00:32 dogbert11 joined 00:35 dogbert17 left 01:11 dogbert17 joined 01:14 dogbert11 left 01:15 sena_kun joined 01:16 Altai-man left 01:53 dogbert11 joined 01:57 dogbert17 left 02:08 dogbert17 joined 02:11 dogbert11 left, dogbert11 joined 02:14 dogbert17 left 02:21 dogbert17 joined 02:23 dogbert11 left 02:41 dogbert17 left 02:44 dogbert17 joined
releasable6 Next release in ≈6 days and ≈15 hours. 1 blocker. Please log your changes in the ChangeLog: github.com/rakudo/rakudo/wiki/ChangeLog-Draft 03:00
03:32 lucasb left 05:14 Altai-man joined 05:16 sena_kun left 05:47 yuplushi left 06:30 yuplushi joined
lizmat Files=1334, Tests=113514, 225 wallclock secs (29.19 usr 8.39 sys + 3146.04 cusr 283.47 csys = 3467.09 CPU) 06:31
08:00 [TuxCM] left 08:43 leont joined 08:52 sena_kun joined 08:53 Altai-man left 09:02 samcv left, samcv joined
leont Remind me, what do I need to do again when «error: Entry '3rdparty/nqp-configure/README.md' not uptodate. Cannot merge.» 09:25
12:51 Altai-man joined 12:53 sena_kun left 13:34 lucasb joined
Geth nqp/nqp-js-S23: daf8df6c99 | (Nicholas Clark)++ | 2 files
Serialization format 23 for nqp-JS

nqp-JS is currently a cross-compiler, and so needs to be able to read the serialization output of MoarVM. As of Feb 2020, commit 9709537d90d61529:
   Serialize the names of closures
   The name is held per code object, meaning different closures may carry a
... (14 more lines)
14:56
nqp: nwc10++ created pull request #655:
Serialization format 23 for nqp-JS
15:03
15:36 squashable6 left 15:38 squashable6 joined 16:51 sena_kun joined 16:54 Altai-man left
Geth ¦ rakudo: Altai-man assigned to Kaiepi Issue Blin 2020.08, round 1 github.com/rakudo/rakudo/issues/3850 18:39
sena_kun .tell Kaiepi hi! Wanted to assign the ticket to both you and nwc10, but for some reason github does not allow that, so pinged instead, so it's not like you are completely responsible for both issues. Thanks! 18:46
tellable6 sena_kun, I'll pass your message to Kaiepi
sena_kun releasable6, status 18:50
releasable6 sena_kun, Next release in ≈6 days and ≈0 hours. 2 blockers. Changelog for this release was not started yet
sena_kun, Details: gist.github.com/9ffc1fb1ad410bcbad...e593de7e1c
gfldex raku: my $*dynvar = 'yes'; sub s($a where ($*dnynvar ~~ 'yes')) { say $a }; s(42); 18:54
evalable6 (exit code 1) Dynamic variable $*dnynvar not found
in sub s at /tmp/vXPN3XcSO6 line 1
in block <unit> at /tmp/vXPN3XcSO6 line 1
gfldex raku: my $*dynvar = 'yes'; sub s($a where {$*dnynvar ~~ 'yes'}) { say $a }; s(42); 18:55
evalable6 (exit code 1) Dynamic variable $*dnynvar not found
in sub s at /tmp/jEui1voZLw line 1
in block <unit> at /tmp/jEui1voZLw line 1
gfldex is this a Rakudobug?
(I hope it is because I got a usecase for that.)
lizmat m: my $*dynvar = 'yes'; sub s($a where {$*dynvar ~~ 'yes'}) { say $a }; s(42); 19:00
camelia 42
lizmat it works if you spell it ok
ah, but you mean if it doesn't ?
