01:11 Ven joined 02:03 colomon joined
timotimo a little profile of hello world on perl6-m shows it spends 6.4% in interp_run, 6.1% in reentrantmutex_lock, 4.3% in reentrantmutex_unlock 09:24
3.8% in string_utf8_decodeand 3.3% in serialization_read_ref
well, a perf profile - meaning it's probabilistic 09:26
but still ...
curious result IMO
JimmyZ I have a basic AA, but I don't know how to get it used :( 09:29
the problem is that I don't do AA between different ins and blocks 09:30
*don't know how to do
timotimo "differens ins"? 09:31
anyway, i'll be AFK for a bit
JimmyZ luajit is by chain 09:32
timotimo maybe a little preprocessor would be interesting that creates C initializer literals or whatever for string constants that we decode from utf8 at run-time at some point
09:38 rurban joined 10:14 Ven joined 10:20 LLamaRider joined 11:40 rurban joined 14:10 Ven joined 14:12 Ven joined 14:20 kjs_ joined 15:49 zakharyas joined 16:30 rurban joined 16:43 Ven joined
japhb timotimo: Did your profile show where those heavy routines were being called *from*? 16:51
TimToady I think they were being called from somewhere in Europe... 16:55
FROGGS_ >.< 17:04
if it starts with 555 you know it is a fake 17:05
japhb Well, mostly: en.wikipedia.org/wiki/555_%28telephone_number%29 17:21
It turns out that not all 555 numbers are fake. I had to deal with that when I wrote a telecom billing system. :-) 17:22
TimToady gotta watch out for those fake fakes... 17:27
japhb You would not believe the uproar if you misbill by even a penny -- and fake numbers are expected to be zero cost -- *even if the phone switch routes them somewhere successfully* 17:29
There are also weird cases like islands changing national identity, areas switching between national and international dialing, and more. It's a deep rabbit hole if you want to get it right and not just punt the hard cases over to customer support .... 17:32
timotimo japhb: perf has some trouble correctly getting the call graphs 17:36
japhb timotimo: Ah, too bad. Would be nice to know what's doing so much locking. (It would be even nicer to know if that locking was contended.) 17:44
18:12 FROGGS__ joined 18:20 ggoebel111111117 joined 18:56 zakharyas joined 21:07 rurban joined 21:22 dalek joined 23:09 kjs_ joined