01:48
ilbot3 joined
08:04
FROGGS joined
10:15
colomon joined
10:58
avar joined
11:06
vendethiel joined
12:21
brrt joined
|
|||
brrt | good * | 12:22 | |
patches coming up | |||
dalek | arVM/even-moar-jit: 6a6fcc0 | brrt++ | src/jit/ (5 files): Move expression tree walking code May be useful in other contexts (such as optimization) |
||
arVM/even-moar-jit: 9a6ca81 | brrt++ | src/jit/ (4 files): Reduce register allocator surface area We will now need way to specify per-tile register requirements, as tiles can no longer try and 'fix' their registers at runtime. We don't actually seem to hit the NYI paths, though. |
|||
arVM/even-moar-jit: fcc270b | brrt++ | src/ (12 files): Replace ExprValue with ValueDescrioptor Internalize RegisterAllocator structure The register allocator only contains temporary structures, and as such does not need to be 'public'. |
|||
brrt | i should probably blog about that | 12:24 | |
the upshot is | |||
we have now the valuedescriptor structure | |||
instead of 'expression values' | 12:25 | ||
since, as you may know, there may be more than one of those | |||
timotimo | yeah | 12:31 | |
brrt | now i need to write the tools to manipulate these things | 12:33 | |
12:44
colomon joined
12:53
colomon joined
|
|||
brrt | in the last bit of refactors, before we really have to change the semantics of the program, i'm pondering to have tile emit functions just take a tile* | 12:58 | |
which, after all, already contains the buffers with the value descriptors and arguments | 12:59 | ||
i'm quite amazed how far we can get by incremental changes | |||
14:47
brrt joined
16:27
tomboy64 joined
16:29
tomboy64 joined
16:30
brrt joined
16:32
tomboy64 joined
17:55
zakharyas joined
18:49
TimToady joined
19:08
colomon joined
19:25
TimToady joined
|
|||
lizmat | ++brrt # blogging | 20:21 | |
jnthn | ooh, a brrog? :) | 20:37 | |
jnthn plans to do that tomorrow | |||
Was going to today, then after I hauled myself out of bed was like...nope, resting day :) | 20:38 | ||
brrt++ # nice post! | 20:47 |