jdv79 anyone know who runs colabti.org? the logger seems to have been downe for a while. 04:03
sena_kun ===> Fetching [FAIL]: WWW:ver<1.005004>:auth<github:perl6-community-modules> from github.com/perl6-community-modules/WWW.git 09:37
What's curious is that it works locally, likely due to a local cache. Can anyone try `zef install WWW`?
I assume github.com/raku-community-modules/...6.json#L24 should be changed and then re-indexed by the ecosystem? 09:38
jmerelo sena_kun: it probably should. I don't want to do it before solving issues, however 09:39
sena_kun: that shouldn't be a problem, however, unless redirections no longer work. 09:40
sena_kun: anyway, as I say, better solve outstanding issues before re-releasing it. 09:42
sena_kun jmerelo, before solving what issues? Redirects should work, yes, but I am still seeing the error for the third time, so apparently they don't somewhere.
jmerelo sena_kun: well, there are a few of them github.com/raku-community-modules/WWW/issues 09:46
sena_kun jmerelo, github.com/raku-community-modules/WWW/issues/19 <- one of the issues is literally what I have right now. And it is probably caused by a redirect not working.
No? 09:47
jmerelo sena_kun: the comment states that there's some issue with authentication. So we'll need to fix the tests (or the underlying code, or both). 09:48
My point is that it's not worth the while to change the link in META.list in the ecosystem and/or the META6.json link (which seems to have been fixed by lizmat, BTW), without looking at that other issues that might make it fail anyway
sena_kun jmerelo, I think this is a fallback behavior due to issues with fetching. 09:49
jmerelo, no, it is worth. There was a module. People relied on that, other modules relied on that. Now fetching is broken due to a link being changed. And you are saying we should not fix fetching, because there are some other issues that might be irrelevant, wishlist ones, etc? 09:50
I am not getting it.
Not sure if a version bump is necessary. 09:56
jmerelo sena_kun: let me work on that this morning. And yes, a version bump is necessary. 10:08
sena_kun jmerelo, I already did github.com/raku-community-modules/...eb30edff47 10:09
jmerelo sena_kun: thanks. I'm still trying to fix a few things... Also need to test if that effectively solves issue #19 10:26
timotimo i have no reason for wanting this, but it randomly occurs to me that when you return something from .^parameterize that also has a .^parameterize you can have types be like MyType[foo][bar][baz] 15:03
Geth rakudo: 8a3e983c99 | (Elizabeth Mattijsen)++ | 2 files
Throw on non-Int result in Iterable index on native array

Because slices on native arrays return the same type of array on which the slice is based, it cannot represent nested indexes like:
   @a[1,2,(3,4)]
   @a[1,2,*]
... (5 more lines)		 15:46
Geth rakudo: 5eb5507640 | (Elizabeth Mattijsen)++ | 2 files
Throw on non-Int result in Iterable index on native array assignment

See previous commit for rationale		 16:20
Geth rakudo: 4d40e23e0f | (Elizabeth Mattijsen)++ | 2 files
Add candidate for native slice assignment from Iterable

Makes slice assignment on a native array from an Iterable about 15x as fast. Also makes it more correct:
If the values for a slice assignment to a native array are not of the same type, it would fall back to the generic slice assignment. This would a. return a List rather than a native array of the same type, and b. would not die on passing nested indices like with other slices on native arrays.		 17:03
MasterDuke lizmat: just curious, do you get a noticeably longer stage parse on the faster-slice-access branch? 17:05
lizmat MasterDuke: yes, because there is a lot more code to parse 17:06
MasterDuke yeah, i wondered if it was enough to make a difference 17:07
lizmat 79603 vs 73339 lines in gen/moar/CORE.c.setting 17:08
timotimo wow 20:29
that's a lot of code right there
lizmat yes 20:52
lizmat that's only to get around the performance issue of a role calling a private method in a consuming class 20:53
fwiw, this only creates more code
if it had been generated as roles and consuming classes, it would still codegen the same amount of code 20:54
would it not?
so the binary size would be the same whether they'd come from generated code like this, or codegenned code from roles and classes 20:55
the only disadvantage, as MasterDuke noted, is increased core setting compilation time 20:56
sortiz Trying to address an issue on Window 10, I find many failures at rakudo's test time. Details in gist.github.com/salortiz/7934a39b7...6fa8f23c71 21:45
I've been out of the loop for many months. Are those glitches expected in Windows? 21:56
Geth problem-solving/solution-250: 5637dc6f8d | Altai-man++ | solutions/documentation/search-categories.md
Provide a solution document for github.com/Raku/problem-solving/issues/250

  "Documentation search categories are not standartized"
Fixes github.com/Raku/problem-solving/issues/250		 22:01
problem-solving: Altai-man++ created pull request #256:
Provide a solution document for #250		 22:02
nine sortiz: no? 22:07
tellable6 2021-01-04T18:40:06Z #moarvm <brrt> nine it may be worth having a MVM_VECTOR_PARAM(x) macro that specifies the vector as a parameter, and maybe an equivalent MVM_VECTOR_ARG(x) that does the same for arguments

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!