| Time |
Nick |
Message |
| 00:04 |
|
pgimeno joined #minetest-dev |
| 00:05 |
pgimeno |
@sfan5 regarding #11880 one possibility is to only save mapgen when the user changes it, as opposed to when a mod changes it |
| 00:05 |
ShadowBot |
https://github.com/minetest/minetest/issues/11880 -- Either save mapgen choice per-game or not at all |
| 00:07 |
pgimeno |
but not saving it at all is also an option; there aren't so many people that create enough worlds as to justify saving it |
| 00:09 |
pgimeno |
other ideas: to not save it when not all options are available; to not save it when the previous setting is not an available option |
| 00:09 |
|
Alias2 joined #minetest-dev |
| 00:09 |
pgimeno |
and to save it per availability combo |
| 00:10 |
pgimeno |
i.e. one setting for games with choice between v5, v6; one separate setting for games with choice between v5, v6, v7, and so on |
| 00:13 |
pgimeno |
the latter is probably a bad idea because it violates the POLA; "why is there a separate mapgen setting for this and that?" |
| 00:22 |
pgimeno |
one other possibility: to have a checkbox "Save this setting as default" off by default |
| 00:32 |
MTDiscord |
<luatic> erlehmann: nan signedness doesn't matter in practice; I know of a way to obtain "positive" nan, but the only way to check for the nan sign is to convert it into a string |
| 00:32 |
MTDiscord |
<luatic> still, this is a massive improvement over the previous serializer which would crash with nan/inf values |
| 00:34 |
pgimeno |
not the only way, copysign can also extract it |
| 00:34 |
MTDiscord |
<luatic> copysign? |
| 00:34 |
pgimeno |
https://en.cppreference.com/w/cpp/numeric/math/copysign |
| 00:34 |
MTDiscord |
<luatic> I'm talking about Lua though |
| 00:35 |
pgimeno |
oh oops, sorry |
| 00:36 |
pgimeno |
LuaJIT doesn't seem to support signed NaN |
| 00:55 |
pgimeno |
which is to say (after reading the context): don't worry about supporting NaNs because for 90% of the users there's no distinction, and anything that needs to distinguish them won't run in LuaJIT |
| 00:55 |
pgimeno |
negative NaNs that is |
| 01:06 |
|
Noisytoot joined #minetest-dev |
| 01:15 |
|
Taoki joined #minetest-dev |
| 01:28 |
|
erlehmann joined #minetest-dev |
| 03:28 |
|
queria^clone joined #minetest-dev |
| 03:34 |
|
queria^clone joined #minetest-dev |
| 04:00 |
|
v-rob joined #minetest-dev |
| 05:00 |
|
MTDiscord joined #minetest-dev |
| 06:18 |
|
v-rob joined #minetest-dev |
| 07:08 |
|
v-rob joined #minetest-dev |
| 07:09 |
|
olliy joined #minetest-dev |
| 08:04 |
|
tekakutli joined #minetest-dev |
| 08:51 |
|
YuGiOhJCJ joined #minetest-dev |
| 08:56 |
|
calcul0n joined #minetest-dev |
| 09:43 |
|
calcul0n_ joined #minetest-dev |
| 10:03 |
sfan5 |
pgimeno: yeah I think not saving it is just the best way |
| 11:23 |
|
tekakutli joined #minetest-dev |
| 11:53 |
|
Fixer joined #minetest-dev |
| 12:13 |
erlehmann |
sfan5 it surely is the easiest way and very predictable. but the mental model of those controls (world name, mapgen, seed) probably needs to be adjusted a bit to accomodate for that. for example: in 5.4.1 (and later?), the mapgen choice will reset when you change the mapgen, then change the map seed. |
| 12:13 |
erlehmann |
so the question is: would not saving mean you always switch to the first mapgen the game finds acceptable? or do you keep the current choice if the game finds it acceptable? |
| 12:14 |
erlehmann |
(of course the easy way out is again not to think about it too hard) |
| 12:23 |
|
pgimeno left #minetest-dev |
| 12:40 |
erlehmann |
anyone of you aware of this kind of rendering bug with the etnaviv driver? http://dump.mntmn.com/screenshot-2021-12-26-13-31-30.png |
| 12:40 |
erlehmann |
camera is inside head |
| 12:43 |
MTDiscord |
<Sublayer plank> huh, is that in first person or third person mode? |
| 12:47 |
MTDiscord |
<luatic> @ROllerozxa that's first person |
| 12:48 |
MTDiscord |
<luatic> I remember Jordach's first person attachment PR having similar issues |
| 12:48 |
MTDiscord |
<Sublayer plank> reminds me of a similar graphical glitch that would occur in the D3D video drivers |
| 12:48 |
MTDiscord |
<luatic> Here it is: https://github.com/minetest/minetest/pull/10360#pullrequestreview-484365382 |
| 12:49 |
MTDiscord |
<luatic> This was the commit that fixed it: https://github.com/minetest/minetest/pull/10360/commits/7b8975120a3f45fb17f24200225abb58a0470733 |
| 12:49 |
MTDiscord |
<luatic> So obviously this is a backface culling issue |
| 12:56 |
erlehmann |
i forwarded it to mntmn maybe that will lead to a solution |
| 14:00 |
|
Extex joined #minetest-dev |
| 15:03 |
|
Taoki joined #minetest-dev |
| 16:24 |
|
proller joined #minetest-dev |
| 18:02 |
|
troller joined #minetest-dev |
| 18:22 |
|
troller joined #minetest-dev |
| 18:53 |
|
troller joined #minetest-dev |
| 18:59 |
|
Noisytoot joined #minetest-dev |
| 21:44 |
|
GreenXenith joined #minetest-dev |
| 22:12 |
|
Fixer joined #minetest-dev |
| 23:19 |
|
appguru joined #minetest-dev |
| 23:54 |
|
MTDiscord joined #minetest-dev |
