jnthn Just a quick heads up in case I don't get to mention it here tomorrow: I'm heading away for a weekend birthday break, and won't be dragging a laptop along with me. I'll be back here on Monday. :-) 01:17
'night o/
timotimo \o/ 01:18
Geth nqp: jstuder-gh++ created pull request #402:
Rakudo Issue 1304: [jvm] Allow 'chain' op to use child as callee		 06:47
rakudo: jstuder-gh++ created pull request #1448:
Issue 1304 Continued: Re-enable negated chaining on the JVM		 06:50
roast: jstuder-gh++ created pull request #381:
Rakudo Issue 1304: Unfudge now passing negated chain tests on JVM		 06:53
nqp: 84278100ac | (Jeremy Studer)++ | src/vm/jvm/QAST/Compiler.nqp
[jvm] Allow 'chain' op to use child as callee

Modify the 'chain' op to allow the option to use the first child as the callee. Before the name of the op served as the operator sub and the children the operands. This modification makes it so that, if there is no name provided to the chain, child 0 serves as the operator and children 1 and 2 the operands. ... (5 more lines)		 07:02
nqp: cd03234e17 | (Zoffix Znet)++ (committed using GitHub Web editor) | src/vm/jvm/QAST/Compiler.nqp
Merge pull request #402 from jstuder-gh/jvm_neg_chaining

Rakudo Issue 1304: [jvm] Allow 'chain' op to use child as callee
rakudo: e72ec51817 | (Zoffix Znet)++ | tools/build/NQP_REVISION
[NQP Bump] Brings 3 commits

NQP bump brought: github.com/perl6/nqp/compare/2018....7-gcd03234 cd03234 Merge pull request #402 from jstuder-gh/jvm_neg_chaining 8427810 [jvm] Allow 'chain' op to use child as callee d71bd73 Fix typo		 07:11
rakudo: version bump brought these changes: github.com/perl6/nqp/compare/2018....7-gcd03234
1f22319016 | (Jeremy Studer)++ | src/Perl6/Actions.nqp

This statement was previously restricted to MoarVM only as the JVM's chain op was not yet set up to handle a child as the callee.
Now, as of commit 84278100acbf589a37f5baa85ffe27d68de52d13 (in the NQP repo), the chain op's JVM implementation works the same way that MoarVM's does, making this negated chaining implementation functional on the JVM backend as well.
  See <github.com/rakudo/rakudo/issues/1304>.
rakudo: 0c1191f6f8 | (Zoffix Znet)++ (committed using GitHub Web editor) | src/Perl6/Actions.nqp
Merge pull request #1448 from jstuder-gh/jvm_neg_chaining

Issue 1304 Continued: Re-enable negated chaining on the JVM		 07:12
roast: d4deb15487 | (Jeremy Studer)++ | S03-operators/relational.t
Unfudge now passing negated chain tests on JVM

  See <github.com/rakudo/rakudo/issues/1304>.
roast: 910f2dfbe8 | (Zoffix Znet)++ (committed using GitHub Web editor) | S03-operators/relational.t
Merge pull request #381 from jstuder-gh/jvm_neg_chaining

