00:36 sena_kun left 01:46 Kaiepi left 01:48 Kaiepi joined 04:17 Xliff left 05:52 TreyHarris left 06:03 sena_kun joined 06:39 sena_kun left 07:48 pochi_ joined 07:51 pochi left
lizmat Files=1290, Tests=109646, 207 wallclock secs (28.09 usr 8.28 sys + 2925.77 cusr 258.99 csys = 3221.13 CPU) 09:39
09:40 TreyHarris joined 09:42 discord6 left, timotimo left 09:56 timotimo joined
|Tux| Rakudo version 2019.11-87-g672c5d403 - MoarVM version 2019.07.1-402-g34813e181
csv-ip5xs0.727 - 0.879
csv-ip5xs-206.383 - 6.413
csv-parser21.078 - 21.263
csv-test-xs-200.425 - 0.451
test5.754 - 6.175
test-t1.764 - 1.822
test-t --race0.805 - 0.883
test-t-2029.108 - 29.710
test-t-20 --race9.496 - 9.634
11:47 Guest38485 joined 12:09 lucasb joined, sena_kun joined 13:02 sena_kun left
Geth rakudo: 74c8017cd1 | (Elizabeth Mattijsen)++ | src/core.c/Str.pm6
Re-imagine Str.parse-base from scratch

  - create shortcut for integers
  - don't parse what you don't need to parse
  - use nqp:: ops directly
  - make it between 3x (with decimal point) and 5x (without) as fast
rakudo: MasterDuke17++ created pull request #3338:
Use Visual Studio 2017 in AppVeyor
lizmat waits for AppVeyor results MasterDuke++
13:16 sena_kun joined
MasterDuke wow it's slow. but looking good so far... 13:38
Geth rakudo: 7ecc5bfd92 | (Daniel Green)++ | appveyor.yml
Use Visual Studio 2017 in AppVeyor

It was switched for MoarMV in
rakudo: 7ce4947211 | MasterDuke17++ (committed using GitHub Web editor) | appveyor.yml
Use Visual Studio 2017 in AppVeyor (#3338)

Use Visual Studio 2017 in AppVeyor
rakudo: 3ce924539a | (Elizabeth Mattijsen)++ | src/core.c/Str.pm6
Make Str.parse-base again 1.5x as fast

By bypassing postcircumfix:<[ ]>. Apparently a native list is returned by nqp::radix_I, so we don't need to do any HLL access.
lizmat so that brings it to 4.5x to 7.5x as fast for today
MasterDuke nice 14:11
14:14 MasterDuke left 14:17 MasterDuke joined
nine Nice to see those "x times faster" commits again :) 14:18
MasterDuke yes indeed. now if we can somehow get those for grammars/regexes... 14:22
[Coke] lizmat++
15:02 sena_kun left
Geth rakudo: b42957cc9f | (Elizabeth Mattijsen)++ | src/core.c/Str.pm6
Make Str.Int a few percent faster

  - use self instead of $!value, this appears most of the gain, not sure why
  - rewrite using ternaries instead of nqp::if
   - they codegen the same now, so go for readability
  - on the slow path, also check for Int, so no unnecessary coercion attempt
15:17 sena_kun joined
lizmat nqp::radix lists flag values: github.com/perl6/nqp/blob/master/d...down#radix 15:26
any reason why these are not nqp::const values ?
and where should I add these ?
(assuming we want constants for them) 15:27
15:55 jmerelo joined 16:01 |Tux| left
jmerelo Did you check out today's tmtvl amazing article on the Advent Calendar on stack frame reduction? rakuadventcalendar.wordpress.com/2...reduction/ 16:20
Great stuff about functional programming
Read and/or spread
16:49 robertle joined 17:02 sena_kun left 17:17 sena_kun joined
AlexDaniel lizmat++ # weekly and speedups! 17:36
squashable6: status 17:41
squashable6 AlexDaniel, ⚠🍕 Next SQUASHathon in 2 days and ≈10 hours (2019-12-07 UTC-12⌁UTC+20). See github.com/rakudo/rakudo/wiki/Mont...Squash-Day
Geth rakudo: 37474a7a24 | (Elizabeth Mattijsen)++ | src/core.c/Str.pm6
Re-imagine Str.Numeric

  - use nqp ops were possible
  - check for empty string first
  - makes all numerification a few percent faster except:
  - makes numerification of the empty string about 13x faster
18:16 tellable6 left 18:18 tellable6 joined 18:33 travis-ci joined
travis-ci Rakudo build failed. Elizabeth Mattijsen 'Make Str.Int a few percent faster 18:33
travis-ci.org/rakudo/rakudo/builds/620170051 github.com/rakudo/rakudo/compare/3...2957cc9f2e
18:33 travis-ci left 18:50 Kaiepi left, Kaiepi joined 18:51 Kaiepi left 19:03 sena_kun left
lizmat restarted the one failing job 19:04
19:18 sena_kun joined
lizmat wonders why there is no proto for val() in the setting 19:18
===SORRY!=== 19:36
Failed to parse the zef config file '/Users/liz/Github/rakudo.moar/install/share/perl6/site/resources'
is what I get when following the directions for installing Inline::Perl5 for spectest 19:37
Kaiepi ^^
Geth rakudo: 2652d9c59b | (Elizabeth Mattijsen)++ | src/core.c/allomorphs.pm6
Sanitize val()'s special cases a bit

