FARPROC is defined with an empty parameter list, which newer versions of GCC consider equivalent to (void). Compilation fails when qSHGetFolderPath gets called with arguments.
The VAO surface cache uses glBufferSubData and triggers a very slow
path in some GLES implementations. Specifically I have observed 10x
frame times under Emscripten with ANGLE/Metal on macOS and with Mali
on Android.
Co-authored-by: Zack Middleton <zack@cloemail.com>
Co-authored-by: WofWca <wofwca@protonmail.com>
Use XDG_DATA_HOME for Flatpak so that fs_homepath is accessible without
special permissions/handling. This changes the home path for Flatpak
from ~/.q3a/ to ~/.var/app/org.ioquake3.ioquake3/data/q3a/.
There is a copy to clipboard option but the error itself wasn't printed
to the console so it wasn't copied to the clipboard.
Unix-like platforms print it to the console so it's written in
crashlog.txt.
The error was only displayed for Team Arena if one of the Team Arena pk3s
were found. Fix it to be display when trying to run Team Arena even with
no Team Arena pk3s present.
Don't project sun shadows (r_forceSun 1) on to view depth equal to 1.0
(nothing drawn or skybox). This caused a full second shadow of the
entire level in tr.screenShadowImage. It would move as the camera far
plane changed and rotate/stretch strangely as the camera view changed.
It was visible in-game on lightmapped transparent surfaces facing the
skybox and happen to overlap the extra shadow of the level.
This affected the glass in wop_padship's underwater room.
Default com_maxfps to 0 under Emscripten. Under Emscripten the browser
handles throttling the frame rate. Manual framerate throttling interacts
poorly with Emscripten's browser-driven event loop.
If getting the display index of the existing window fails, fallback to
the default display size instead of falling back to 640x480.
emscripten r_mode -2 worked on window creation but vid_restart fellback
to 640x480.
Also make viewing client.html directly use client-config.json in the
same directory so it's possible to "live edit" both the files without
having to run make.
List files for multiple games in a single client-config.json file so
that com_basegame argument can pick different game data.
Use ioquake3.html?com_basegame=demoq3 (or tademo) to run the Quake 3 or
Team Arena demo. They require new QVMs from baseq3/missionpack to run.
Modify the client HTML file when copying it to the build directory to
apply current CLIENTBIN and BASEGAME. It always loads engine/data from
the current directory (no need to try to locate the build directory).
This enables several things:
* Optionally load pk3 files from a web server at runtime instead of bundling them with Emscripten at build time
* Set command line arguments via URL param
* It's not ugly
This mainly targets OpenGL ES 2.0 but it also supports compiling GLSL as
ESSL 3.00. It's missing support for framebuffer objects which should be
possible on ES 2. (Though using renderbuffers instead of textures.)
opengl1 cvars that are not supported will display a message and disable
the cvar. This has not been reviewed for new opengl2 cvars. Enabling
cvars may cause rendering issues. Some of the broken cvars may be
possible to support using OpenGL ES 3 features.
The game displays okay with the default cvars.
Fix six image skybox having a black border around the sides of the sky
when using OpenGL 1.1 (using CL_CLAMP instead of GL_CLAMP_TO_EDGE).
It's technically visible in q3dm10 but it's more obvious in Team Arena
maps such as mpteam6.
Don't draw the world scene to a separate FBO from the rest of the
screen.
This fixes the world scene having HOM instead of seeing through to the
previously drawn content. World of Padman uses this to have a separate
3D scene for the sky and world in wop_padship for dynamic skybox.
This also makes r_ext_framebuffer_multisample apply to HUD models
instead of depending on r_ext_multisample (which doesn't work on Linux
with some Intel graphics).
Newer mingw-w64 changed default Windows version compatibility.
Need to specify older behavior for EnumProcesses() to avoid "missing
K32EnumProcesses()" runtime error.
When r_ext_framebuffer_multisample > 0 was used, flares behind the
mirror in q3tourney6 were incorrectly visible. This was because it
checks scene depth in the depth prepass which only drew opaque
surfaces. It also needs to contain depth for mirror/portal surfaces.
tcGen environment generates texcoords in range of 0.0 to 1.0 and they
need to be offset to the position/size in the merged lightmap atlas.
