timo i've gotta take a closer look at this r3tool 00:48
08:45 sena_kun joined
Geth roast: 3a463fe779 | (Elizabeth Mattijsen)++ | S32-list/reverse.t
Add test for #3595
 09:31
rakudo/main: ce5e02cd20 | (Elizabeth Mattijsen)++ | src/core.c/Range.rakumod
Make Inf .. xxx produce Nils

Fixes #3596		 09:44
roast: 2010266f68 | (Elizabeth Mattijsen)++ | S02-types/range.t
Tests for #3596
roast: 8d5e3cd1b3 | (Elizabeth Mattijsen)++ | S06-advanced/dispatching.t
Add test for #3611
 10:06
rakudo/main: dd62b98e09 | (Elizabeth Mattijsen)++ | t/12-rakuast/xx-fixed-in-rakuast.rakutest
Add test for #3626
 10:28
rakudo/lizmat-uniname: 5f71b3a253 | (Elizabeth Mattijsen)++ | src/core.c/unicodey.rakumod
Make Int.uniname return a Failure on unassigned codepoints

As discussed in github.com/rakudo/rakudo/issues/3636
This however breaks tests in t/spec/S15-unicode-information/uniname.t so a decision needs to be made.		 10:48
rakudo: lizmat++ created pull request #5658:
Make Int.uniname return a Failure on unassigned codepoints
rakudo/main: 258a1b54bc | (Elizabeth Mattijsen)++ | src/core.c/Range.rakumod
Make +((5..2).reverse) produce correct value

Namely 0
In response to github.com/rakudo/rakudo/issues/3637		 11:09
¦ rakudo: lizmat self-assigned `printf` doesn't honor `%f` widths github.com/rakudo/rakudo/issues/3733 12:42
rakudo/main: 6adcecd939 | (Elizabeth Mattijsen)++ | src/core.c/Rational.rakumod
Make sure the numerator is shown on N / 0
 13:09
14:06 notna joined 14:12 bartolin joined 16:18 notna left
Geth rakudo/main: 53a5ba759c | (Elizabeth Mattijsen)++ | src/core.c/Rational.rakumod
Centralize creation of 0 denominator Rats

To make it easier to add additional logic		 17:03
bartolin hi, #raku-dev! I haven't been around for a long time. Seeing the flood of notifications about closed issues (lizmat++) made me think I should try to get involved again 17:33
one conrete question: on my vm (FreeBSD 14.1) integration/advent2012-day21.t hogged up all memory very fast. is that a known problem? (it's part of the make target "stresstest") 17:37
timo oh hai bartolin 17:45
bartolin hej timo. I noticed your commit that sped up raku-ast-compiler.nqp for the jvm backend. very, very nice :) 17:48
timo yeah that was fun to investigate 17:49
bartolin timo++ (also for caring for the package for Debian)
timo <3
unfortunately, the ram usage for building rakudo-jvm is a bit too much for the default azure vms we get on our pipelines, so we can't really get to testing that 17:58
bartolin yeah, I noticed that the Xmx for the build had to be increased again. the ram usage was one of the many things I wanted to give a look. but I don't have relevant experience with profiling, so it never made it to the top of my list 18:07
Geth rakudo/lizmat-saneRats: 82e0cfb01c | (Elizabeth Mattijsen)++ | 2 files
Make Rationals remember where they divided by 0

Inspired by github.com/rakudo/rakudo/issues/3738
This stores a handled Failure with the location where the Rational was created, and throws that whenever it is being stringified.
... (14 more lines)		 18:40
rakudo: lizmat++ created pull request #5659:
Make Rationals remember where they divided by 0
roast: a6983a6d03 | (Elizabeth Mattijsen)++ | 2 files
Add test for #3777
 18:57
[Coke] R#3702 was closed as completed, but I don't see a merged PR or commit that addressed it? 19:01
linkable6 R#3702 [closed]: github.com/rakudo/rakudo/issues/3702 [CLI] USAGE statements should be formatted by default
[Coke] tbrowder gave it a thumbs up... 19:02
lizmat I closed it in favour of #4223 19:04
R#4223
linkable6 R#4223 [open]: github.com/rakudo/rakudo/issues/4223 [CLI] --help output for scripts with subcommands
[Coke] ah, thanks 19:22
21:05 lizmat left 21:07 lizmat joined 21:30 MasterDuke joined
MasterDuke m: say (reduce * * *, (1993 xx 10_000)) %% 2 21:33
camelia False
MasterDuke m: say (reduce * * *, (1 .. 10_000)) %% 2 21:34
camelia True
MasterDuke those have interesting profiles. they take pretty much the same amount of time (it's mostly spent in the multiplication), but the allocations are wildly different 21:36
the `xx` version has 19985 Ints as the top allocated and then 1144 BOOTHashes. the `..` version has 103380 BOOTHashes and then 69920 Ints and then 41314 Scalars 21:38
`xx` version: In total, 8181 call frames were entered and exited by the profiled code. Inlining eliminated the need to create 26738 call frames (that's 76.57%).
the `..` version: In total, 208319 call frames were entered and exited by the profiled code. Inlining eliminated the need to create 68261 call frames (that's 24.68%). 21:39
the optimized qast doesn't look terribly different 21:45
those BOOTHashes in the `..` version are mostly coming from 21:48
reify-at-least, SETTING::src/core.c/List.rakumod:40, 50005
push-exactly, SETTING::src/core.c/Iterator.rakumod:28,40004
push-exactly, SETTING::src/core.c/Rakudo/Iterator.rakumod:2137, 10001
22:35 guest555 joined
Geth rakudo/main: 9b31d7c66a | (Elizabeth Mattijsen)++ | t/12-rakuast/xx-fixed-in-rakuast.rakutest
Add test for #3786
 22:35
roast: 4bb9f2f2f7 | (Elizabeth Mattijsen)++ | S32-num/rat.t
Remove debug code

Spotted by bartolin++		 22:48
tbrowder please remind me how to run tests and skip the #? tests 22:54
23:07 guest555 left 23:08 [Coke] left
timo fudgeandrun is responsible for implementing #? 23:11
what do you need exactly?
"make spectest" will `fudgeall` for you before running the tests 23:12
23:13 [Coke] joined
tbrowder i was just trying to run the two i’m working on. 23:40
but i can run them all, thnx
timo "make t/spec/bla/foo.t" should also fudge and run it 23:47
tbrowder ah, thnx, that’s what i was looking for, haven’t done that in a long time! 23:50

Powered by Cro and the Raku® Programming Language.

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

Thank you!