|
07:16
ilbot3 joined
09:41
cognominal joined
11:42
kjs_ joined
16:34
Ven joined
17:04
kjs_ joined
17:17
FROGGS joined
17:32
Ven joined
|
|||
| dalek | arVM: 151b7f7 | jnthn++ | src/spesh/graph. (2 files): A better fix for leak on inlining handlers. This one doesn't wrongly free handlers that came from another spesh candidate, which was the cause of problems with the previous attempt to fix this leak. |
17:32 | |
| jnthn | I valgrund the CORE.setting build this time :) | 17:33 | |
| Which of course took forever, especially as I did leak tracking. | 17:35 | ||
| dalek | arVM: b05f0e6 | jnthn++ | src/mast/driver.c: Don't leak buffer when saving bytecode to file. We'd exit very soon after this anyway, but good to keep a clean sheet. |
17:43 | |
| arVM: 90648b3 | jnthn++ | src/6model/serialization.c: Some missing cleanup in bounded serialization. |
17:50 | ||
| jnthn | Those were the biggest leaks in CORE.setting compilation | ||
| dalek | arVM: 81d7f3b | jnthn++ | src/6model/parametric.c: Fix leak in type parameterization. |
17:56 | |
| lizmat | jnthn: so it's time for a MoarVM / nqp bump ? | 18:06 | |
| jnthn | Might be wise to do a build/spectest with the changes first | 18:08 | |
| But I think it should be pretty alright | |||
| lizmat | ok, I'll give it a try | 18:09 | |
|
21:43
kjs_ joined
23:35
kjs_ joined
|
|||