This also needs to be after tcMods so they apply for the original range.
This fixes tcMod scale used by main_q3abanner and shinygrate1_4.
This issue was visible on the blue monitor (comp3c) near the plasmagun
in q3dm0 and the Quake III Arena banner in q3dm17.
Affected shaders in Quake 3 and Team Arena maps:
textures/base_wall/comp3
textures/base_wall/comp3b
textures/base_wall/comp3b_dark
textures/base_wall/comp3c
textures/base_wall/main_q3abanner
textures/base_wall/shinygrate1_4
textures/sfx/teslacoil
All of the shaders are used by q3dm0 but other maps also use some.
This fixes the texcoord range to be 0.0 to 1.0 for external lightmaps in
shaders that also have an internal lightmap. (Instead of 2.0 to 3.0 for
example.) This only affects clampMap as repeating wrapping sampled the
correct location. I haven't found such a shader but I need to get real
lightmap index for future merged lightmap hacks.
Using r_drawSunRays 1, r_hdr 0, cg_viewsize 50 caused the sun rays to
only draw properly in the bottom right quarter of the world viewport.
The scissor rectangle for the world viewport was applied to the first
FBO_FastBlit() in RB_SunRays().
Set glScissor() in FBO_FastBlit() as it's done in FBO_Blit() and
FBO_BlitFromTexture(). Sun rays probably worked correctly with r_hdr 1
because the blit for HDR changed the scissor state.
Shader stage tcMods for matrix and turb effects need to be applied in
order for turb to be correct and all tcMod turb need to be applied
instead of only the last one.
Quake 3's textures/liquids/slime1 had tcMod turb and then tcMod scale.
OpenGL2 applied the matrix first and then turb which had the wrong result.
On non-Windows, compiling QVM tools failed if dagcheck.md had CRLF line
endings and compiling QVMs failed if game source had CRLF line endings.
Also made Windows open the files as binary (no automatic CRLF to LF) so
it behaves the same as on non-Windows.
If mapLightScale and shadowScale weren't included at the end of
q3gl2_sun, the next shader line was skipped by SkipRestOfLine().
COM_ParseExt() with allowLineBreaks=qfalse, returns "" once and
then goes to the next line anyway. (Doesn't work well multiple
"optional" tokens. It looks like a vanilla bug.)
Fixes World of Padman wop_trashmap.
For r_mergeLightmaps 1
- Fix tcMod transform on "map $lightmap" stages
- Fix external lightmap image texcoords if shader also has an internal
lightmap
For r_sunlightMode 1
- Fix "tcGen lightmap" stages with blendFunc GL_SRC_ALPHA GL_ONE being
converted to white image + modulate lightmap and drawing incorrectly
Generating cubemaps set backEnd.viewParms.isMirror = qtrue while the
loading screen said 'loading... maps/q3dm1.bsp' and it just stayed like
that until done loading (no additional messages or item icons) because
all 2D drawing was culled due to flipped culling for isMirror.
This was noticed because a recent commit fixed RB_ShowImages() to be
drawn to the screen instead of draw into renderFbo and never blit to
the screen. Now the loading screen draws over it as expected.
Mentioned commit: a81df34905
"OpenGL2: Fix border for cg_viewsize using HDR/FB-MSAA w/postProcess"
Fix r_ext_framebuffer_multisample > 0 causing the screen to always be
solid black when using AMD Windows driver.
The AMD Windows driver requires binding renderbuffer for it to be valid.
The OpenGL2 renderer uses GL_EXT_direct_state_access that shouldn't
require this. It would be required for Core/GL_ARB_direct_state_access.
It seems like a driver bug.
If using renderFBO (r_hdr 1 or r_ext_framebuffer_multisample > 0) and
r_postProcess 1, the viewport border (cg_viewsize < 100) which is drawn
before the world scene was drawn to the renderFBO but not copied to the
default framebuffer.
Now 2D before the world scene is drawn to the default framebuffer
directly for r_postProcess 1 like 2D drawn after the world scene.