  - let dispatch handle the difference between List and Slip
  - show any Mu value, specifically IterationEnd (which may help in debugging)
  - simplify reporting on other single values
19:39 jmerelo left 19:46 MasterDuke left 19:55 squashable6 left 19:57 squashable6 joined 20:02 squashable6 left, MasterDuke joined 20:04 squashable6 joined 20:28 lucasb left 21:02 sena_kun left 21:17 sena_kun joined
Geth rakudo: 6c85e69101 | (Elizabeth Mattijsen)++ | src/core.c/Str.pm6
Abstract partse-base Failure into a method

  - making size of other methods smaller, so more likely to inline
  - add some comments
21:57 Kaiepi joined
Kaiepi hey guys, so you know that pattern you sometimes see in rakudo where a class has a parameterize metamethod that mixes in a paremetric role that just adds a method returning its parameter? 22:00
m: class Named { method ^parameterize(Named:U ::T, Str:D $name) { T.^add_method: 'name', anon method name(T: --> Str:D) { $name }; T } }; class Foo is Named['foo'] { }; say Foo.name 22:01
camelia foo
Kaiepi yeah, you don't need to mixin the role for that
at least, i don't think it's needed 22:02
nvm, just tested doing that with two classes doing that and it doesn't work
still, i wonder if there are ways to avoid needing to make a mixin for each parameterization of a parametric class 22:07
lizmat m: my %h = a => slip(1,2,3); .say for %h<a> # I wonder if this is intended behaviour, or an artefact, jnthn? 22:20
camelia 1
lizmat afk& 22:21
vrurg Does anybody knows if it's possible to connect comma to an external debug session? 22:34
sena_kun vrurg: it is not. 22:35
vrurg :(
sena_kun vrurg: we surely should wishlist it when the time for debugger fixes comes
though I am aware of some more "interesting" issues than this feature missing. :( 22:36
vrurg Perhaps there is a plugin for emulating a terminal in debug session?
sena_kun is confused
vrurg pardon... an ANSI terminal
sena_kun you mean REPL of debugger as timotimo++ module does or?
vrurg No, vt* emulation. ANSI, vt100, etc. 22:37
sena_kun no idea then, sorry
vrurg I'm writing a simplistic text-UI framework. 22:38
Exactly the case where debug prints are not really useful and I hoped for some debugger support. :) 22:39
sena_kun yeah, that's the case... maybe something more sophisticated than prints, e.g. modules.raku.org/dist/LogP6:cpan:ATROXAPER may help? 22:42
22:42 lucasb joined
vrurg Maybe... It's a bit overcomplicated for a simple task though. 22:55
23:02 sena_kun left 23:03 sena_kun joined, sena_kun left 23:24 Xliff joined
Xliff . 23:24
m: .say for DateTime.new('2019-12-12T12:00:00Z') ... DateTime.new('2019-12-12T15:00:00Z') 23:27
camelia 2019-12-12T12:00:00Z
No such method 'succ' for invocant of type 'DateTime'. Did you mean any of these?

in block <unit> at <tmp> line 1
Xliff m: .say for DateTime.new('2019-12-12T12:00:00Z') ... DateTime.new('2019-12-12T12:00:00Z')
camelia (timeout)2019-12-12T12:00:00Z
Xliff AlexDaniel: I've been saying I'd write the damned thing if someone would give me specs. 23:38
I'm sick of ppl asking "What web framework" when the tools are there to develop one.
Funny... no one seems to want to take me up on it. :p
AlexDaniel Xliff: just take a web framework for other language and do the same but in raku style? 23:39
Xliff Um. No.
AlexDaniel so you want someone to do it from scratch, or what?
Xliff First and foremost, you really can't do that. Unconsciously, when you port, you do something using the original languages paradigm. 23:40
Kinda like confirmation bias.
AlexDaniel no but it gives you a list of objectives for sure
Xliff With a spec, you adjust what you do based on the language you have chosen.
Um. No.
Again. I don't want to port. 23:41
AlexDaniel I'm not telling you to port, I'm saying there are existing projects that get the job done, and you can see what they do/can do
Xliff Do you know how horrible things would be if I tried to write specs from WordPress code?
And I'm telling you that I don't want to go through that process.
I shouldn't have to.
If you know what a "decent web framework" does, you can write it down and pass it to someone to create. 23:42
I just don't have the bandwidth to do everything. Much of it is consumed with other projects (which you can guess... :p)
.oO( GIMP bindings... yum! )
AlexDaniel weekly: Xliff++ wants to create a Raku web framework. Help from others is needed to create an initial list of objectives/spec. colabti.org/irclogger/irclogger_lo...12-03#l177 23:44
notable6 AlexDaniel, Noted! (weekly)
Xliff Oh. You. BASTARD! ;p 23:51