07:28
librasteve_ joined
08:14
sena_kun joined
|
|||
Geth | rakudo/main: f055c31e02 | (Elizabeth Mattijsen)++ | src/core.c/Str.rakumod Move setting of local $/ The fact that $/ was set inside an if, instead of inside the body, may be the reason this failed on the JVM. So move it into the body of the helper sub (even though we may not actually need it). This in response to #5744 |
10:50 | |
lizmat | bartolin: ^^ could you check ? | ||
bartolin | lizmat: nice, that indeed seems to help. I haven't run a full spectest yet, but S05-transliteration/trans.t is clean. \o/ | 11:31 | |
lizmat | I guess the scoping rules on JVM are subtly different from MoarVM | ||
13:03
sena_kun left
|
|||
[Coke] | gist.github.com/coke/0a5f70703ba4f...fd8de8332d - two blin failures since release. | 14:17 | |
BioPerl6 Cro::APIToken | 14:25 | ||
each on a different trans commit | 14:26 | ||
bartolin | I think the failure from BioPerl6 is something I saw yesterday as well. (But didn't look closely.) | 14:46 | |
m: say "x".trans("y" => "") | |||
camelia | x | ||
bartolin | m: say "x".trans(("y" => "",)) | ||
camelia | MVMArray: Index out of bounds in block <unit> at <tmp> line 1 |
||
Geth | rakudo/main: bcd7beaf7d | (Elizabeth Mattijsen)++ | src/core.c/Str.rakumod Fix multi-needle .trans without replacements Spotted in github.com/rakudo/rakudo/issues/5745. Note that .trans("U" => "", "u" => "") is a very inefficient way to do .trans("Uu" => ""). |
15:35 | |
lizmat | it's about 5x as slow | 15:39 | |
[Coke] | c: HEAD say 2 | 15:42 | |
committable6 | [Coke], ¦HEAD(bcd7bea): «2» | ||
[Coke] | that was quick, lizmat++ | 15:45 | |
Geth | roast: a58a69cc08 | (Elizabeth Mattijsen)++ | S05-transliteration/trans.t Add test for #5745 |
16:03 | |
21:57
sena_kun joined
22:18
guifa joined
22:56
sena_kun left
|