MasterDuke well, i now have `my $kinds = 4; say $kinds` surviving under RakuAST, and `my $23kinds` throwing the correct exception (X::Syntax::Variable::Numeric). however...it loses 290 other spectests... 00:49
but it does get us t/spec/S15-literals/identifiers.t 00:54
gist.github.com/MasterDuke17/0b0d6...cd0c15fe9d is the current status. this change just loses 16 spectests, but doesn't gain any. if i uncomment the added bit in `token variable`, the loss increases to 292, but it also gains one new one (t/spec/S15-literals/identifiers.t) 01:02
nine timo: you don't have to play with the MOP do fall for this. 08:37
m: my $T := Int; dd array[$T]
camelia "Can only parameterize array with a native type, not Int"
Geth nqp/main: 755945a5d1 | (Elizabeth Mattijsen)++ | tools/templates/MOAR_REVISION
Bump MoarVM to get mimalloc bump, MasterDuke++
 09:20
rakudo/main: 29064c7113 | (Elizabeth Mattijsen)++ | tools/templates/NQP_REVISION
Bump NQP to get MoarVM mimalloc bump, MasterDuke++
 09:30
lizmat m: array[Int] # what's wrong with this, nine ? 09:31
camelia ===SORRY!=== Error while compiling <tmp>
Can only parameterize array with a native type, not Int
at <tmp>:1
------> array[Int]<HERE> # what's wrong with this, nine ?
Geth rakudo/main: 6ba184927b | (Elizabeth Mattijsen)++ | src/core.c/RakuAST/Deparse.rakumod
RakuAST: fix deparse snafu on subsets
 09:38
nine lizmat: this is wrong with it: 09:46
m: my $T := Int; array[$T]
camelia ( no output )
nine m: my $T := Int; say array[$T].^name
camelia Str
lizmat aaahhh
appears that 08b449e1ab36ad18c781 was the original reason 09:51
linkable6 (2018-07-26) github.com/rakudo/rakudo/commit/08b449e1ab Make parameterization errors show where they occur
lizmat to fix github.com/rakudo/rakudo/issues/2032
so I guess we need to have some heuristic to determine we're at compile time or not 09:58
looks like our fudging logic doesn't support .rakutest extensions yet 10:29
nine Why not just throw an exception like in the original code? 10:47
It's quite common for us to re-pack such exceptions at a higher level in the call stack 10:49
Geth rakudo/main: 4b884364dc | (Elizabeth Mattijsen)++ | src/core.c/native_array.rakumod
Make native array parameterization die

To prevent it from just returning a string on a failed parameterization at runtime, spotted by nine++		 11:04
nine lizmat: the same is in Buf and there's still the workaround code for turning strings into exceptions 11:53
lizmat the workaround code is in World, and I'm not spending time on that 11:55
will fix the Buf case
lizmat is apparently not awake yet, because can't find the Buf case ? 12:09
[Coke] wonders if whateverable has a million old gists lying about. 13:25
Geth rakudo/main: a22d6740d4 | (Elizabeth Mattijsen)++ | src/Perl6/World.nqp
Remove parameterization returning Str workaround code

By popular request		 14:54
rakudo/main: bf4646e065 | (Stefan Seifert)++ | 3 files
RakuAST: report failure to parameterize types at BEGIN time
 18:44
rakudo/main: ef387e8600 | (Stefan Seifert)++ | src/Raku/ast/type.rakumod
RakuAST: report trying to parameterize non-parametric type
MasterDuke patrickb: i saw github.com/soloterm/screen recently and thought there might be something useful there for your terminal work 19:04
so if i change `$*IN_DECL` into the correct `$*IN-DECL` in my patch, then i lose 16 spectests but gain one 19:06
Geth rakudo/main: ce32eb4157 | (Stefan Seifert)++ | src/Raku/Actions.nqp
RakuAST: throw typed exception on failure to adverb
rakudo/main: 716449d64c | (Stefan Seifert)++ | src/Raku/Grammar.nqp
RakuAST: throw somewhat more helpful error on infix:(&)
nine is still working on t/spec/S32-exceptions/misc.t 19:07
MasterDuke the one that i gain is t/spec/S15-literals/identifiers.t 19:08
Xliff_ nine: To answer your last question: 1,240,000. And I am allowed to find things daunting if its in a different domain! :D 19:17
MasterDuke some of the spectests i lose are surprising. in t/spec/S32-hash/perl.t, expected: '(my Int %{Str} = :a(1), :b(2))', got: '(my Int % = :a(1), :b(2))' 19:19
so missing a parameterization
but then later in the same file. expected: '$(my Any %)', got: '$(my Any %{Any})' 19:20
now adding a parameterization
anyway, afk for a bit
nine Xliff_: holy cow! 19:24
Geth rakudo/main: 48c6805100 | (Stefan Seifert)++ | src/Raku/ast/code.rakumod
RakuAST: report redeclaration of our-scoped routines
 19:31
rakudo/main: 5a45937c44 | (Stefan Seifert)++ | src/Raku/ast/code.rakumod
RakuAST: fix internal compiler error on redeclaration of regexes
rakudo/main: 551e2a6191 | (Stefan Seifert)++ | src/Raku/ast/package.rakumod
RakuAST: prohibit changing repr of stubbed packages

Fixes: class Foo { ... }; class Foo is repr('Uninstantiable') { } # to fail		 19:39
Xliff_ Moo. 19:51
nine: So, since we last talked, I have worked on the following: github.com/Xliff/raku-GOffice, github.com/Xliff/raku-GCR, github.com/Xliff/raku-Gnumeric 19:54
Oh... I also wrote a teleprompter in Raku: github.com/Xliff/raku-Teleprompter 19:55
Still trying to get things zef-ready and if anyone wants to help, that's the area I need it in.
Geth rakudo/main: 7458db3faa | (Stefan Seifert)++ | src/Raku/ast/variable-access.rakumod
RakuAST: report variables in sink context
 20:22
rakudo/main: be1d946342 | (Stefan Seifert)++ | src/Raku/ast/expressions.rakumod
RakuAST: don't sink targets of metainfix assign
rakudo/main: 44c3655134 | (Stefan Seifert)++ | src/Raku/ast/expressions.rakumod
RakuAST: don't sink operands of smart match

Smart match can have lots of side effects, so it's often done in sink context and warnings about this are counter productive.
Geth rakudo: MasterDuke17++ created pull request #5837:
RakuAST: Throw right exceptions for declaring numeric and match variables		 22:58

Powered by Cro and the Raku® Programming Language.

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

Thank you!