This does not work with a setup where the same backend is driving both renderers.
Most of this is now routed through 'screen', and the decision between renderers has to be made inside the actual render functions.
The software renderer is still driven by a thin opaque interface to keep it mostly an isolated module.
This was originally invented to fix the sprite offsets for the hardware renderer.
Changed it so that it doesn't override the original offsets but acts as a second set.
A new CVAR has been added to allow controlling the behavior per renderer.
No more locking insanity! :)
There are no locking counters or other saveguards here that would complicate the implementation because there's precisely two places where this buffer must be locked - the RenderView functions of the regular and poly SW renderer which cannot be called recursively.
In its current form this is quite useless. What's really needed is to require a lock on the RenderBuffer for the 3D scene, but since this is not needed for the 2D stuff anymore it can be done far simpler.
This was done mainly to reduce the amount of occurences of the word FTexture but it immediately helped detect two small and mostly harmless bugs that were found due to the stricter type checks.
When running in a confined environment (such as a snap) it may not be
possible to write to directories such as ~/.config. By using the $HOME
variable instead of the '~' shortcut, the confined environment can pass
an alternative 'home' directory with write privelges.
I have only changed this for posix/unix and haven't touched code for
MacOS, as I don't know if that behaves differently
Since this is a non-standard function it's better kept to as few places as possible, so now DirEntryExists returns an additional flag to say what type an entry is and is being used nearly everywhere where stat was used, excluding a few low level parts in the POSIX code.
- now that the frame buffer stores its render time, the 'ms' return from I_GetTimeFrac is not needed anymore, we may just as well use the globally stored value instead.
The only feature this value was ever used for was texture warping.
Since this calls I_WaitVBL, which resets the frame time, it was essentially just like calling a real-time timer anyway and nothing in it required a specific 0-timepoint.
The same applies to the ZScript interface. All it needs is a millisecond-precise timer with no semantics attached.
* store the frame time in the current screen buffer from where all render code can access it.
* replace some uses of I_MSTime with I_FPSTime, because they should not use a per-frame timer. The only one left is the wipe code but even this doesn't look like it needs either a per-frame timer or a timer counting from the start of the playsim.
- moved timer definitions into their own header/source files. d_main is not the right place for this.
- removed some leftover cruft from the old timer code.
- Added: d_main.cpp now searches for "gzdoom_optional_assets.pk3" - this can be changed in version.h for fork authors.
- Updated forum links to point to ZDoom.org.
The old code went through a list of predefined file names and looked each of them up in a list of predefined directories until it found a match. This made it nearly impossible to add custom IWAD support because the list of valid file names could not be extended.
This has now been switched around to run a scan for matching files on each given directory. With this approach it can look for *.iwad and *.ipk3 as IWAD extensions as well and read an IWADINFO out of these files that can be added to the internal list of IWADs, making it finally possible to define custom IWADs without having to add them to the internal list.
(This isn't fully tested yet so some errors may still occur.)
This fixes two issues:
* timer related texture animations are not being recreated multiple times if a scene renders multiple viewpoints (e.g. camera textures or portals.)
* interpolation is smoother when maps have a high think time of multiple milliseconds. A good map to see the difference would be ZDCMP2 which has a think time of 4-5 milliseconds. With the timer taken in real time after the thinkers have run and VSync on this resulted in alternating time slices of 11 and 21 ms between frame interpolations instead of an even 16 as should be done for smooth 60 fps because roughly every second frame was offset by those 5 ms.
It now works the following way:
(0) - Force off (ZDoom defaults)
(1) - Force on (Doom defaults)
(2) - Auto off (Prefer ZDoom defaults - if DEHACKED is detected with no ZSCRIPT it will turn on) (default)
(3) - Auto on (Prefer Doom defaults - if DECORATE is detected with no ZSCRIPT it will turn off)
For some files that had the Doom Source license attached but saw heavy external contributions over the years I added a special note to license all original ZDoom code under BSD.
This is to ensure that the Class pointer can be set right on creation. ZDoom had always depended on handling this lazily which poses some problems for the VM.
So now there is a variadic Create<classtype> function taking care of that, but to ensure that it gets used, direct access to the new operator has been blocked.
This also neccessitated making DArgs a regular object because they get created before the type system is up. Since the few uses of DArgs are easily controllable this wasn't a big issue.
- did a bit of optimization on the bots' decision making whether to pick up a health item or not.
See https://forum.drdteam.org/viewtopic.php?t=7588
Processing order is now the same as in Chocolate Doom
prBoom+ loads separate files after all WAD lumps though
This makes sense but would change loading sequence existed in ZDoom for years
- activated the RenderOverlay event, now that it can be called from the correct spot, i.e. right after the top level HUD messages are drawn. The system's status output will still be drawn on top of them.
