| Time |
Nick |
Message |
| 00:06 |
MTDiscord |
<MisterE> We need cube particles |
| 00:08 |
MTDiscord |
<MisterE> We don't need arbitrary-mesh particles, but we do need basic cubes |
| 00:08 |
MTDiscord |
<MisterE> That can rotate, have velocity, acc, etc |
| 01:26 |
|
Baytuch_2 joined #minetest-dev |
| 01:34 |
|
kilbith joined #minetest-dev |
| 01:34 |
|
YuGiOhJCJ joined #minetest-dev |
| 01:37 |
|
vampirefrog joined #minetest-dev |
| 01:48 |
|
YuGiOhJCJ joined #minetest-dev |
| 05:00 |
|
MTDiscord joined #minetest-dev |
| 05:33 |
|
fluxionary joined #minetest-dev |
| 06:10 |
|
fluxionary joined #minetest-dev |
| 06:27 |
|
ivanbu joined #minetest-dev |
| 06:29 |
|
appguru joined #minetest-dev |
| 08:18 |
|
pmp-p joined #minetest-dev |
| 08:35 |
|
calcul0n joined #minetest-dev |
| 09:40 |
|
olliy joined #minetest-dev |
| 10:38 |
|
vampi__ joined #minetest-dev |
| 11:36 |
|
TheCoffeMaker joined #minetest-dev |
| 12:24 |
|
jwmhjwmh joined #minetest-dev |
| 12:38 |
jwmhjwmh |
Any idea why this executable can't find symbols from libMinetest.so at link time? https://github.com/TurkeyMcMac/minetest/blob/0fbc7648d4b943d739a673e5c453ae9a45239ecb/android/native/jni/Android.mk#L300-L304 |
| 12:45 |
|
TheCoffeMaker joined #minetest-dev |
| 13:20 |
|
Fixer joined #minetest-dev |
| 13:38 |
jwmhjwmh |
Nevermind, I'm trying a different approach. |
| 13:52 |
sfan5 |
you can't run an executable on android |
| 14:06 |
jwmhjwmh |
Damn. I suppose I have to use some Android-specific API to launch a process. Why can't Android just be like normal Linux, for God's sake? |
| 14:30 |
sfan5 |
the api isn't the issue, it's POSIX |
| 14:31 |
sfan5 |
the issue is that afaik you can only ship shared libraries with apps, and the (unpacked) apk location is the only place you are allowed to load/execute stuff from |
| 14:31 |
sfan5 |
for example extracting an executable to the app data folder and executing it won't work |
| 14:50 |
pgimeno |
I've heard about termux but I don't know if it's too limited for this purpose |
| 14:53 |
sfan5 |
I believe the sandboxing changes totally broke termux too |
| 14:54 |
MTDiscord |
<ROllerozxa> termux bypasses it by targetting a lower API level, but that also means they can't update the google play version anymore |
| 15:01 |
|
qur joined #minetest-dev |
| 15:15 |
|
olliy joined #minetest-dev |
| 15:22 |
jwmhjwmh |
Perhaps libMinetest.so could be compiled as an executable with dynamic symbols. |
| 15:27 |
rubenwardy |
I suspect you may need to create the process in Android land and run C++ using the NDK |
| 16:10 |
jwmhjwmh |
If someone with Android experience wants to try to get it working, the process is launched in src/script/csm/csm_controller.cpp |
| 16:34 |
|
Desour joined #minetest-dev |
| 16:55 |
|
LandarVargan joined #minetest-dev |
| 18:35 |
|
jwmhjwmh joined #minetest-dev |
| 18:47 |
|
proller joined #minetest-dev |
| 20:19 |
|
qur joined #minetest-dev |
| 20:23 |
|
Fixer joined #minetest-dev |
| 21:21 |
MTDiscord |
<GoodClover> What purpose do peer ids server in the network protocol? MT ignores packets from different addr+port combos, so it's not for if that changes. Is there some weird situation where multiple clients could be behind the same addr+port? i.e. why not just use the ip+port as an ID? |
| 21:21 |
MTDiscord |
<GoodClover> Unless it's a historical thing/in case we ever need it thing? |
| 21:24 |
|
proller joined #minetest-dev |
| 21:29 |
|
MTDiscord1 joined #minetest-dev |
| 21:30 |
MTDiscord1 |
<GoodClover> hmm I suppose using addr+port as an index would be slow compared sequential IDs the server controls which can be an easy array |
| 21:36 |
MTDiscord |
<GoodClover> I take that back, peer ids use a std::map |
| 21:41 |
|
Baytuch joined #minetest-dev |
| 21:42 |
MTDiscord |
<GoodClover> Hmm, in very old code peer ids were used between clients to talk about other players. IIUC other players are all objects now as far as the client is concerned? |
