| Time |
Nick |
Message |
| 01:27 |
Soni |
dropping windows support is easy thanks to qemu |
| 01:27 |
Soni |
well, "dropping" windows support |
| 01:29 |
Soni |
so how is wasm support coming along? |
| 02:36 |
MTDiscord |
<warr1024> I'm in favor of dropping Windows support in the same sense that one drops the bass. |
| 02:59 |
|
[MTMatrix] joined #minetest-dev |
| 02:59 |
|
[MTMatrix] joined #minetest-dev |
| 03:00 |
|
[MTMatrix] joined #minetest-dev |
| 05:00 |
|
MTDiscord joined #minetest-dev |
| 06:16 |
|
fluxionary joined #minetest-dev |
| 07:35 |
|
json joined #minetest-dev |
| 07:49 |
|
calcul0n__ joined #minetest-dev |
| 08:03 |
nrz |
Haha wasm is interesting, i checked if it can replacé our lua, but no |
| 08:09 |
MTDiscord |
<paradust> Soni: https://minetest.dustlabs.io/ |
| 08:09 |
MTDiscord |
<paradust> I've been trying to get it upgraded to 5.8.0 but hard-stuck on some shader problems |
| 08:51 |
Mantar |
Not bad, runs surprisingly okay in the browser! But Exile's not supposed to use v7 mapgen, it's hard enough without a dozen giant cliffs in your way! |
| 09:04 |
MTDiscord |
<paradust> Mantar: I need to find a way to fix that. The auto-launch bypasses the main menu, but it's the main menu which provides the list of allowed mapgens |
| 09:07 |
celeron55 |
just set the mg_name setting |
| 09:07 |
celeron55 |
(it's dumber than you might expect) |
| 09:08 |
celeron55 |
(at least i think that's how it works) |
| 09:11 |
MTDiscord |
<paradust> yea I guess I just need to read game.conf early and look for allowed_mapgens and disallowed_mapgens and pick accordingly |
| 09:18 |
MTDiscord |
<paradust> nrz: In theory, you could make a Lua JIT which emits webassembly. It would be trickier and less performant than a native JIT, but provides an extra layer of protection. |
| 09:20 |
MTDiscord |
<warr1024> It could turn out more efficient in the long run. Specialized JITs for Lua don't get as much community development as web tech stuff. |
| 09:27 |
|
TLuna joined #minetest-dev |
| 09:38 |
|
MTDiscord joined #minetest-dev |
| 10:33 |
|
vampirefrog joined #minetest-dev |
| 10:37 |
sfan5 |
@paradust I think the mingw people are aware of a minimal example but they might not know that segment heap makes it ~50% worse |
| 10:37 |
sfan5 |
in either case it seems they don't have a real fix beyond some bandaid either |
| 10:53 |
sfan5 |
hmm some of the rust people said that mingw + clang does not have this issue |
| 10:55 |
|
TLuna joined #minetest-dev |
| 11:03 |
|
appguru joined #minetest-dev |
| 11:05 |
sfan5 |
sounds worth a try |
| 11:05 |
sfan5 |
note to self: https://github.com/mstorsjo/llvm-mingw/blob/master/wrappers/clang-target-wrapper.sh |
| 11:41 |
|
YuGiOhJCJ joined #minetest-dev |
| 12:27 |
|
TLuna joined #minetest-dev |
| 13:13 |
|
TLuna joined #minetest-dev |
| 13:19 |
sfan5 |
I added an unittest for the broken TLS and it works (= fails) even in wine \o/ |
| 15:05 |
|
Desour joined #minetest-dev |
| 15:28 |
|
TLuna joined #minetest-dev |
| 16:43 |
|
calcul0n joined #minetest-dev |
| 16:50 |
|
fluxionary joined #minetest-dev |
| 17:06 |
sfan5 |
after fighting the toolchain for a while I've verified that mingw+clang has working thread-local storage |
| 17:19 |
sfan5 |
so obvious question: do we want to switch to that? |
| 17:45 |
|
TLuna joined #minetest-dev |
| 18:06 |
|
Desour joined #minetest-dev |
| 18:26 |
MTDiscord |
<jordan4ibanez> clang is way better |
| 18:26 |
ROllerozxa |
clang sounds cool |
| 19:13 |
|
vampirefrog joined #minetest-dev |
| 20:33 |
|
calcul0n joined #minetest-dev |
| 21:11 |
|
appguru joined #minetest-dev |
| 22:21 |
|
appguru joined #minetest-dev |
| 23:09 |
|
Desour joined #minetest-dev |
| 23:32 |
|
panwolfram joined #minetest-dev |
| 23:32 |
MTDiscord |
<paradust> what platforms (other than web) use the OGLES2 driver? |
| 23:36 |
Desour |
phones |
| 23:50 |
|
imi joined #minetest-dev |
