|
github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm Set by AlexDaniel on 12 June 2018. |
|||
|
01:05
lucasb left,
leont left
03:13
kawaii left
03:18
kawaii joined
04:18
benchable6 left,
nativecallable6 left,
sourceable6 left,
evalable6 left,
coverable6 left,
linkable6 left,
tellable6 left,
shareable6 left,
unicodable6 left,
statisfiable6 left,
releasable6 left,
bloatable6 left,
greppable6 left,
committable6 left,
quotable6 left,
notable6 left,
squashable6 left,
bisectable6 left
04:19
sourceable6 joined
04:20
statisfiable6 joined,
releasable6 joined,
notable6 joined,
linkable6 joined,
unicodable6 joined,
bloatable6 joined,
evalable6 joined,
coverable6 joined,
benchable6 joined
04:21
shareable6 joined,
quotable6 joined,
greppable6 joined,
committable6 joined,
tellable6 joined,
nativecallable6 joined,
squashable6 joined,
bisectable6 joined
05:00
frost-lab joined
05:11
frost-lab left
05:20
frost-lab joined
06:14
Altai-man joined
08:20
Altai-man left
08:29
patrickb joined
10:30
Altai-man joined
|
|||
| Geth | MoarVM: cognominal++ created pull request #1398: fix #1395 : On Apple Silicon with arch arm64, does not attempt to JIT |
11:09 | |
| linkable6 | MOARVM#1395 [open]: github.com/MoarVM/MoarVM/issues/1395 Apple Silicon nqp build fails unless MVM_JIT_DISABLE is set | ||
| nine | Thinking about that throw-away frame with lexicals for BEGIN time code. The compiler basically knows when a lexical can only be found outside the current comp unit. So I guess it should be able to simply generate different code for accessing those. | 12:50 | |
| But what about when the same code will be executed again at runtime? Right now we go some extra steps to not compile dynamically compiled frames again. But what if we did just do that? The by far most expensive part of compilation - parsing - won't be done again. | 12:53 | ||
| Apparently we don't go that extra step in NQP | 13:01 | ||
|
13:07
sena_kun joined
13:09
Altai-man left
|
|||
| MasterDuke | we do in rakudo, but not nqp? | 13:38 | |
| nine | yep | 13:44 | |
| A possible reason might be that in NQP there's just a lot less BEGIN time evaluation, so it doesn't matter as mutch. Most common in Raku code are probably role bodies | |||
|
14:08
frost-lab left
15:46
domidumont joined
17:06
Altai-man joined
17:09
sena_kun left
17:18
leont joined
17:46
domidumont left
19:15
zakharyas joined
19:23
zakharyas1 joined
19:25
zakharyas left
21:07
sena_kun joined
21:09
Altai-man left
22:29
patrickb left
22:42
zakharyas1 left
23:01
sena_kun left
|
|||
| Geth | MoarVM: fbd386a022 | (Stéphane Payrard)++ | Configure.pl fix #1395 : does not try to JIT with Apple Silicon and arch being arm64 |
23:44 | |
| MoarVM: 7bdfb38b2e | (Stéphane Payrard)++ | Configure.pl Replace die by hardfail |
|||
| MoarVM: 51cc76c16a | (Jonathan Worthington)++ (committed using GitHub Web editor) | Configure.pl Merge pull request #1398 from cognominal/lipo fix #1395 : On Apple Silicon with arch arm64, does not attempt to JIT |
|||
| linkable6 | MOARVM#1395 [closed]: github.com/MoarVM/MoarVM/issues/1395 Apple Silicon nqp build fails unless MVM_JIT_DISABLE is set | ||