| Time |
Nick |
Message |
| 00:00 |
Zughy[m] |
can confirm, not being able to type è or é is so annoying |
| 00:02 |
sfan5 |
luk3yx: your steps are a bit fuzzy but I cannot find a bug with the craftpreview boat model in either first or third person |
| 00:03 |
sfan5 |
anyway just open issues, we'll eventually find out the reason |
| 00:03 |
luk3yx |
sfan5: Both lines look 1px too short |
| 00:04 |
sfan5 |
they are |
| 00:04 |
luk3yx |
And the boat needs to be intersecting some node |
| 00:05 |
luk3yx |
https://i.imgur.com/TQGrytP.png |
| 00:05 |
sfan5 |
how could a gui item possibly intersect an in-world node |
| 00:05 |
luk3yx |
I have no idea but they seem to |
| 00:06 |
sfan5 |
the change missing in the model code is probably this then https://github.com/minetest/minetest/commit/3579dd21867598ff30867ebd68b690c85ba14f9b#diff-841c1b824e0ac7bf7749b4e4d239685d529ef79054b6a91dd1730c4b38c625e4L42-L47 |
| 00:06 |
kilbith |
it's not a "GUI item", it's a GUI Scene, it's entirely possible that scenes can conflict |
| 00:08 |
DS-minetest |
such intersection happens when the depth buffer isn't reset, no? |
| 00:08 |
sfan5 |
(that link is supposed to show the changes to src/client/sky.cpp in case that isn't working) |
| 00:08 |
|
Fractalis joined #minetest-dev |
| 00:10 |
kilbith |
luk3yx, try that: https://paste.ubuntu.com/p/dCY3gVYPPm/ |
| 00:10 |
kilbith |
on my end this seems to work correctly with that patch |
| 00:12 |
kilbith |
well even without it, there is no glitch on my end |
| 00:13 |
kilbith |
https://i.imgur.com/zEiWxSU.png |
| 00:13 |
kilbith |
everytime you have a different bad surprise with the model |
| 00:14 |
kilbith |
looks like some vertices (?) aren't drawn in the correct order |
| 00:16 |
sfan5 |
that looks like what'd happen without the depth buffer and no backface culling |
| 00:16 |
sfan5 |
but I suppose both of those are enabled |
| 00:19 |
kilbith |
ok we apparently need to enable both EMF_BACK_FACE_CULLING and EMF_ZWRITE_ENABLE on this scene |
| 00:19 |
kilbith |
this *seems* to fix it |
| 00:20 |
sfan5 |
that breaks transparent models, doesn't it? |
| 00:20 |
kilbith |
didn't test |
| 00:22 |
kilbith |
well yes it does break the transparent models |
| 00:29 |
sfan5 |
since the gui may render it's own independent 3d scene maybe you do have to call driver->clearZBuffer at some point |
| 00:30 |
sfan5 |
the hud also does this before rendering item stacks |
| 00:35 |
luk3yx |
Okay now I'm even more confused |
| 00:35 |
luk3yx |
If I have any items in my inventory the model looks fine |
| 00:38 |
luk3yx |
kilbith: The issue I was having still occurs with your fix |
| 00:39 |
luk3yx |
If I only have one inventory item and drop it in the trash slot the issue reappears the moment I do |
| 00:40 |
kilbith |
here is the reason why |
| 00:40 |
kilbith |
inventory item is a scene from drawItemStack |
| 00:40 |
kilbith |
... which calls up driver->clearZBuffer() |
| 00:41 |
luk3yx |
Interesting, I can reproduce this issue on 5.4.0 |
| 02:10 |
|
hlqkj1_ joined #minetest-dev |
| 02:32 |
|
BuckarooBanzai4 joined #minetest-dev |
| 05:00 |
|
MTDiscord joined #minetest-dev |
| 05:24 |
|
olliy joined #minetest-dev |
| 06:06 |
|
olliy_ joined #minetest-dev |
| 08:00 |
|
ShadowNinja joined #minetest-dev |
| 08:53 |
|
calcul0n_ joined #minetest-dev |
| 09:03 |
|
YuGiOhJCJ joined #minetest-dev |
| 10:03 |
|
hlqkj joined #minetest-dev |
| 11:00 |
|
YuGiOhJCJ joined #minetest-dev |
| 11:44 |
|
tech_exorcist joined #minetest-dev |
| 11:47 |
|
proller joined #minetest-dev |
| 11:58 |
|
tech_exorcist joined #minetest-dev |
| 11:58 |
|
z812 joined #minetest-dev |
| 12:07 |
|
calcul0n__ joined #minetest-dev |
| 12:46 |
|
olliy joined #minetest-dev |
| 13:05 |
|
celeron55_ joined #minetest-dev |
| 13:06 |
|
TC01 joined #minetest-dev |
| 13:06 |
|
basxto2 joined #minetest-dev |
| 13:11 |
|
hlqkj1 joined #minetest-dev |
| 13:25 |
|
celeron55 joined #minetest-dev |
| 14:10 |
|
hlqkj1_ joined #minetest-dev |
| 14:28 |
|
Wuzzy joined #minetest-dev |
| 15:01 |
|
z812_ joined #minetest-dev |
| 15:35 |
|
ircSparky joined #minetest-dev |
| 15:46 |
|
numzero joined #minetest-dev |
| 15:59 |
|
Fixer joined #minetest-dev |
| 16:11 |
|
Fractalis joined #minetest-dev |
| 16:39 |
MTDiscord |
<appguru> backface culling is not the issue here |
| 16:39 |
MTDiscord |
<appguru> it's all about the z buffer |
| 17:06 |
|
DS-minetest joined #minetest-dev |
| 17:08 |
|
ircSparky_ joined #minetest-dev |
| 17:22 |
|
MTDiscord joined #minetest-dev |
| 17:54 |
|
eugenefil joined #minetest-dev |
| 18:04 |
|
olliy joined #minetest-dev |
| 18:37 |
|
proller joined #minetest-dev |
| 18:54 |
|
eugenefil joined #minetest-dev |
| 19:05 |
|
proller joined #minetest-dev |
| 21:04 |
|
proller joined #minetest-dev |
| 21:17 |
|
Fractalis joined #minetest-dev |
| 22:49 |
sfan5 |
numzero: I suppose you don't have a better idea for a fix here? http://sprunge.us/3x3vEs?diff |
| 23:03 |
|
ivanb joined #minetest-dev |
| 23:08 |
numzero |
sfan5: for the exact behavior described, no |
| 23:08 |
numzero |
btw why are you patching 2D line drawing? does anyone ever use it? |
| 23:09 |
sfan5 |
we draw the crosshair using it |
