unchanged in the match-refactor branch 16:15
18:00 Altai-man_ joined 18:03 sena_kun left 18:40 JJMerelo left 19:17 ufobat joined 19:21 Xliff joined 19:29 ufobat_ joined
dogbert17 lizmat: a few other modules are still failing but you probably already know this 19:32
lizmat yeah, Altai-man_ provided me with another Blin run
19:33 ufobat left 20:01 sena_kun joined 20:03 Altai-man_ left 20:13 MasterDuke joined 20:18 patrickb joined
[Tux] Rakudo version 2020.06-7-gf1960baa9 - MoarVM version 2020.06-6-gbf6af07de
csv-ip5xs0.972 - 1.080
csv-ip5xs-209.356 - 9.516
csv-parser32.800 - 33.848
csv-test-xs-200.427 - 0.433
test9.892 - 10.271
test-t2.444 - 2.461
test-t --race1.267 - 1.703
test-t-2041.325 - 42.117
test-t-20 --race11.623 - 11.676
(during the zoom session listening to jnthn)
sena_kun ouch 20:56
[Tux] tux.nl/Files/20200626230356.png system load. So that explains it 21:04
21:17 ufobat_ left
moritz zoom can be a real resource hog 21:17
22:01 Altai-man_ joined 22:03 sena_kun left 22:22 patrickb left 22:37 lichtkind_ left
Geth rakudo/match-refactor: 82223076a5 | (Elizabeth Mattijsen)++ | src/core.c/Match.pm6

Checking the capnames hash by itself is *not* enough, one needs to actually look at the $!cstack as well. Which it now does.
rakudo/match-refactor: 5b626a4d2b | (Elizabeth Mattijsen)++ | src/core.c/Match.pm6
Fix typo/thinko
rakudo/match-refactor: c56c996f4f | (Elizabeth Mattijsen)++ | src/core.c/Match.pm6
Fix :exists further

Captures that are expected to return more than one value, are always considered to exists, as they always return (a potentially empty) Array. Also, since EXISTS-KEY is probably going to be used much more, make that contain the base logic, and have EXISTS-POS refer to that.
Altai-man_ lizmat, another blin? 22:41
lizmat Altai-man_: yes please! 22:42
Altai-man_ Pancake party...