The mod list on Windows would search the root of the drive if
fs_basepath, fs_homepath, fs_steampath, or fs_gogpath are blank ("")
(which is usually the case).
The issue is in the low-level Sys_ListFiles() but it only affects the
mod menu, on Windows. It cannot be abused by console commands to
list system files outside of the virtual filesystem.
---
If a directory at the root of the drive of the working directory
contained a pk3 file, the directory was listed in the mods menu. The
virtual filesystem doesn't add blank directory names to the search path
so it cannot load mods from the drive root. (Unless of course you set a
fs_*path cvar to "C:\".)
Sys_ListFiles() with blank directory caused Windows to use "\*" for the
search path and "\" prefix means root of drive. Unix opendir("") failed
so nothing was listed for blank directory.
Sys_ListFilteredFiles() with blank directory _and_ specifying subdirs
could access any directory (on Windows and Unix-like) but no code uses
this or makes it accessible.
These functions are only used for initializing the virtual filesystem
and listing mods. They are not accessible by anything else such as a
console command. Only the mods menu, on Windows, is affected.
The macOS client and server were completely unusable when run from a
terminal. They blocked forever in `[NSApp run];` which was called by
Sys_InitProtocolHandler(). `applicationDidFinishLaunching` was never
called to exit the NSApp run loop.
Use SDL's SDL_DROPFILE event to receive URLs to handle on macOS instead.
This also handles URLs while the game is running (connect to new server)
instead of nothing happening when clicking a link while the game is
running.
This lets the user click a link in a web browser to very easily join a Quake 3 multiplayer game. As browser-based matchmaking websites become more popular, this makes it a lot more convenient and simple to play Quake 3 with others.
The links have the following URI format: quake3://connect/example.com:27950. The format has been designed to be flexible to allow more types of links in the future and avoiding having to make a breaking change. At the moment, "connect" is the only supported command.
Fix crash when scrolling using touchpad on macOS 10.6 (I fixed it
in SDL 2.0.18) and fix failing to start on x86 due to ioq3 SDL dylib
having absolute path instead of "@executable_path/libSDL2-2.0.0.dylib".
SDL 2.24.0 removed macOS 10.6 support (last with x86). Offically at
least, they also dropped x86_64 on macOS 10.7 and 10.8. So ioq3 UB1
is limited to SDL 2.0.22.
- Update x86 and x86_64 to SDL 2.0.22. Compiled with 10.13 SDK as
10.14 SDK dropped running on older than macOS 10.9 and I ran
"strip -x" on the libraries. Ad-hoc signed with PyOxidizer rcodesign.
- PPC is still MAN-AT-ARMS' SDL 2.0.1 build but I ran "strip -x"
on it to reduce the file size.
- ARM64 build was removed from UB1 as UB2 with SDL 2.24.0 should
be used instead.
Update mingw and macOS SDL libraries to 2.24.0. (MSVC .lib files and
Xcode project haven't been updated.)
There is now a separate SDL dylib for macOS Universal Binary 1 and 2
because SDL 2.24.0 dropped support for macOS 10.6 (which is last to
support x86). Documentation says SDL 2.24.0 also dropped macOS 10.7
and 10.8 too. So UB2 is now 10.9 or later while 10.5 to 10.8 live in
UB1.
Compiled using osxcross with macOS 11.3 SDK. Ran "strip -x" on Windows
and macOS libraries. Ad-hoc signed dylib using PyOxidizer rcodesign to
fix digest checksums after strip.
The first chunk of Wavelet sound was decoded as ADPCM and then
following chunks were decoded as Wavelet. It played as static
noise.
Using Wavelet compression requires engine code changes to enable.
When in_restart is called or SDL controller added/remove event is
received, IN_InitJoystick() tried to update the list but instead
of updating the cvar Cvar_Get() just printed a developer message:
Warning: cvar "in_availableJoysticks" given initial values:
"first joystick
", "first joystick
second joystick
"
When using a non-default sound configuration (such as 6 channels), after
a long time (about 4.5hours for 6 channels at 22050 Hz) an overflow will
occur in `S_TransferPaintBuffer`, causing an out of bounds write into
the dma buffer.
