|
00:55
vendethiel joined
02:26
nebuchadnezzar joined
04:35
vendethiel joined
05:59
Peter_R joined
06:14
rurban joined
06:50
rurban joined
07:02
rurban joined
07:15
rurban joined
07:52
lizmat joined
08:16
vendethiel joined
08:55
FROGGS joined
|
|||
| FROGGS | o/ | 08:56 | |
|
10:11
tgt joined
10:13
colomon joined
10:55
FROGGS joined
11:30
TimToady joined
12:09
Ven joined
12:25
brrt joined
|
|||
| brrt | \o | 12:25 | |
| register allocation is hard | 12:27 | ||
| what if i'm too dumb for all this? | |||
| then i'll still have to find a solution :-) | 12:30 | ||
| maybe it's better to say | |||
| what if my earlier ideas were wrong | |||
| one of these ideas: 'you can do everything you want using generic tree traversal | 12:37 | ||
| i'm fairly sure i can use the SSA calculations from spesh in some wy | 12:55 | ||
| just don't know how yet | |||
| hmm..i'll be back later | 13:35 | ||
|
13:35
brrt left
|
|||
| cognominal | what is a PHI ops? | 16:00 | |
| got it : fr.wikipedia.org/wiki/Static_singl...nment_form | 16:01 | ||
| jnthn | .oO( what, not "forme d'assignment static single"? :) ) |
16:21 | |
| dalek | arVM/even-moar-jit: e1a0e20 | brrt++ | src/jit/expr (2 files): Implement sp_p6ogetvt_o as expression template This required the addition of a few new jit ops, notably do (for a sequence of statements yielding the last), all (logical and of a series) and any (logical or). and and or become their corresponding binary operations. I note that I have not found some way to deal with operand types (i.e. size) yet, although the sizes are explicit in loads and stores. |
||
| jnthn is gonna have to learn this sexpr language :) | 16:32 | ||
|
16:34
harrow joined
|
|||
| cognominal | jnthn, I am not sure that has even been translated. That could be : forme statique d'affectation unique | 16:59 | |
| * ever | 17:00 | ||
|
21:50
zakharyas joined
22:28
TEttinger joined
|
|||