|
01:23
vendethiel joined
01:46
vendethiel joined
02:03
colomon joined
02:14
vendethiel joined
02:22
colomon joined
02:46
vendethiel joined
02:49
ilbot3 joined
03:10
vendethiel joined
04:10
colomon joined
04:27
vendethiel joined
07:57
FROGGS joined
08:07
zakharyas joined
08:08
kjs_ joined
09:19
kjs_ joined
09:48
rurban_ joined
10:33
kjs_ joined
10:41
donaldh joined
10:43
kjs_ joined
11:29
kjs_ joined
12:55
zakharyas joined
|
|||
| JimmyZ saw jnthn++ said building a spesh graph *as* we trace, is it something a bit like the luajit one? | 14:41 | ||
| so we can avoids many branches? | |||
| *avoid | |||
| jnthn | Traces don't have branches, that's kinda the point... :) | ||
| But yeah, it's along those lines. | 14:42 | ||
| This stuff is *not* on the cards until The Big Three are done, though... | 14:43 | ||
| timotimo | in order to find out the class name of a P6Opaque repr'd thing, i'd have to invoke the "name" method that its HOW provides on the type object, yes? | ||
| JimmyZ | aye, I'm expecting that(Trace spesh) :P | ||
| timotimo | +"from C code" | 14:44 | |
| oh! GTG! | |||
| jnthn | But you can't invoke from C code really, as that's a nested runloop | 14:45 | |
|
14:48
donaldh joined
|
|||
| timotimo | oh, bleh :) | 15:24 | |
| was there at some point a desire to have a field in the P6opaque repr data to hold the name of a class or something like that | 15:25 | ||
| p6opaque repr data could be shared between different classes with the same structure, though | |||
|
15:28
kjs_ joined
16:42
kjs_ joined
17:12
rurban_ joined
18:44
tadzik joined
18:58
mj41 joined
19:09
FROGGS joined
20:08
kjs_ joined
20:44
kjs_ joined
22:04
FROGGS_ joined
22:37
tgt joined
23:02
tgt joined
|
|||