|
github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm Set by AlexDaniel on 12 June 2018. |
|||
|
00:06
patrickz left
00:41
MasterDuke left
|
|||
| Geth | MoarVM/experimental-hllize-speshing: c026ae48d6 | (Timo Paulssen)++ | src/spesh/optimize.c WIP further hllize speshializations. can't really test it, tho |
01:08 | |
|
07:39
zakharyas joined
07:42
MasterDuke joined
08:30
patrickb joined
|
|||
| MasterDuke | looks like our libatomic_ops is kind of outdated. upstream is now at 7.6.10 (soon to 7.6.12 github.com/ivmai/libatomic_ops/blo...ChangeLog) and we're only at 7.6.2 | 08:34 | |
|
08:36
Altai-man_ joined
|
|||
| MasterDuke | it's odd, if `MVM_string_utf8_c8_encode_C_string` is in my list of allocating functions to check (even if it's the only one!) then the plugin reliably causes a segfault in gcc when compiling src/io/filewatchers.c | 08:45 | |
|
10:27
sena_kun joined
10:28
Altai-man_ left
10:49
MasterDuke left
11:20
zakharyas left
|
|||
| Kaiepi | what are MVM_BC_get_I64 and MVM_BC_get_N64 for? | 11:37 | |
| timotimo | they grab an int64 or num64 literal out of the bytecode | 11:51 | |
| Kaiepi | when would they be used over .i64/.n64 in interp.c? | 11:52 | |
|
11:53
patrickb left
11:54
MasterDuke joined
|
|||
| timotimo | you mean on GET_REG? | 11:57 | |
| Kaiepi | yeah | 12:03 | |
| timotimo | GET_REG reads a 16 bit integer from the bytecode at an offset and gives you a pointer to the register in the current frame's work area ("stack frame") selected by that integer value | 12:04 | |
| Kaiepi | ah | 12:06 | |
| so that's probably not what i want | |||
| timotimo | depends entirely upon what you want to do | 12:07 | |
| Kaiepi | i'm writing an op for a function that takes a couple MVMuint32, but these are user input | 12:08 | |
| what i have atm is the op itself takes a couple uint64 and casts them to MVMuint32 to pass to the function, and on the rakudo side of things i use nqp::unbox_u and UInt to make sure the input is valid | 12:09 | ||
| but i'm not really sure if this is the right way to go | |||
| timotimo | you will need a register, then | 12:22 | |
| well, multiple | 12:25 | ||
|
12:26
Altai-man_ joined
12:27
zakharyas joined
12:29
sena_kun left
|
|||
| MasterDuke | is there ever a case when you're freeing a struct member (e.g., `MVM_free(foo->bar)`) that you shouldn't (or wouldn't want t) also set the member to `NULL` (i.e., use `MVM_free_null(foo->bar)`)? | 12:31 | |
| jnthn | If it will never be accessed again, setting it to NULL is a waste of time. | 12:32 | |
| MasterDuke | so if before an MVM_exception_throw_* it's unnecessary? | 12:44 | |
|
12:47
squashable6 left
12:49
squashable6 joined
12:54
domidumont joined
12:57
Kaiepi left
13:02
Kaiepi joined,
Kaiepi left
13:03
Kaiepi joined
13:09
domidumont left
|
|||
| Geth | MoarVM: MasterDuke17++ created pull request #1291: Add gcc plugin to check for missing free before throw |
13:19 | |
|
14:27
sena_kun joined
14:29
Altai-man_ left
15:05
Kaeipi joined
15:09
Kaiepi left
15:27
Kaeipi is now known as Kaiepi
15:51
konvertex joined
16:26
Altai-man_ joined
16:28
sena_kun left
16:45
squashable6 left
16:48
squashable6 joined
17:55
squashable6 left
17:56
squashable6 joined
18:07
zakharyas left
18:27
sena_kun joined
18:28
lucasb joined
18:29
Altai-man_ left
19:30
zakharyas joined
|
|||
| nine | MasterDuke: you can't say that in general. Exceptions can be caught and the free'd thing may be in a long lived data structure | 20:21 | |
|
20:26
Altai-man_ joined
20:29
sena_kun left
20:52
zakharyas left
22:27
sena_kun joined
22:29
Altai-man_ left
|
|||
| MasterDuke | what about if you free the struct at the end? do you ever need to NULL the members after you free them? | 22:37 | |
|
23:33
lucasb left
|
|||