Perl 6 language and compiler development 2.0 | Logs at irclog.perlgeek.de/perl6-dev/today | For toolchain/installation stuff see #perl6-toolchain | For MoarVM see #moarvm
Set by moderator on 8 May 2018.
00:59 |Tux| joined 01:58 ilbot3 joined
moderator Perl 6 language and compiler development 2.0 | Logs at irclog.perlgeek.de/perl6-dev/today | For toolchain/installation stuff see #perl6-toolchain | For MoarVM see #moarvm
Geth rakudo: 755627cb28 | skids++ | t/02-rakudo/repl.t
Add test for RT#123380 (REPL tests go in rakudo tree)
 02:14
rakudo: 16bc0cce97 | (Zoffix Znet)++ (committed using GitHub Web editor) | t/02-rakudo/repl.t
Merge pull request #1806 from skids/rt123380test

Add test for RT#123380 (REPL tests go in rakudo tree)
synopsebot RT#123380 [open]: rt.perl.org/Ticket/Display.html?id=123380 [BUG] A script that gives different results when executed from a file and when pasted into REPL
02:38 ufobat___ joined
Geth rakudo: thundergnat++ created pull request #1807:
Add fast .Str multi for FatRat, split Rational.Str into Rat/FatRat Str		 03:07
roast: 9d50144e74 | skids++ | S03-operators/reduce-le1arg.t
unskip passing test to close RT #124552

At some point this started to work
synopsebot RT#124552 [new]: rt.perl.org/Ticket/Display.html?id=124552 Roast rakudo skip/todo test:./S03-operators/reduce-le1arg.t line:43 reason: 'expected Any but got Mu instead'
05:19 geekosaur joined 05:20 geekosaur joined
[Tux] Rakudo version 2018.04.1-59-g16bc0cce9 - MoarVM version 2018.04.1-90-g7a521eed4
csv-ip5xs0.887 - 0.976
csv-ip5xs-208.799 - 8.999
csv-parser36.789 - 36.889
csv-test-xs-200.453 - 0.459
test8.809 - 8.989
test-t2.335 - 2.360
test-t --race0.977 - 1.011
test-t-2042.297 - 42.316
test-t-20 --race14.594 - 15.101
 06:58
yoleaux 7 May 2018 23:45Z <MasterDuke> [Tux]: could you generate some graphics for test-t-20 and test-t-20 --race ? i feel like those haven't gotten much faster, which is odd, cause i'd expect it to be more visible with them
MasterDuke, tux.nl/Talks/CSV6/speed4r.html (again)
07:13 brrt joined
lizmat Files=1239, Tests=76314, 316 wallclock secs (15.58 usr 5.57 sys + 2176.66 cusr 223.50 csys = 2421.31 CPU) 07:51
brrt lizmat: have you tried the jit-stack-walker on OS X 07:55
lizmat brrt: no, I haven't
brrt could you please :-)
I'm thinking of making it a compile time option 07:56
lizmat ok, then I don't understand what it does
brrt it is a branch :-) 07:57
what it does is:
currently every basic block in the JIT *has* to store the current position in the current thread frame 07:58
because otherwise the (relatively rare) exception mechanism and dynamic lexical mechanism don't know which inline / frame handler we are in, and fail 08:03
i.e. we *eagerly* load the current position 08:04
the idea of the jit-stack-walker branch is to make that less eager
lizmat ok, so if there are no exceptions to handle, that's a win 08:05
brrt correct
lizmat does "return" count as an exception ?
brrt on rakudo, yes
well, in some cases 08:06
lizmat ok, so this would penalize everybody who uses explicit "return" in their code
when not necessary
brrt not necessarily
i mean, stack walking is a pretty cheap operation
given that the first few pages of the stack are already on cache
and every basic block entry is quite a bit more frequent than frame return 08:07
lizmat ok, was just checking
brb
brrt (what it does is inspect the stack for the return address in the compiled JIT code, so we *know* the current position exactly) 08:08
lizmat have to ba afk for bit& 08:15
Geth nqp: 474ff3cb87 | (Elizabeth Mattijsen)++ | tools/build/MOAR_REVISION
Bump Moar to get more jitted opcodes
 10:25
¦ nqp: version bump brought these changes: github.com/MoarVM/MoarVM/compare/2...3-g372d058
10:47 brrt joined 10:54 AlexDaniel joined 11:05 AlexDani` joined 11:25 Kaiepi joined
Geth rakudo: a16eeac7b0 | (Elizabeth Mattijsen)++ | tools/build/NQP_REVISION
Bump NQP to get the latest JITted opcodes
 12:05
rakudo: version bump brought these changes: github.com/perl6/nqp/compare/2018....5-g474ff3c
ee1ba1536a | (Elizabeth Mattijsen)++ | 2 files

