Build 7792
SVN-
General
Fix translucent masked wall sorting issue in Polymost
-
General
Voxel fix
-
General
Rename "autosaves" to "checkpoints" in menu
SVN handles and Git names are merged into one person.
Fix translucent masked wall sorting issue in Polymost
Voxel fix
Rename "autosaves" to "checkpoints" in menu
Fix stupid bug with sprite shadows in Polymost
Voxel fixes from Nuke
Fix issue with shader compilation
Fix sequential GNUmakefile build
Support separate .glsl shaders in Windows MSVC builds
osdfuncs.cpp: fix OSD background sometimes not rendering with the correct gl state
Support clamping in polymost1Frag.glsl and use it for sprites
Separate out polymost1Frag and polymost1Vert shaders from polymost.cpp and add rules for building in the GNUmakefile.
Additionally, git ignore the shader's generated headers folder.
Fix volume slider bug introduced in r7773
QOL change: print save filename to console for manual savegames, too, not just ones generated through the script
Voxel fixes from Nuke.YKT
Don't attempt to animate sprites drawn as voxels by picnum
This is just syntax and formatting
No functional changes.
Rename stupid structure member I probably named when I was high
Stop trying to make CTRL-ALT-DEL happen
It's not going to happen.
Patch from Nuke.YKT to replace audiolib's volume table with floating point calculations
This fixes an issue where rapidly changing the player's angle could result in the audio "popping" due to the sudden change to a new position in the volume table, and the resulting difference in volume between one sample and the next. This also fixes a DC offset issue when applying the volume level to certain sample types.
Untie EVENT_ALTFIRE from the conditions required to trigger EVENT_FIRE
Add gamefunc_Alt_Weapon and EVENT_ALTWEAPON back in
Make shadows on voxels in Polymost look like they do in classic
Even more minor fix to previous commit
Minor fix to last commit
Add tilehasmodelorvoxel() function and fix instances where voxels for multi-sided sprites would get flipped based on the viewing angle
Fix audio position updating in automap mode
Prevent addlogvar from accepting constants, which causes a crash
Make cache1d try a little harder to find a suitable block before bailing with "CACHE SPACE ALL LOCKED UP!"
Make cacheAgeEntries() do a more consistent amount of work per frame
Swap indexes for gamefunc_AutoRun and gamefunc_Alt_Fire
Normally I wouldn't be in favor of changing indexes that are already exposed to scripts around, but gamefunc_AutoRun is a locally handled key that doesn't have any script events associated with it, and gamefunc_Alt_Fire is pretty important to have up near the top wherever gamefuncs are listed.
Replace gamefunc_Alt_Weapon with a gamefunc_Alt_Fire that works as expected
Improve depth fighting by pushing out the near plane and improve near plane clipping by depth clamping.
Suggest reducing SCISDIST when there's time to debug SCISDIST issues.
If GL_ARB_clip_control & GL_ARB_depth_clamp are available, reflect that in glinfo
Update GLAD to 0.1.30 and add GL_ARB_clip_control, GL_ARB_depth_clamp, and GL_ARB_shader_texture_lod extensions
game.cpp: temporarily limit A_Spawn() owner changes from r7738 to locators
Mapster32 Polymost cursor selection omission
Disable -malign-double on i386 as libraries would need to be compiled with it.
Fix 32-bit Polymost issues by enabling SSE instructions and setting GCC to use SSE floating point math exclusively.
Additionally, enable SSE2 to better match 64-bit builds and use -malign-double for the potential performance benefit.
Gump redux
Fix builds on platforms other than win32--oops
Clarify incompatible savegame log message
I am not a smart man
Fix building with GCC 8
Reintroduce mouse selection patch from Nuke.YKT
Revert "Patch from Nuke.YKT to improve mouse selection in the editor in Polymost"
Looks like commits made to public SVN after this commit in my local tree resulted in a merge error
Patch from Nuke.YKT to improve mouse selection in the editor in Polymost
actors.cpp: Change demo camera path following behaviour to search for locators based on a hitag channel to avoid locator lotag contention.
Now, destination and subject locators should be assigned one hitag channel. Destination locators will be visited starting from lotag 0 and counting upward. Subject locators can have any lotag within the same hitag channel, and may also be a destination if the lotag is not separated from the destination locator ids. Destination locators must now specify their subject locator in their owner field by lotag.
CON_USE_COMPUTED_GOTO eval macro micro-optimization
This is almost as fast as ditching the bounds check here entirely.
Neat Polymost y-shearing patch from Nuke.YKT
Fix old issue with VM_CONDITIONAL macro
Rename Gv_GetArrayCountFromFile() to Gv_GetArrayCountForAllocSize()
Rename Gv_SetVarX() to Gv_SetVar() (overloaded)