if a dynvar doesn't exists, it returns a failure 19:01
gfldex ohh
in that case i may have a problem elsewhere
lizmat m: my $*dynvar = 'yes'; sub s($a where {($*somethingelse // '') eq 'yes'}) { say $a }; s(42);
camelia Constraint type check failed in binding to parameter '$a'; expected anonymous constraint to be met but got Int (42)
in sub s at <tmp> line 1
in block <unit> at <tmp> line 1
19:02
lizmat alternately, probably better: is possible, defined a define a "default" $*dynvar at the lowest possible level 19:03
gfldex i'm using `CALLERS::<$*colored-exceptions> ~~ 'yes' // 'no'` to deal with undeclared dynvars 19:04
lizmat why so difficult? 19:05
gfldex I don't want to force users of a module to define a dynvar that may not even be used. 19:06
lizmat ($*FOO // '') eq 'yes'
that will be True if it is defined and has "yes" in it, otherwise False ? 19:07
gfldex raku: sub s(){ ($*FOO // '') eq 'yes' }; s(); 19:08
evalable6
gfldex that will help. thanks a lot 19:09
lizmat yw 19:13
19:19 lichtkind joined 19:32 sena_kun left, sena_kun joined
lizmat m: class Array is Array { method AT-POS($pos) { say "fetching #$pos"; nextsame } }; my @a is Array = 1,2,3; say @a[0] 20:27
camelia fetching #0
1
lizmat m: class Array is Array { method AT-POS($pos) { say "fetching #$pos"; nextsame } }; my @a = 1,2,3; say @a[0]
camelia 1
lizmat looks to me the codegen is not looking up Array in the right context 20:28
20:51 Altai-man joined 20:53 sena_kun left 21:20 HarmtH joined
Kaiepi .tell sena_kun, got tests to pass for Data::Dump again. it depended on Parameter.default returning Any, which that commit breaks in favour of returning Code. making the pr rn 21:41
tellable6 Kaiepi, I'll pass your message to sena_kun
ugexe theres quite a bit of api changes in that commit. like should Parameter.name *really* return an empty string instead of Nil if its Nil? That doesnt seem right to me 22:48
method name() { nqp::isnull_s($!variable_name) ?? Nil !! $!variable_name } 22:49
method name(Parameter:D: --> Str:D) { nqp::isnull_s($!variable_name) ?? '' $!variable_name }
going from the first to the second seems wrong from me just quickly glances at it
maybe there is some good reason though, alas i dont have time to go through the original PR at the moment.
Altai-man Kaiepi, ^ 22:50
ugexe the above isnt related to the breakage btw, just in general 22:52
Altai-man agrees empty string to Nil change is, well, a change
Kaiepi i figured '' would be appropriate to return since there are other methods (usage-name iirc) that return empty strings in similar cases. i'm not very attached to the idea though 22:53
tellable6 2020-08-16T18:46:17Z #raku-dev <sena_kun> Kaiepi hi! Wanted to assign the ticket to both you and nwc10, but for some reason github does not allow that, so pinged instead, so it's not like you are completely responsible for both issues. Thanks!
ugexe usage name looked to return Nil as well, but you changed it to return '' 22:54
only thing i could think of is you wanted to add that --> Str:D and thought it couldnt return Nil 22:55
Kaiepi oh, it was the sigil/twigil/etc. methods i was thinking of
ugexe well that makes me slightly less concerned, although i'm not sure why *those* don't use Nil 22:58
s/slightly//
releasable6 Next release in ≈5 days and ≈19 hours. 2 blockers. Please log your changes in the ChangeLog: github.com/rakudo/rakudo/wiki/ChangeLog-Draft 23:00
Kaiepi some of those look to have been originally written by lizmat. i went along with it because i like the idea of the methods always returning an object of some sort, but i can see why returning Str instead would be preferable 23:05
s/object/instance/ 23:06
i'd see what she has to say on this 23:07
ugexe ah alright, then i suspect everything is ok 23:08
23:24 lucasb left 23:25 lichtkind left
ugexe why did you change to returning `Code` instead of `Any`? 23:29
method default() { nqp::isnull($!default_value) ?? Code 23:32
method default(Parameter:D: --> Code:_) { nqp::isnull($!default_value) ?? Code
er mispaste
but essentially it feels like i'd expect `Any` more than `Code` for a undefy Parameter.default 23:38
a better question might be Any vs Parameter, but Code seems out of place to me compared to those two 23:39
23:40 leont left