| Geth | nqp: MasterDuke17++ created pull request #880: Reduce $?CLASS lookups |
00:39 | |
|
00:42
timo left
00:55
kjp left
01:08
kjp joined
06:20
kjp left,
hankache_ left,
djinni` left,
finanalyst left,
apogee_ntv left,
hankache_ joined,
apogee_ntv joined
06:21
kjp joined,
djinni` joined
06:24
kjp left,
ShimmerFairy left,
coleman left,
Pixi left
06:25
ShimmerFairy joined,
kjp joined,
Pixi joined
06:27
finanalyst joined
|
|||
| Geth | nqp/main: da5e7771dc | MasterDuke17++ (committed using GitHub Web editor) | src/QRegex/Cursor.nqp Reduce $?CLASS lookups (#880) Looking it up once and putting it in a local decreases the number of times it's looked up via MVM_frame_find_lexical_by_name. Oddly, my example `raku -e 'my %h; my $c = "rakudo/gen/moar/CORE.c.setting".IO; for ^1 { for $c.lines { %h{$0}++ if /^\s+(\d+)/ } }; say "$_.value():$_.key()" for %h.sort(*.value)'` only sees a 2x reduction, when I would expect something like a 5x reduction. |
06:33 | |
|
06:42
coleman joined
|
|||
| Geth | rakudo/main: e77aaaef81 | (Elizabeth Mattijsen)++ | tools/templates/NQP_REVISION Bump NQP to get $?CLASS lookups fix, Masterduke++ |
07:15 | |
| rakudo/lizmat-81: bff083cf6c | (Elizabeth Mattijsen)++ | src/core.c/Any-iterable-methods.rakumod Reduce bytecode size of .are Remove the return constraint, it's clear from the code that it either returns True or a Failure Sadly a "target has a :noinline instruction - ins: takeclosure" is still preventing it from being inlined. Couldn't figure out why. |
08:06 | ||
| rakudo: lizmat++ created pull request #6858: Reduce bytecode size of .are |
|||
| nqp/main: e5e5a905f1 | (Elizabeth Mattijsen)++ | src/QRegex/Cursor.nqp Some more $?CLASS -> $CLASS shortcuts |
09:05 | ||
| rakudo/lizmat-81: fb90dafeb3 | (Elizabeth Mattijsen)++ | tools/templates/NQP_REVISION Bump NQP for more $?CLASS fixups |
09:56 | ||
| lizmat | argh | ||
| Geth | rakudo/main: 50379ea28a | (Elizabeth Mattijsen)++ | tools/templates/NQP_REVISION Bump NQP for more $?CLASS fixups |
||
|
10:54
hankache joined
10:58
hankache_ left
11:41
Pixi left
11:58
Pixi joined
|
|||
| ShimmerFairy | lizmat: created that issue on the array constructor arguments on rakudo's bug tracker. If that's not where you expected me to file the bug (say, you were thinking of it as a spec issue instead), just let me know. | 12:06 | |
| lizmat | thanks, github.com/rakudo/rakudo/issues/6859 is fine | 12:07 | |
|
12:43
hankache left
12:48
hankache joined
14:54
[Coke] left
16:51
[Coke] joined
18:05
samebchase left
18:33
MasterDuke joined
18:59
MasterDuke left
|
|||
| lizmat | bisectable6: class Foo { has @.bar }; say Foo.new(|{:bar([1, 2, 3])}) | 19:50 | |
| bisectable6 | lizmat, Will bisect the whole range automagically because no endpoints were provided, hang tight | ||
| lizmat, ¦6c (105 commits): «Foo.new(bar => [[1, 2, 3],])» | 19:51 | ||
| lizmat, Nothing to bisect! | |||
|
21:05
finanalyst left
|
|||