Parrot 4.6.0 "Wild Parrots of Telegraph Hill" | parrot.org/ | Log: irclog.perlgeek.de/parrot | #parrotsketch meeting Tuesday 19:30 UTC
Set by moderator on 18 July 2012.
01:00 wagle joined 02:09 woosley joined 02:36 Psyche^ joined 02:45 benabik joined 05:06 Khisanth joined 05:43 plobsing joined, wagle joined 06:41 fperrad joined 07:12 brrt joined 07:35 muixirt joined
muixirt good morning 07:35
any parrot internals wizards awake? 07:36
07:49 muixirt_ joined 07:51 muixirt__ joined 07:54 muixirt__ joined
sorear muixirt: try your question, I'm not a wizard but I know the oddest things sometimes 08:05
muixirt hi sorear 08:06
well, I wonder how opcode dispatch works these days
is the "fast" runcore the default?
sorear I think the fast runcore is the default and most/all of the others were scrapped 08:07
08:08 GeJ joined
muixirt looks at runcore/cores.c : runops_fast_core 08:10
I wonder this line is good for: Parrot_pcc_set_pc(interp, CURRENT_CONTEXT(interp), pc); 08:11
I commented it out to see what happens, but nothing spectacular, so probably this funcion isn't executed 08:15
oh well, my C skills are practically inexistent :( 08:16
sorear: this loop and the dispatch method are not especially fast (?) 08:20
sorear It's called the fast core because we used to have a slow core that did bounds checking. 08:21
It's _not_ particularly fast.
muixirt sorear: how does dynops work? Do they switch the interp->code->op_func_table or is there only one op_func_table and customs ops get appended? 08:35
09:35 brrt left 09:48 brrt1 joined 12:54 PacoAir joined 13:16 bluescreen joined 13:28 preflex_ joined 13:58 JimmyZ joined 14:49 awwaiid joined 14:51 bluescreen joined 15:23 dmalcolm joined 16:55 tuxit joined 17:06 PacoAir joined 17:16 darbelo joined 17:30 rurban joined 17:34 benabik joined 17:51 jashwanth joined
benabik ~~ 17:57
Oh joy. "make: command not found" 19:04
Curse you Apple
Updating Xcode apparently deleted the command line tools. 19:06
19:06 dngor joined
dalek kudo/nom: 08c72b7 | jnthn++ | src/Perl6/Actions.pm:
Fix --doc (moritz++).
21:26
p: f0c85fd | jnthn++ | / (2 files):
Give NQP support for mixins.
p: fe87905 | jnthn++ | / (3 files):
Move arity and flat out to a role that we mix in on the handful of QAST nodes that need it. Shaves 16 bytes (on 64-bit) off the vast majority of nodes.
p: ffe1cba | jnthn++ | src/QAST/ (2 files):
The childorder attribute only applies to QAST::Op, so move it there; another 8 bytes of many nodes.
alvis msg Coke I apologize for the delay in replying to your offer. (It's just been rather hectic and not likely to let up anytime soon.) Anyway, I need priveleges to parrotdev.blogspot.com such that I can archive the site and, then, after storing it somewhere on parrot.org, delete it. Or leave it to someone else to delete; either way is fine with me. 22:04
msg Coke That said, the earliest I can get to is, depending on several unknown factors, mid-to-late next week. So, there's no real rush. I'll be away for the next several days, btw. 22:05
dalek p: 26a8bc1 | jnthn++ | src/QAST/ (4 files):
Sort out compile time value handling, so Rakudo can do it without using annotations.
kudo/nom: c1747df | jnthn++ | tools/build/NQP_REVISION:
Bump NQP_REVISION for memory usage reductions and better compile time value support.
22:07
kudo/nom: f23a454 | jnthn++ | src/Perl6/ (6 files):
Refactor to use the improved compile time value support, eliminating many uses of annotations.
22:13 preflex_ joined 22:53 kid51 joined 23:49 kid51 joined