| Time |
Nick |
Message |
| 00:46 |
|
Sompi joined #luanti-dev |
| 04:39 |
|
sugarbeet joined #luanti-dev |
| 04:58 |
|
SFENCE joined #luanti-dev |
| 05:00 |
|
MTDiscord joined #luanti-dev |
| 05:10 |
|
SFENCE joined #luanti-dev |
| 05:30 |
|
SFENCE joined #luanti-dev |
| 07:13 |
|
mrcheese_ joined #luanti-dev |
| 07:19 |
|
mrcheese joined #luanti-dev |
| 07:19 |
|
mrcheese joined #luanti-dev |
| 07:32 |
|
YuGiOhJCJ joined #luanti-dev |
| 07:33 |
|
mrcheese joined #luanti-dev |
| 08:00 |
|
mrcheese_ joined #luanti-dev |
| 10:21 |
|
sugarbeet joined #luanti-dev |
| 11:35 |
sfan5 |
merging #16686, #16704, #16705 in 20m |
| 11:35 |
ShadowBot |
https://github.com/luanti-org/luanti/issues/16686 -- [no squash] Call fsync in safeWriteToFile() by sfan5 |
| 11:35 |
ShadowBot |
https://github.com/luanti-org/luanti/issues/16704 -- Fix minimap rendering regression by appgurueu |
| 11:35 |
ShadowBot |
https://github.com/luanti-org/luanti/issues/16705 -- Bump all used GitHub Actions by AFCMS |
| 11:53 |
|
Eragon joined #luanti-dev |
| 12:13 |
[MatrxMT] |
<Zughy> sfan5 |
| 12:13 |
sfan5 |
my clock runs slower |
| 13:01 |
|
ivanbu joined #luanti-dev |
| 14:27 |
|
mrcheese joined #luanti-dev |
| 15:39 |
|
SFENCE joined #luanti-dev |
| 16:04 |
|
SFENCE joined #luanti-dev |
| 16:06 |
|
SFENCE joined #luanti-dev |
| 16:42 |
|
turtleman joined #luanti-dev |
| 16:52 |
|
SFENCE joined #luanti-dev |
| 17:54 |
SFENCE |
Luanti migration to iOS... iOS have endian.h, but not be16toh, htobe16 and similar functions used in serialize.h for readU16, writeU16 and similar functions. |
| 17:55 |
SFENCE |
If I add conditional preprocesor code for iOS, Sounds it as good solution for it? |
| 18:05 |
|
Wuzzy joined #luanti-dev |
| 18:10 |
MTDiscord |
<luatic> serialization code for integers should not need those at at all, you just use bitops to get the lowest/highest byte... |
| 18:12 |
MTDiscord |
<luatic> wait actually we have those fallbacks already in place, now i see what you mean |
| 18:13 |
MTDiscord |
<luatic> i guess abridging the condition from HAVE_ENDIAN_H to something else makes sense, yeah |
| 18:15 |
MTDiscord |
<luatic> what's your concern with doing that? |
| 18:48 |
|
SFENCE_arch joined #luanti-dev |
| 19:08 |
|
SFENCE joined #luanti-dev |
| 19:11 |
SFENCE |
My concern is to do something like this: https://pastebin.com/VqCrMPeP |
| 19:11 |
SFENCE |
for iOS at least. |
| 19:36 |
|
mrcheese joined #luanti-dev |
| 21:14 |
|
SFENCE joined #luanti-dev |
| 21:45 |
|
turtleman joined #luanti-dev |
| 21:49 |
|
SFENCE joined #luanti-dev |
| 22:22 |
|
SFENCE joined #luanti-dev |
| 22:55 |
|
SFENCE joined #luanti-dev |
| 23:32 |
MTDiscord |
<luatic> well that would indeed be suboptimal i think, but you don't have to do that? |
| 23:32 |
|
panwolfram joined #luanti-dev |
| 23:32 |
MTDiscord |
<luatic> as said, if you look at serialize.h, there's an #if HAVE_ENDIAN_H ... #else ... |
| 23:33 |
MTDiscord |
<luatic> so if you just change the #if condition to take into account that despite having endian.h, apple doesn't define these functions under these names, that'll work |
| 23:33 |
MTDiscord |
<luatic> that's the lazy solution |
| 23:34 |
MTDiscord |
<luatic> but really i think it probably is a decent solution |
| 23:36 |
MTDiscord |
<luatic> clang and gcc at least seem to recognize the pattern used for the fallback, optimizing it to a bswap, so there's no performance hit |
| 23:37 |
MTDiscord |
<luatic> so there's no need to complicate the code here. if anything, it should probably be simplified by only using the fallback code |
| 23:38 |
MTDiscord |
<luatic> if for some reason you do really want to use the stdlib builtins for iOS, you should abstract over these byteswap util funcs in porting.h |
| 23:50 |
|
mrcheese_ joined #luanti-dev |
| 23:51 |
|
mrcheese_ joined #luanti-dev |
| 23:52 |
|
SFENCE joined #luanti-dev |