The problematic line is:
```
out_idx = (s_paintedtime * dma.channels) % dma.samples;
```
With `s_paintedtime` large enough, the result of the multiplication will
overflow to a negative number (since `s_paintedtime` is signed), and the
index into the output buffer will be negative.
The engine function pointer to vmMain used variadic arguments but
the vmMain function in the dylib has explicit arguments. Evidently
the arguments are stored on the stack and/or registers differently.
Found by fretn.
GL_DEPTH_TEXTURE_MODE defaults to GL_LUMINANCE. Code sets it here as
part of a workaround old nvidia drivers (apparently? the URL in comment
is dead). GL_DEPTH_TEXTURE_MODE was removed in OpenGL 3.0 / 3.2 Core
contexts and may be treated as an error.
The CD key defaulted to spaces. Q3 UI uses text overwrite mode by
default so typing the CD key writes over the spaces. Team Arena UI
uses text insert mode by default so to enter the CD key the user
had to delete the spaces or toggling overwrite mode by pressing
the insert key.
Now the CD key defaults to empty so in Team Arena you can just type
it in.
Printing GL_EXTENSIONS list might be cut off for OpenGL contexts
before 3.0 due to glConfig.extensions_string being a limited length.
Instead get the full extensions list directly.
This was already fixed for OpenGL 3.0 and later contexts.
Using LUMINANCE with OpenGL 3.2 Core contex results in all images being
solid black (in the menu, world, and the console). In the end the whole
screen was solid black except the intro video.
Store images in RGB and RGBA instead. This doesn't affect r_grayscale
visually. If there is issues with it now possibly using compressed
formats, then it would already be an issue for r_grayscale 0.9.
Unlike the Steam and GOG support I'm not doing a registry search as I don't see anything in the registry that would show us the path. If we find more official ways to search for it later we can address that, in the meantime the path is just hardcoded to where it's installed by default.
The ppc arch is the 2.0.1 that was there originaly.
The i386 arch is 2.0.16, now compiled with an SDK later than 10.9 so the Notarization process is cool with it.
So now all post-ppc SDL libraries are 2.0.16.
Also works on Apple Silicon. Specific signing values are in a non-committed file, and the ub2 script only notarizes if a "notarize" flag is passed in on the command line.
NOTE: the SDL dylib currently only has x86_64 and arm64, will need extra work to graft those back in and keep the Notary service happy.
- Reorganized Xcode project
- Added missionpack support to Xcode project
- Incorporated changes from MaddTheSane for Apple Silicon support https://github.com/maddthesane/ioq3
- Built SDL 2.0.14 for Apple Siicon, added to existing SDL dylib
- Built SDLMain for Apple Silicon, added to existing libSDL2main.a
qvmcall64 has its own custom calling convention due to pushing all non-volatile registers to the stack. The game uses set/longjmp which on Windows uses "RtlUnwindEx" to unwind the callstack. qvmcall64 cannot be unwound by default due to the custom calling convention. To allow unwinding, we need to add custom SEH unwind data to the function.
Not all non-volatile registers were actually saved and restored, leading to a few registers being trashed after calling the vm instructions.
All non-volatile registers have been added.
The affect of my 2017 commit c3e64d3806
"Make bots only use q3tourney6 crusher to kill their enemy"
made it easier to beat the last map of the Quake 3 single player
campaign. It was mainly intended for CTF and Team Arena gametypes
variants of the map. (Team deathmatch is usually treated as deathmatch
with teams rather than an actual team based gametype.)
Restore original difficulty for non-CTF/Team Arena gametypes; bot will
crush player even if bot hasn't seen the player since either they or
player respawned. [This is not good in CTF where all five bots on the
team suddenly flip around and shoot at the crusher trigger.]
Team deathmatch still uses my added behavior of not triggering crusher
if there is a teammate under the crusher.
Pulseaudio audio capture didn't stop when paused on Debian 8 but works
on Debian 9 when using the same manual SDL build. So it seems to have
been an issue in pulseaudio, not SDL.
