| Geth | rakudo: ugexe++ created pull request #6117: Fix coercion in parameterized Quanthashes |
01:18 | |
|
07:45
finanalyst joined
08:39
finanalyst left
|
|||
| Geth | rakudo/main: 3b8b6d61d4 | (Elizabeth Mattijsen)++ (committed using GitHub Web editor) | 2 files Add a worry for using the unimplemented TEMP phaser (#6105) As discussed in github.com/Raku/problem-solving/issues/511 |
09:48 | |
| rakudo/main: 26155d3768 | (Nick Logan)++ (committed using GitHub Web editor) | src/Raku/Grammar.nqp Fix spurious sigspace warning for rx:s// in RakuAST frontend (#6116) The `quote:sym<rx>` and `quote:sym</ />` tokens in the RakuAST grammar were missing the `:my $*WHITESPACE-OK := 0` dynamic variable declaration. When the `:s` adverb was parsed, `adverb-rx2str-control` tried to set `$*WHITESPACE-OK := 1` but the assignment silently failed (wrapped in try) because the ... (9 more lines) |
09:52 | ||
|
10:49
sjn left
12:06
sp19833_ joined
12:36
sp19833_ left
12:37
sp19833_ joined
13:03
sp19833_ left
13:04
sp19833_ joined
13:45
sjn joined
13:46
sp19833_ left
14:53
sp19833_ joined
14:55
sp19833_ left,
sp19833_ joined
14:59
sp19833_ left
15:01
sp19833_ joined
15:06
sp19833_ left,
donaldh joined
15:07
donaldh_ joined
15:11
sp19833_ joined
15:12
donaldh_ left
15:28
sp19833_ left
|
|||
| lizmat | m: my %h{Int()} = ^10; dd %h<0>, %h{0} # feels to me that should give the same result | 15:47 | |
| evalable6 | Any element of element of %h = Any Any element of %h = 1 |
||
| lizmat | or should it? | ||
| inspired by research into #6117 | 15:48 | ||
| github.com/rakudo/rakudo/pull/6117 | |||
| fwiw, the same issue also applies to QuantHashes that are parameterized with coercing types | 15:49 | ||
| patrickb ab5tract ugexe timo thoughts? | |||
| I'm sorta in two minds about this atm: | 15:52 | ||
| 1. make it work, so %h<0> would return 1 | 15:53 | ||
| 2. don't allow coercing types in parameterization of (Quant)Hashes | |||
| I like 1, but this will mean overhead for the non-coercing parameterized (Quant)Hashes | 15:54 | ||
| I like 2, because it creates clarity and prevents hidden overhead | 15:55 | ||
| fwiw, there are *no* tests for quanthashes parameterized with a coercive type | 16:31 | ||
| sadly there *are* tests for object hashes | 16:44 | ||
| m: my %h{Int(Str)} = "10" => 42; say %h<10>; say %h{10} | |||
| evalable6 | (Any) 42 |
||
| lizmat | see also github.com/Raku/old-issue-tracker/issues/6176 | 16:46 | |
| librasteve_ | rakudoweekly.blog/2026/04/13/2026-...gs-busses/ | 18:00 | |
|
20:04
librasteve_ left
|
|||
| lizmat | librasteve++ | 22:08 | |
|
23:00
MasterDuke joined
|
|||
| MasterDuke | timo: wonder how fast stage parse would be with one of these? dl.acm.org/doi/10.1145/3716368.3735157 | 23:01 | |
|
23:24
MasterDuke left
|
|||
| Geth | roast: schultzdavid++ created pull request #885: +test (r)index for lists of overlapping needles (Rakudo issue #6104) |
23:30 | |