This allows using the UI scale or its own value, like all other scaling values.
In addition there is a choice between preserving equal pixel size or aspect ratio because the squashed non-corrected versions tend to look odd, but since proper scaling requires ununiform pixel sizes it is an option.
- changed how status bar sizes are being handled.
This has to recalculate all scaling and positioning factors, which can cause problems if the drawer leaves with some temporary values that do not reflect the status bar as a whole.
Changed it so that the status bar stores the base values and restores them after drawing is complete.
- better handling of ForceScale for the fullscreen HUD that doesn't mess around with CVARs.
- moved the mug shot into the status bar, because this is global state that needs to be shared between different pieces of code which want to display a mug shot.
- SBARINFO should work off the current status bar settings instead of the ones stored in its script object
- decided to ditch the widget system I had started to lay out. As it turns out that would make things far more complicated and slower than they need to be.
Note that the Strife status bar does not draw the health bars yet. I tried to replace the hacky custom texture with a single fill operation but had to find out that all the coordinate mangling for the status bar is being done deep in the video code. This needs to be fixed before this can be made to work.
Currently this is not usable in mods because they cannot initialize custom status bars yet.
Since the true color software renderer also handles them there is no point keeping them on the GL side.
This also optimized how they are stored, because we no longer need to be aware of a base engine which doesn't have them.
There's simply never enough of them and they are used far too infrequently to justify the hassle of tagging along two TMaps per class.
For what they provide, single global lists that handle all player classes at once are fully sufficient.
The math in DCanvas::FillBorder does not always work out so better clean the entire screen before drawing a fullscreen image to ensure that the menu blend is always drawn over something valid.
CVAR with this flag can be set in console or from command line when sv_cheats is enabled
There is no such restriction for changing its value from ACS, via SetCVar() and related functions
'cheat' modifier can be used in CVARINFO lump to create variable of this kind
The idea is, when status bars are moved to ZScript that only this small wrapper class needs to be dealt with and the implementation can be left alone. SBARINFO is far too complex to be scriptified, but having it inherit directly from DBaseStatusBar and access its member variables severely limits the options of dealing with the status bar code. This way, it only accesses some globally visible functions in DBaseStatusBar and no variables.
- renamed the global ST_X and ST_Y variables because it is far too confusing and error-prone to have the same names inside and outside DBaseStatusBar.
This was done to ensure it can be properly overridden in scripts without causing problems when called during engine shutdown for the type and symbol objects the VM needs to work and to have the scripted version always run first.
Since the scripted OnDestroy method never calls the native version - the native one is run after the scripted one - this can be simply skipped over during shutdown.
The original implementation just printed a mostly information-free message and then went on as if nothing has happened, making it ridiculously easy to write broken code and release it. Changed it to:
* Any VMAbortException will now terminate the game session and go back to the console.
* It will also print a VM stack trace with all open functions, including source file and line numbers pointing to the problem spots. For this the relevant information had to be added to the VMScriptFunction class.
An interesting effect here was that just throwing the exception object increased the VM's Exec function's stack size from 900 bytes to 70kb, because the compiler allocates a separate local buffer for every single instance of the exception object.
The obvious solution was to put this part into a subfunction so that it won't pollute the Exec function's own stack frame. Interesting side effect of this: Exec's stack requirement went down from 900 bytes to 600 bytes. This is still on the high side but already a lot better.
- instead add a list of SpecialInits to VMScriptFunction so this can be done transparently when setting up and popping the stack frame. The only drawback is that this requires permanent allocation of stack objects for the entire lifetime of a function but this is a relatively small tradeoff for significantly reduced maintenance work throughout.
- removed most #include "vm.h", because nearly all files already pull this in through dobject.h.
* everything related to scripting is now placed in a subdirectory 'scripting', which itself is separated into DECORATE, ZSCRIPT, the VM and code generation.
* a few items have been moved to different headers so that the DECORATE parser definitions can mostly be kept local. The only exception at the moment is the flags interface on which 3 source files depend.
- _FPU_GETCW is defined for more than just x87. Don't use it if the
control word for the target architecture doesn't support _FPU_EXTENDED
or _FPU_DOUBLE defined, e.g. pretty much anything but x87. If I had been
using glibc on PowerPC instead of Apple's libc, I probably would have
noticed this sooner, since _FPU_GETCW is part of glibc.
- handle a 'restart' CCMD a bit more controlled. Instead of throwing an exception in the CCMD handler it now just flags D_DoomLoop to return.
If the exception is thrown within the CCMD this can easily happen deep inside the renderer when it calls NetUpdate. But since the software renderer with its use of global variables is not equipped to be yanked out of lile this it could leave broken data behind that caused glitches or even crashes on subsequently played maps.