- New msvc142 directory (keeping with naming after version of C compiler)
- Added SDL2.lib/SDL2main.lib back in (compiled from SDL 2.0.8 source from libsdl.org)
- Post build step to copy SDL2.dll to debug directory
There is no need for a tentative definition in this case. By providing
an explicit initializer, the resulting definition has the same effect
as before, but is no longer tentative. This avoids easy errors
resulting from multiple tentative definitions in different translation
units. (Previous compilers permitted those common errors, but GCC 10
rejects them by default.)
My 2017 commit f7c3276 ("Fix g_teamAutoJoin and g_teamForceBalance")
caused BroadcastTeamChange() to be run an extra time when bots connect.
This caused a duplicate center print message to be sent to all clients.
g_teamAutoJoin 1 sent an extra center print when any player or bot
connects. Though that is the original behavior from Team Arena.
Team set during connection no longer sends broadcast as it will be
sent later in ClientConnect().
Reported by Pascal Brochart (pbrochart) as causing clients to be
kicked due to server command overflow if many bots join at the
same time.
If bots join before the local client (dedicated server) the bots would
not be listed in the Q3A team orders menu and so they could not be given
orders using the menu.
The Q3A team orders menu got local client's team while looping through
all clients. Ignoring any bots with clientnum lower than the local
client. Get local client's team before the loop so all bots are
listed.
The lightning trail is fired from the real non-predicted entity for
the local client so it can show the server location. The real client
entity was always used to check if a player entity is firing the
lighting gun. This potentially causes corpses and players on single
player podiums with lightning gun to render the lightning trail if
the real (living) client is firing the lightning gun.
I'm not aware of this causing any issues without modifying the game
logic. Corpses have no weapon and it's not possible to fire during
single player intermission.
GL_CLAMP (clamp to border) was changed to GL_CLAMP_TO_EDGE in 2008
(f2baf359). In 2018 (ce1d5406) I made OpenGL 1.2 be required since
GL_CLAMP_TO_EDGE is used.
Restore support for GL_CLAMP in order to support OpenGL 1.1 like vanilla
Quake 3 does. This should allow using the default Microsoft Windows
GDI Generic OpenGL 1.1 driver (untested but it won't fail the version
check at least).
From gpuinfo.org, it looks like drivers stopped advertising support for
GL_SGIS_texture_edge_clamp so use a version check in addition to the
extension check.
r_allowExtensions 0 disables using GL_CLAMP_TO_EDGE in the opengl1
renderer. GL_CLAMP support wasn't added to the opengl2 renderer.
For lerped frames (refEntity_t frame not equal oldframe) IQM joint
matrices may have incorrect axis scale. This can cause significant model
distortion. The matrix lerp is linear causing each vector to move in a
straight line between frames instead of arcing like a circle. Each joint
frame can have a different scale so can't just normalize the joint
matrix.
Store joints as quaternions and spherical lerp between them and then
convert to a matrix. For my test model, setting up the skeleton is four
times slower now but it still seems to be fast enough to be usable.
r_cubeMapping requires textureCubeLod() which is only in OpenGL 3.0
(GLSL 1.30) and later. It's not in OpenGL ES 3.0 / GLSL ES 3.00.
This needs to be checked before R_InitImages() so can't just check in
GLSL_InitGPUShaders().
Override the video mode list in the Team Arena data files with detected
modes from SDL like in ioquake3's Q3 UI. Add the aspect ratio to the
end of the video resolution (i.e., "640x480 (4:3)"). Add the current
(custom) video mode to the list.
Before when using a custom resolution in the menu you could not change
the video mode using the mouse because the resolution text was blank.
Now custom video resolution is displayed and can be clicked.
Team Arena HUD's team member info box can be cycled through using
nextTeamMember and prevTeamMember commands. nextTeamMember command loops
around when end of list is reached but prevTeamMember command got stuck
at last entry in list (team overview).
My commit last month "Fix SDL audio playback with surround sound" broke
16-bit stereo sound. S_TransferStereo16() still assumed that dma.samples
was a power of two. I also cleaned up code related to the previously
mentioned commit.
