|
00:26
vendethiel joined
01:04
vendethiel joined
01:30
vendethiel joined
01:48
ilbot3 joined
03:06
colomon joined
07:02
zakharyas joined
07:04
Ven joined
07:10
FROGGS joined
07:14
brrt joined
08:42
FROGGS joined
09:48
camelia joined
10:24
camelia joined
10:39
Ven joined
|
|||
| dalek | arVM: 42e3d2e | timotimo++ | / (8 files): add op "getcodelocation"; gets file and line of a MVMCode |
10:58 | |
| JimmyZ | timotimo: I think ^^ is a bit wrong | 11:47 | |
| timotimo | OK, what do i have to change? | ||
| JimmyZ | you can't put MVM_repr_box_str to the args directly | 11:48 | |
| that's is gc move problem, said two year ago. :P | 11:49 | ||
| timotimo | well, this is the convenience wrapper from reprconv.c | ||
| oh? | |||
| JimmyZ | you put filename_key in the arg stack, and then box allloces, and filename_key moved by gc . | 11:50 | |
| you will got null | |||
| timotimo | oh | ||
| yes | |||
| i think maybe i'll use temp root push calls instead of MVMROOT | 11:52 | ||
| JimmyZ | yes, I like temp root push too :). | 11:53 | |
| timotimo | JimmyZ: gist.github.com/timo/c4f0a66295777...qqqqqqqqq9 looks better now? | 12:02 | |
| gist.github.com/timo/c4f0a662957771347a31 (minus the cat) | |||
| jnthn | temp root push is just hard to see the pairing. | 12:05 | |
| So I discourage it generally. | |||
| JimmyZ | yeah, but could avoid two temp push by re-order the code, allocaces it and then use the var | ||
| I mean bind it earlier | 12:06 | ||
| jnthn | Givne it's hugely unlikely this is hot path code, I'd write it to be obviously correct :P | ||
| timotimo | so ... is this code correct-ish now? | 12:09 | |
| JimmyZ | MMVM_gc_root_temp_push(tc, (MVMCollectable **)&linenumber_boxed); is not needed | 12:10 | |
| timotimo | ah, you are right | 12:11 | |
| that's the only problem? | |||
| JimmyZ | so yeah, it is correct , methinks :P | 12:12 | |
| timotimo | thanks :) | ||
| dalek | arVM: 8f46067 | timotimo++ | src/6model/reprs/MVMCode.c: needed more roots, use temp push instead of MVMROOT |
||
| timotimo | might bump later, too | 12:18 | |
| dalek | arVM: 1d592be | (Jimmy Zhuo)++ | src/6model/reprs/MVMCode.c: fix build, fix some MVMROOT issues |
12:25 | |
| timotimo | argh | 12:31 | |
| thank you | |||
|
12:34
travis-ci joined
|
|||
| travis-ci | MoarVM build passed. Jimmy Zhuo 'fix build, fix some MVMROOT issues' | 12:34 | |
| travis-ci.org/MoarVM/MoarVM/builds/57792400 github.com/MoarVM/MoarVM/compare/8...592be53a7f | |||
|
12:34
travis-ci left
|
|||
| JimmyZ | MVMObject is always a trap, big C trap :P | 12:39 | |
| timotimo | that's right | 12:40 | |
| and also putting declarations in the middle of the curlied spaces :( | 12:43 | ||
| that's my kryptonite | |||
| JimmyZ: thanks for catching my mistakes; i was in between places when i pushed that; didn't try to build even though i should have | 13:18 | ||
|
17:52
FROGGS joined
18:27
oetiker joined
18:33
vendethiel joined
18:42
zakharyas joined
|
|||
| dalek | arVM: a599f01 | (Francois Perrad)++ | src/strings/normalize.h: move up MVM_unicode_normalizer_process_codepoint_full put the declaration before its first use in MVM_unicode_normalizer_process_codepoint fix the following warning: src/strings/normalize.h:86:5: warning: implicit declaration of function ‘MVM_unicode_normalizer_process_codepoint_full’ [-Wimplicit-function-declaration] return MVM_unicode_normalizer_process_codepoint_full(tc, n, in, out); ^ |
19:21 | |
| MoarVM: 597701b | FROGGS++ | src/strings/normalize.h: | |||
| MoarVM: Merge pull request #201 from fperrad/normalize_h | |||
|
21:09
dalek joined
21:15
colomon joined
22:00
lizmat joined
22:21
vendethiel joined
|
|||
| timotimo | i can't really get the number of Int, Scalar and BOOTCode allocations down | 23:08 | |
| it's like a darn waterbed | 23:09 | ||