Rakudo Issue 1304: Unfudge now passing negated chain tests on JVM
lizmat Files=1233, Tests=76124, 311 wallclock secs (14.81 usr 5.14 sys + 2149.22 cusr 201.72 csys = 2370.89 CPU) 08:57
that's 28 seconds less than my last run, I believe, nine++
AlexDaniel I'm assuming that it will speed up the toaster as well 10:57
lizmat AlexDaniel: it probably will 10:58
afk for rest of day&
[Tux] WHAT? .... :) 12:45
Rakudo version 2018.01-40-g0c1191f6f - MoarVM version 2018.01-21-gd3ce626b4
csv-ip5xs1.033 - 1.062
csv-ip5xs-2011.102 - 11.298
csv-parser11.877 - 12.589
csv-test-xs-200.424 - 0.459
test10.615 - 10.753
test-t2.537 - 2.572
test-t --race1.093 - 1.116
test-t-2046.291 - 46.658
test-t-20 --race16.238 - 16.859
2018-01-26 15:13:38 test-t 2.687
2018-01-19 13:33:23 test-t 2.683
2018-01-26 15:16:48 test-t 2.608
2018-01-27 13:39:21 test-t 2.572
2018-01-27 13:42:33 test-t 2.537
result twittered 12:47
timotimo here's to another 100x 13:13
DrForr +2 13:16
(1 for the 100x, 1 for yay in general!) 13:17
timotimo m: say 16.238 / 46.291 13:18
camelia 0.350781
timotimo that's not so bad; [Tux], you're running the --race ones on a 4-core machine?
[Tux] Linux 4.4.104-39-default [openSUSE Leap 42.3] HP Z220CMT Xeon(R) CPU E3-1245 V2 @ 3.40GHz/1965(8) x86_64 11889 Mb 13:22
timotimo i'm not sure, is the (8) the number of cores there? :| 13:26
[Tux] www.cpubenchmark.net/cpu.php?cpu=I...mp;id=1191 13:37
No of Cores: 4 (2 logical cores per physical)
Linux shows 8 CPU's
timotimo ah, i wonder what it'll use for the degree in race in that case 13:38
Geth rakudo: b188cc8265 | (Stefan Seifert)++ | 2 files
Speed up resolve-unknown-repos

No need to turn the repo-chain into an array just to traverse it.
timotimo because for 8 cores, 3x speedup isn't so great any more :)
Geth rakudo: 838beab9d9 | (Stefan Seifert)++ | 4 files
Avoid unnecessary check of PrecompilationId's validity

In many cases we can trust that the precompilation id is already valid,
  because it's derived from the compiler's id or is generated by way of sha1
or comes from the name of a file named after a precompilation id. So avoid the costly regex parsing of the id in those cases and just trust the value.		 13:46
rakudo: baa1393b6d | (Stefan Seifert)++ | src/vm/moar/ops/container.c
Clean up handler successors after throwish ops become non-throwing

Mirrors MoarVM commit f80d610cc4		 13:58
rakudo: 3ed2fbd550 | (Stefan Seifert)++ | lib/NativeCall.pm6
Improve dynamic optimization of nativecast

Using multi candidates for the Int and Num special cases, the dynamic optimzier will be able to remove a type check altogethr in some cases.		 14:01
nine Oh, this is crazy. We can safe another 15ms with this added to src/core/IO/Handle.pm: my class buf8 does Buf[uint8] is repr('VMArray') { } 15:15
Because it turns out, almost all of the time spent slurping a source file to calculate its checksum is wasted on punning the buf8 role. 15:16
Oooh...a much less intrusive and therefor better way is to add a "buf8.new;" to Buf.pm. This way we create the punned version at compile time. 15:43
timotimo nice. 16:15
Geth rakudo: db1e067e3d | (Stefan Seifert)++ | src/core/Buf.pm
Reduce startup overhead of file I/O and module loading a bit

Punning the buf8 role costs a couple of milliseconds when loading modules. We can move this overhead to compile time of the setting instead.
rakudo: a9a9e1c97e | (Stefan Seifert)++ | src/core/CompUnit/Repository/FileSystem.pm
Safe another millisecond by avoiding .repo-chain

We have better and faster ways to iterate of a linked list		 16:28
nine Why don't we initialize $*PERL when compiling the setting? Seems like it cannot change afterwards anyway as all its properties are determined at the time of the settings compilation 16:37
AlexDaniel nine++ # awesome work 17:07
nine Initializing $*PERL during setting compilation is harder than it sounds. But that's just because it throws some exception and I can't find a way to get at the message. 17:36
"An exception occurred while evaluating a BEGIN" but which one?
mst nine: apply time traveling debugger? 17:42
cygx o/ 23:36
nine: install-core-dist.pl fails with "No such method 'prefix' for invocant of type 'Str'" on my MinGW build 23:37
nine: could that be related to your recent commits?

Powered by Cro and the Raku® Programming Language.

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

Thank you!