FOSS gamedev and creative worlds

Hello, I'm using minetest under Linux Mint 11 amd64. The controller of my graphics card (or what lspci says):
VGA compatible controller: Intel Corporation 82945G/GZ Integrated Graphics Controller (rev 02)

Well, the problem is that minetest's screen randomly (mostly at night or in dark places) turns black like this (without the mouse):


I can still open the menu and disconnect or close it, but that user can never play again.


I have tested it with minetest-c55 (development and stable, compiled form source and form ppa), and minetest delta, it happens with all of them. I don't think it's minetest's fault, but it is the only program which I have this problem with, so it has to be related with it.

Am I the only one with this problem?


EDIT: Sometimes it happens at midday, so it is not related with "light"

EDIT: I get buried

https://bbs.archlinux.org/viewtopic.php?id=116471
That's a common problem with the Intel drivers for Linux. a quick search via duckduckgo/google "intel linux to dark" shows solutions.

Nonononono You are just spawned in a underground block and you have to enable free move to fly out of it.. Well at least that was the problem in my case

How do I enable free move?

Press 'K' then 'Space' to come back at the surface.
Also, you can change these keys in the main menu.

Thanks!!!
That was my problem =D. I don't know why, but I was like 20 metres underground O.o

Yeah it happens sometimes. Weird and old bug...

Well, I can play now, but it's still annoying having to be coming back to the surface every 3 minutes. If If it is an old bug, and I have it updated with the ppa, why does it happen? Or it has never been fixed?


EDIT: With the last stable it happens less :D

This happens with me too. I am usually 3 blocks under where I saved. Very annoying. I've looked at the log and it says that the correct location is sent and received, but the Y position is usually 5.0 below what it should be. The only thing I think i could be is difference with OpenGL on Linux and Directx on Windows.

Later commits in git have something to do with this bug. So you may try to build the latest source and test if the bug is fixed.

The fix is server-side, so if you play on the public server new client version won't help.

bookwar wrote:Later commits in git have something to do with this bug. So you may try to build the latest source and test if the bug is fixed.

The fix is server-side, so if you play on the public server new client version won't help.

Anyway, it sometimes happen

bookwar wrote:Later commits in git have something to do with this bug. So you may try to build the latest source and test if the bug is fixed.

The fix is server-side, so if you play on the public server new client version won't help.

I've tested this in latest source pulled from git and it still happens.
I've made a small change to server.cpp at line 2035 to offset the problem:
player->setPosition(player->getPosition()+v3f(0,5,0));

I wonder if the saved location is from where the feet are and when it is loaded it is set to where the "camera" set.

FYI, if the problem primarily happens when you login to a server (or to a local game), it seems to be a bug with the way gravity and collision detection works. When the client environment doesn't update for about half a second (that's often the case when you login), depending on the floating point implementation of your compiler you might glitch through the ground. There is some intermediate NaN result that breaks the physics formulas.

We discussed this bug in the IRC channel not long ago. See this log starting at around 14:00. Here's my patch: yay

kahrl wrote:FYI, if the problem primarily happens when you login to a server (or to a local game), it seems to be a bug with the way gravity and collision detection works. When the client environment doesn't update for about half a second (that's often the case when you login), depending on the floating point implementation of your compiler you might glitch through the ground. There is some intermediate NaN result that breaks the physics formulas.

We discussed this bug in the IRC channel not long ago. See this log starting at around 14:00. Here's my patch: yay

It also often happens just in random moments of time.

Occasionally taking a screenshot caused this to happen to me.

kahrl wrote:FYI, if the problem primarily happens when you login to a server (or to a local game), it seems to be a bug with the way gravity and collision detection works. When the client environment doesn't update for about half a second (that's often the case when you login), depending on the floating point implementation of your compiler you might glitch through the ground. There is some intermediate NaN result that breaks the physics formulas.

We discussed this bug in the IRC channel not long ago. See this log starting at around 14:00. Here's my patch: yay

This was my thinking too, but I had no idea how to test or debug that this happening. Thank you.

That happens with me too, fedora with AMD Radeon graphics.

It doesn't happen to me now (Mint 12 amd64). I thought it was fixed : /