slipher commented Jul 27, 2020

Travis failure is real.

slipher commented Jul 27, 2020

Thanks for the explanations. I disagree that only developers will want to change renderer options - the user might want to tweak options to try to …

DolceTriade commented Jul 27, 2020

ya, we should start moving as much as we can into Lua, however, in this case, i don't think we have the infrastructure there yet, so we can just me…

illwieckz commented Jul 26, 2020

Yep, I experienced similar need and I thought about bot fill, because basically it seems to fit all the needs. Even playing against a team full of …

slipher commented Jul 26, 2020

Perhaps 'bot fill' could be made a votable command, with selectable team and number. This command sets a target number of players for one or both t…

slipher commented Jul 23, 2020

Partially. This shows registering a C++ callback (albeit in a weird way, I think there are utilities for doing that), and making an event handler t…

Dimitriio commented Jul 22, 2020

Hi @slipher, @DolceTriade, I saw in the sample luainvaders of librocket something that look like what you were saying. The example show a Lua Inter…

illwieckz commented Jul 20, 2020

David Airlie suggested to try to run the game with R600_DEBUG=nohyperz environment variable, I tried on both AGP Radeon HD 2600 PRO (RV630) and PCI…

illwieckz commented Jul 19, 2020

Note: Half Lambert lighting is only meant to be used on models (like rim lighting). Currently the code may be used on other surfaces too and that m…

illwieckz commented Jul 19, 2020

@slipher note that in the future we may want to rebuild the shaders and reload assets as a background task when the user toggles options. Maybe tha…

illwieckz commented Jul 19, 2020

unless it is possible to use both the enabled and disabled variants without restarting To sum it up: when all cvars are enabled, the enabled and …

illwieckz commented Jul 19, 2020

We need permutations because a single scene can display for example a texture with lightmaps and normal maps (a bsp surface that has light maps) an…

slipher commented Jul 19, 2020

Well we have to think whether it is a good tradeoff to save a few milliseconds in shader compilation, but require restarting the game to change an …

illwieckz commented Jul 18, 2020

Bug was reported on Mesa side: gitlab.freedesktop.org:mesa/mesa#3290

slipher commented Jul 18, 2020

As @DolceTriade tells me, the right way to do this stuff is to have a Lua script that builds the table, and some presentation-agnostic way of makin…

slipher commented Jul 18, 2020

LGTM

Dimitriio commented Jul 18, 2020

Well, the source of the datagrid is only one array, so if we want one datagrid to be shown (with the sorting), we need to merge them, atleast in th…

illwieckz commented Jul 18, 2020

Hi, that's nice to see! 👍 Do we need to merge them? We may just list LAN servers atop Internet server list, but in this case the issue is: how to n…

Dimitriio commented Jul 18, 2020

Hello @illwieckz, i am working on it. I have done for now the UI. As you can see, i removed the tabs and have added the "Source" column. But i have…

illwieckz commented Jul 18, 2020

I tried with four TeraScale 1 GPU, and two of them reproduce the issue: Sapphire Radeon HD 2600 PRO | RV630 HP Radeon HD 3650 | RV635 PRO (the one…

illwieckz commented Jul 17, 2020

Now lightmaps aren't loaded when light mapping is disabled, same for “light style” light maps.

illwieckz commented Jul 16, 2020

We may also not load lightmaps when lightmapping is disabled but #296 must be fixed first (“light styles” dynamic lightmaps are still displayed eve…

illwieckz commented Jul 14, 2020

we may also enforce the option to also downscale the lightmap if tests show this is required too Tests show this is not required. Hardware that h…

illwieckz commented Jul 14, 2020

I added a commit editing lioghtMapping GLSL shader to not compile the dynamic light code when dynamic lighting is disabled. This dynamic light code…

illwieckz commented Jul 13, 2020

Also, on that computer with so much feature disablement, GLSL shader size goes from 8.5MB to 2MB (size of the GLSL cache).

illwieckz commented Jul 13, 2020

Well, this is just a log, can't break things.

illwieckz commented Jul 13, 2020

More logs about this GPU: SDL_Init( SDL_INIT_VIDEO )... Using SDL Version 2.0.10 SDL using driver "x11" Initializing OpenGL display Display aspect:…

illwieckz commented Jul 13, 2020

Unfortunately, the problem does not occur when compiling shader but when using them, so #354 is helpless to identify the shader that fails. Also, w…

illwieckz commented Jul 13, 2020

All of this seems to be printed by the driver itself: r300 FP: Compiler Error: ../src/gallium/drivers/r300/compiler/r300_fragprog_emit.c::emit_alu(…

illwieckz commented Jul 13, 2020

travis failure is a travis issue

slipher commented Jul 13, 2020

LGTM

illwieckz commented Jul 12, 2020

In fact the Mesa driver type sets a GLSL macro that is not used anymore. And in fact it looks like the code setting this type may never detect Mesa…

illwieckz commented Jul 12, 2020

Also this code does not work anymore: Daemon/src/engine/sys/sdl_glimp.cpp Lines 1356 to 1357 in 6b94e2e AMD renderer string does not con…

illwieckz commented Jul 12, 2020

This confirms GL 2.1 hardware is considered as old one: Daemon/src/engine/sys/sdl_glimp.cpp Lines 1359 to 1360 in 6b94e2e Pre-TeraScale …

illwieckz commented Jul 12, 2020

else if ( Q_stristr( glConfig.renderer_string, "rv770" ) || […] { glConfig.hardwareType = glHardwareType_t::GLHW_ATI_DX10; I have access to an RV…

illwieckz commented Jul 12, 2020

We may just assume all GPUs are modern ones, then put special conditions when people report bugs.

illwieckz commented Jul 11, 2020

This hardware seems to support GL_EXT_framebuffer_object instead of ARB_framebuffer_object. Khronos' OpenGL wiki says: Warning: This section descr…

illwieckz commented Jul 11, 2020

Only two extensions seems to be required: ARB_half_float_vertex and ARB_framebuffer_object, on such hardware or driver, the first one is there but …

illwieckz commented Jul 11, 2020

I see alternative code for missing GL_ARB_texture_float, GL_EXT_gpu_shader4, GL_EXT_texture_integer, GL_ARB_texture_rg, GL_ARB_texture_gather, and G…

illwieckz commented Jul 11, 2020

There is a lot of missing extensions: ...GL_ARB_texture_float not found ...GL_EXT_gpu_shader4 not found ...GL_EXT_texture_integer not found ...GL_A…

illwieckz commented Jul 11, 2020

On a side note, Xonotic/DarkPlaces (an OpenGL 2.1 game if I'm right) manages to render things on this hardware. This is garbage (like gnome shell i…

illwieckz commented Jul 11, 2020

The Warn: SDL_GL_CreateContext failed: Could not create GL context: GLXBadFBConfig message may not be related to the segfault, this is a warning be…

illwieckz commented Jul 8, 2020

As said in #347, we may want to add a cvar to prevent to enforcement to occur, to be able to keep the ability to test if the issue occurs on given …

illwieckz commented Jul 7, 2020

That really looks like a driver bug. When enabling tri display (r_showtris 1), the line repaint the background texture with the foreground texture:

illwieckz commented Jul 7, 2020

We may also want to name the faulty shader and dumps to make debugging easier.

illwieckz commented Jul 7, 2020

Note that the game starts properly and displays the menu properly, the compiler error occurs when loading maps.