🦋 Welcome to the IRC channel of the core developers of the Raku Programming Language (raku.org #rakulang). This channel is logged for the purpose of history keeping about its development | evalbot usage: 'm: say 3;' or /msg camelia m: ... | Logs available at irclogs.raku.org/raku-dev/live.html | For MoarVM see #moarvm
Set by lizmat on 8 June 2022.
00:14 Guest72 left 01:29 nine left 01:30 nine joined 02:31 Guest72 joined 08:15 finanalyst joined 09:02 finanalyst left
lizmat m: dd infix:<,>(1); my $op = &infix:<,>; dd $op(1) 09:32
camelia 1
(1,)
lizmat that feels... inconsistent
looking at the QAST of: dd infix:<,>(1) it QASTs as: dd 1 09:35
so which one is correct? opinions ab5tract nine ? 09:42
Geth rakudo/main: 9564c4fdc6 | (Elizabeth Mattijsen)++ | src/core.c/metaops.rakumod
Return to original approach of triangle METAOP_REDUCE_LIST

As described in #5882, commit ba2431f4456ad43 actually changed the internal logic of the triangle metaop list reducer to check for the first value: if that indicated the end of the values, then the iterator wouldn't actually be returned, but an optimized version for empty lists. ... (13 more lines)
10:30
roast: 3b28f5e8c7 | (Elizabeth Mattijsen)++ | S13-overloading/metaoperators.t
Add test for R#5882
10:33
linkable6 R#5882 [open]: github.com/rakudo/rakudo/issues/5882 "Triangular comma" metaoperator [\,] unexpectedly puts first element inside a list
nine lizmat: with RakuAST both print (1,) 10:35
lizmat ah, good, so another test in rakuast-fixed is warranted :-)
meh, doesn't appear to be testable 10:50
easily 10:51
ab5tract: looks like the subset error fixes broke xx-fixed-in-rakuast.rakutest test 100 10:55
Geth rakudo/main: 94ebbccdbf | (Elizabeth Mattijsen)++ | t/12-rakuast/xx-fixed-in-rakuast.rakutest
Add test for infix:<,>(42) behaviour

As spotted in irclogs.raku.org/raku-dev/2025-05-22.html#09:32
10:57
rakudo/main: 79a332b52d | (Elizabeth Mattijsen)++ | t/12-rakuast/xx-fixed-in-rakuast.rakutest
Oops, forgot to todo the test without RakuAST
10:59
12:09 Guest72 left
ab5tract lizmat: I don't comprehend 12:23
tellable6 2025-05-22T08:00:24Z #raku <wayland76> ab5tract: To clarify, I wouldn't want to introduce extra stuff into Raku constructors either; it'd be a module.
lizmat ab5tract: RAKUDO_RAKUAST=1 raku t/12-rakuast/xx-fixed-in-rakuast.rakutest fails 12:24
well, actually dies
aah... maybe not... 12:25
hmmm... the output is weird... 12:26
looks like tests following test 100 are now part of its subtests ? 12:27
ab5tract I didn't take this to be related to my subset changes, as it seemed wildly broken for some reason
ie, this is the error I get "Type check failed in assignment to $grammar; expected Mu but got Raku::Grammar (Raku::Grammar.new() ...)" 12:28
lizmat is this on HEAD?
ab5tract yeah
recompiling to be sure 12:29
maybe it's mad at -Ilib
lizmat this is what I see: gist.github.com/lizmat/c6f61d8cda9...b6440c2ef2
the test in line 46 is what I just added
hmmm... ab5tract indeed doesn't appear to be anything related to subsets, sorry for the noise 12:35
what I don't get, is that without RakuAST it runs 108 tests 12:36
and with only 100? But still comes out ok ?
ab5tract well the test does seem to be broken in that there is no expectation that the thing would die
`throws-like` shouldn't imply death 12:37
I guess I can toss a use fatal in there though
now, as to how the further tests get nested as subtests, I have no clue 12:41
hmm, it seems that there is something unfortunate happening with the `unk_routines` tests 12:49
I don't think `throws-like` appreciates being put into a try 12:53
but without it we get `This representation (P6int) cannot unbox to a native string (for type BOOTInt)`
also, I don't understand why the EVAL'd code is reproduced in the output for throws-like in the subset tests, but not in other uses of throws-like 12:56
lizmat ab5tract: I'll put it on my list to figure out, thanks for looking! 13:01
Geth rakudo/main: b29ab0f523 | ab5tract++ | t/12-rakuast/xx-fixed-in-rakuast.rakutest
Force new subset-type-without-default test to die

Without `use fatal`, this code doesn't satisfy the current expectations of `throws-like`, which needs a proper "death" to feel satisfied.
Catching is a bit of an issue in general with worries, but that's for tackling at a later date.
13:11
lizmat ab5tract++ 13:20
afk&
15:06 librasteve_ joined 19:23 Guest72 joined 19:45 librasteve_ left 20:06 Guest72 left 20:07 Guest72 joined 21:20 Guest72 left 21:23 Guest72 joined 22:51 Guest72 left 23:08 Guest72 joined