Based on (my @a = ^10).splice(5). Larger return values will be faster still, smaller will see it approach to noise levels.
AlexDaniel is the failure in t/09-moar/00-misc.t known? 12:07
oh, actually I'm probably doing something wrong… 12:08
AlexDaniel retries
lizmat AlexDaniel: I see it also :-( (now) 12:11
AlexDaniel yeah 12:22
12:22 |Tux| joined
Geth rakudo: 4ab2f2cd73 | (Elizabeth Mattijsen)++ | 2 files
Make native array.splice(offset,size) about 25% faster

Based on (my @a = ^10).splice(3,5). Larger return values will be faster still, smaller will see it approach to noise levels.		 12:46
13:03 FROGGS joined
Geth rakudo: da646aa2ee | (Elizabeth Mattijsen)++ | 2 files
Extract CLONE_SLICE to none-type specific part

  - because with nqp::slice now we can
  - saves about 5K from the .moarvm file		 13:05
13:36 skids joined 13:43 travis-ci joined
travis-ci Rakudo build errored. Elizabeth Mattijsen 'Make native array.splice(offset,size) about 25% faster 13:43
travis-ci.org/rakudo/rakudo/builds/376818027 github.com/rakudo/rakudo/compare/e...b2f2cd73af
13:43 travis-ci left
buggable [travis build above] ✓ All failures are due to: timeout (3 failures). 13:43
14:26 travis-ci joined
travis-ci Rakudo build errored. Elizabeth Mattijsen 'Extract CLONE_SLICE to none-type specific part 14:26
travis-ci.org/rakudo/rakudo/builds/376825119 github.com/rakudo/rakudo/compare/4...646aa2ee6e
14:26 travis-ci left
buggable [travis build above] ✓ All failures are due to: timeout (3 failures). 14:26
Geth rakudo: e0ad668e22 | (Elizabeth Mattijsen)++ | src/core/native_array.pm6
Fix native array slices with containerized Range

It was ignoring the containerizations of the range, which made it act differently from the non-native array slice with Range.		 15:12
roast: 94f8510908 | (Elizabeth Mattijsen)++ | 3 files
Add tests for native array slices with containerized Range
 15:21
16:02 travis-ci joined
travis-ci Rakudo build passed. Elizabeth Mattijsen 'Fix native array slices with containerized Range 16:02
travis-ci.org/rakudo/rakudo/builds/376878858 github.com/rakudo/rakudo/compare/d...ad668e228d
16:02 travis-ci left
Geth nqp/truffle: fca59795e4 | (Paweł Murias)++ | 7 files
[jvm] Build with the truffle api included

It's not used for anything just yet		 16:21
nqp/truffle: 10d02e3104 | (Paweł Murias)++ | 10 files
[jvm] Build and run the simplest truffle node
16:31 brrt joined 19:20 robertle joined
Geth nqp: 291bf8a249 | (Elizabeth Mattijsen)++ | tools/build/MOAR_REVISION
Bump Moar to revert inlining of ctxlexpad
 20:06
¦ nqp: version bump brought these changes: github.com/MoarVM/MoarVM/compare/2...4-gcbf26fa
rakudo: c66a6f0ec5 | (Elizabeth Mattijsen)++ | tools/build/NQP_REVISION
Bump NQP to hopefully fix #1810
 20:21
¦ rakudo: version bump brought these changes: github.com/perl6/nqp/compare/2018....6-g291bf8a
synopsebot RAKUDO#1810 [open]: github.com/rakudo/rakudo/issues/1810 Recent regression "ctxlexpad needs an MVMContext"
MasterDuke [Tux]: hm. tux.nl/Talks/CSV6/speed4r.html does show a definite improvement. what about the --20 versions? 20:22
21:04 AlexDaniel joined
Geth 6.d-prep: 88100707be | (Zoffix Znet)++ (committed using GitHub Web editor) | TODO/FEATURES.md
List Complex division for IEEE-fication along with Nums
 22:44
22:47 Kaiepi joined 22:49 skids joined
Geth roast: bdf67b698b | skids++ | S06-other/main-usage.t
Add tests for RT#126532
 23:45
synopsebot RT#126532 [new]: rt.perl.org/Ticket/Display.html?id=126532 [BUG] Internal error when passing whitespace through a named argument in Rakudo
23:58 |Tux| joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!