|
02:48
ilbot3 joined
03:47
vendethiel joined
04:25
ggoebel17 joined
04:30
harrow joined
07:02
vendethiel joined
07:15
domidumont joined
07:20
domidumont joined
08:13
zakharyas joined
09:50
domidumont joined
09:57
zakharyas joined
10:56
lizmat joined
17:10
harrow joined
17:16
domidumont joined
18:26
lizmat joined
18:49
vendethiel joined
19:22
lizmat joined
19:30
vendethiel joined
19:45
lizmat joined
19:48
FROGGS joined
19:52
lizmat_ joined
20:28
brrt joined
20:45
vendethiel joined
|
|||
| brrt | jnthn: note that get_string is a macro that accesses cu->body.strings directly | 20:47 | |
| in the jit i mean | |||
| timotimo | well, that could cause trouble :) | 20:48 | |
| brrt | making that a function may be difficult | 20:50 | |
| not every place that uses it, is register-clean | 20:53 | ||
|
21:01
domidumont joined
|
|||
| brrt | oh, btw, JSON::Fast breaks under even-moar-jit | 21:04 | |
| i'm intending to build some kind of bisecting tool | 21:05 | ||
| to pass env vars, 'don't compile after seq nr $x', and 'don't compile basic blocks beyond $y with expressions' | |||
| should be easy to do and pinpoint the breakage | 21:06 | ||
| anyway, off i am | 21:13 | ||
|
21:45
vendethiel joined
|
|||
| jnthn | brrt: Yes, that's on my todo list already...I'll just ask for the string while JIT-compiling, which means it'll certainly be available. :) | 22:02 | |
| Thanks for spotting, though :) | |||
| timotimo | oh, right, those are just the at-compile-time-fixed strings | 22:16 | |
| jnthn | aye | 22:18 | |
| timotimo | in that case, doing it at jit time is absolutely fine | 22:20 | |
| the jit would be rather happy having the strings just as a pointer | 22:21 | ||
| jnthn | That's how it JIT's em usually, I think :) | 22:32 | |
| jnthn continues on his way for sleep :) | 22:33 | ||
| lizmat | gnight jnthn | ||
| jnthn | o/ | ||