This cuts down on as much message noise as possible, outputs everything to a file specified as a parameter and then quits immediately, allowing this to run from a batch that's supposed to check a larger list of files for errors.
Multiple outputs get appended if the file already exists.
- This isn't a real file or even a name, but the game would try and load
it, including running through various permutations, potentially resulting
in loading the current directory as an archive.
This requires quite a bit more thorough cleanup. I got it to the point where the titlepic appears after restarting, but it still crashes when starting the game so there's more data that needs to be cleaned up...
Conflicts:
src/p_mobj.cpp
(This stops right before moving the conversation IDs into MAPINFO because that feature is quite conflict-heavy and will have to merged by itself.)
Conflicts:
src/d_main.cpp
src/info.cpp
src/p_local.h
(Had to merge this all by itself because it was creating too many merge conflicts when combined with other stuff.
UpdateSounds will not be called during screen wipes and the entire setup of this function suggests that this is not advisable at all.
The OpenAL stream updates were done deep inside this function implicitly.
This caused music to stop while a wipe was in progress. So in order to allow uninterrupted music playback during screen wipes the music updates need to be handled separately from sound updates and be called both in the main loop and the wipe loop.
I think that the OpenAL music updating should be offloaded to a separate thread but at least it's working now without causing interruptions during wipes.
- Old mess:
* Execute autoexec files right away.
* Execute -exec files right away.
* Execute command line commands right away.
- If, during any of the above, an unknown command or a set of an
unknown variable is encountered, store it for later.
- Pullin commands are directly executed and add to the list of files
to load.
* Do a little setup, including parsing CVARINFOs.
* Retry saved commands in case CVARINFO added a cvar they refer to.
- New, less messy, mess:
* Parse autoexec files into an array.
* Parse -exec files.
* Parse command line commands.
- During all of the above, exec commands are also parsed into the
array immediately rather than being saved for execution later.
- Pullin commands are parsed into a different array. The pullin
command doesn't actually do anything directly anymore.
* Add all the pullin files to the list of files to load.
* Do a little setup, including parsing CVARINFOs.
* Execute every command that was parsed in the preceding steps.
This has an important implication:
Previously the config was loaded before IWADINFO so in order to allow the config to access the data this had to be switched around.
This means that zdoom.pk3 will not be looked for in the global IWAD search paths anymore, but since it shouldn't be there to begin with it should be an acceptable compromise.
As a result the old 'Group' property could be removed and all other means to get a section name were disabled.
As an example, if the code gets 'doom.doom2.commercial' it will use the following sections in this order:
global.autoload
doom.autoload
doom.doom2.autoload
doom.doom2.commercial.autoload.
- Multi-directory archives (e.g. zips) now support filtering lumps
depending on the loaded IWAD. The search rules are the same as for the
Autoload entries in the user's ini. For instance, if you are playing
Doom 2, the following filters will be applied:
* "filter/doom2/*"
* "filter/doom/*"
They will be renamed to strip out the "filter/doom2/" and "filter/doom/"
parts and will be ordered so they take precedence over any files not
inside a filter/ directory. Any files inside another filter/ directory
(e.g. "filter/hexen/*") will be ignored.
- Fixed: Using +set cvarname and +cvarname on the command line would not
work if cvarname was defined in CVARINFO. This should be the proper way to fix
it. Rather than move all command line execution after loading CVARINFO,
keep command line execution before wads are loaded. If an attempt is
made to set an unknown cvar or to run an unknown command (which could
potentially be shorthand for setting an unknown cvar), save it and try
running it again after all CVARINFOs have been handled.
Conflicts:
src/CMakeLists.txt
src/b_think.cpp
src/g_doom/a_doomweaps.cpp
src/g_hexen/a_clericstaff.cpp
src/g_hexen/a_fighterplayer.cpp
src/namedef.h
src/p_enemy.cpp
src/p_local.h
src/p_mobj.cpp
src/p_teleport.cpp
src/sc_man_tokens.h
src/thingdef/thingdef_codeptr.cpp
src/thingdef/thingdef_function.cpp
src/thingdef/thingdef_parse.cpp
wadsrc/static/actors/actor.txt
wadsrc/static/actors/constants.txt
wadsrc/static/actors/shared/inventory.txt
- Added register reuse to VMFunctionBuilder for FxPick's code emitter.
- Note to self: Need to reimplement IsPointerEqual and CheckClass, which
were added to thingdef_function.cpp over the past year, as this file no
longer exists in this branch.
Even when '+logfile' argument was omitted, the console would print 'Could not start log', because 'logfile != NULL' was used as a check for the presence of '+logfile' argument, but the internal buffer of FString is never NULL, so the right check is 'logfile.isNotEmpty()'.
While I'm at it, I fixed another bad check for 'pagename'.
