Parrot 5.6.0 "Psittacosaurus" | parrot.org/ | Log: irclog.perlgeek.de/parrot | #parrotsketch meeting Tuesday 19:30 UTC
Set by moderator on 3 August 2013.
03:43 preflex_ joined 05:49 FROGGS joined 05:58 FROGGS joined 07:08 FROGGS joined 07:22 denis_boyun joined 09:15 denisboyun joined 09:25 FROGGS joined 10:44 denisboyun joined 11:30 denisboyun joined
dalek kudo/nom: a04ebf2 | jnthn++ | src/Perl6/Optimizer.nqp:
Some optimizer code cleanups, commenting.
11:59
kudo/nom: c2ef115 | jnthn++ | src/vm/parrot/guts/bind.c:
Compile-time analysis of sigs with opt params.

This means the optimizer doesn't trip up on protos like ($a?, $b?), meaning we can inline +, < and so forth for natives again.
12:31 denisboyun joined 13:03 kid51 joined 13:33 rurban joined 14:37 Psyche^ joined 14:39 FROGGS joined 15:23 benabik joined 15:26 PacoLinux joined
dalek p: 8e8ae36 | (Tobias Leich)++ | src/vm/parrot/QAST/Compiler.nqp:
we cant trust boxed_primitive when inlinable is false, jnthn++
15:28
15:36 mtk joined 15:59 denisboyun joined
dalek kudo/nom: fe0813a | jnthn++ | src/vm/jvm/runtime/org/perl6/rakudo/ (2 files):
Implement compile time bind analysis on JVM.

Gets compile time multiple dispatch analysis working. For some reason, inlining doesn't happen yet, though.
16:26
kudo/nom: 59e521e | (Elizabeth Mattijsen)++ | src/core/ (2 files):
Make List.(uniq|squish) up to spec
16:53
16:57 rurban joined 18:56 rurban1 joined 19:31 rurban1 joined 20:11 bluescreen joined 20:46 denis_boyun joined 20:59 perlite_ joined 21:02 rurban1 joined
denis_boyun Util: Hi are you here? 21:03
23:27 rurban1 joined 23:31 benabik joined
dalek p: d6f0f6f | jnthn++ | src/QAST/BVal.nqp:
Fix QAST::BVal dumping; ruoso++.
23:36
kudo/nom: 8d851b7 | (Elizabeth Mattijsen)++ | src/core/List.pm:
Reimplement :with for List.(uniq|squish) (keep :as available)

Yes, it was removed from the spec, but the alternative for doing [eqv] based uniqueness using *.perl is 1. still not possible now (because .perl doesn't always give an accurate representation), and 2. very resource intensive, because a whole structure needs to be stringified first before being able to find out that they're different. Whereas [eqv] is very quickly able to find out whether two structures are different.
So, do I believe it belongs in the spec? Yes! :-)
23:37
23:57 benabik joined