Having everything lumped together made this a maintenance hassle because it affected how the level has to be stored.
This hasn't been tested yet, so it may not work as intended!
The original place in I_CheckNativeMouse is unsafe because that function can get called from the system message queue which can result in a bad global state of the VM for such a call because it can be recursively invoked from code that may temporarily alter some settings.
currentUILevel is now primaryLevel.
For ZScript, currentVMLevel was added. This is also exported as 'level' and will change as needed.
This also means that no breaking deprecations will be needed in the future, because in order to sandbox a level only 4 variables need to be handled: level, players, playeringame and consoleplayer.
The remaining global variables are not relevant for the level state.
The static 'level' has been mostly removed from the code except some places that still need work.
I think these were the last two still missing it, all remaining uses of the global level variable are in code that doesn't get run through a level tick and are supposed to access the primary level.
Two files were split:
g_level.h contained both the game data definitions and some prototypes belonging to the game logic. These were split up.
decallib.cpp contained both the data and the animation thinkers. The thinkers are now in their own file.
There is one exception in ACS for a net arbitrator check.
Aside from this the bot_observer CVAR was also removed. This was never implemented properly and could stomp upon custom player settings.
The playsim really has no idea what the renderer is supposed to do here and the current system has some serious issues that eventually need addressing. So it is better to just set a flag that an actor needs to have its view interpolation reset if being used as a camera and let the render code deal with it.
This will keep the playsim clean of future changes to this feature.
src/posix/cocoa/i_video.mm:559:2: error: use of undeclared identifier 'atterm'
src/posix/sdl/hardware.cpp:85:28: error: ‘atterm’ was not declared in this scope
src/posix/sdl/i_input.cpp:261:10: error: ‘gamestate’ was not declared in this scope
src/posix/sdl/i_input.cpp:261:23: error: ‘GS_LEVEL’ was not declared in this scope
src/posix/sdl/i_input.cpp:263:48: error: ‘GS_INTERMISSION’ was not declared in this scope
src/posix/sdl/i_input.cpp:263:80: error: ‘GS_FINALE’ was not declared in this scope
src/posix/sdl/i_system.cpp:115:25: error: ‘atterm’ was not declared in this scope
src/posix/sdl/i_system.cpp:372:30: error: invalid use of incomplete type ‘const struct dirent’
src/posix/sdl/i_system.cpp:375:42: error: ‘findstate_t’ has not been declared
src/posix/sdl/i_system.cpp:391:12: error: request for member ‘current’ in ‘* fileinfo’, which is of non-class type ‘int’
src/posix/sdl/i_system.cpp:392:12: error: request for member ‘count’ in ‘* fileinfo’, which is of non-class type ‘int’
src/posix/sdl/i_system.cpp:392:56: error: request for member ‘namelist’ in ‘* fileinfo’, which is of non-class type ‘int’
src/posix/sdl/i_system.cpp:393:22: error: ‘alphasort’ was not declared in this scope
src/posix/sdl/i_system.cpp:393:31: error: ‘scandir’ was not declared in this scope
src/posix/sdl/i_system.cpp:394:16: error: request for member ‘count’ in ‘* fileinfo’, which is of non-class type ‘int’
src/posix/sdl/i_system.cpp:401:31: error: ‘findstate_t’ has not been declared
src/posix/sdl/i_system.cpp:403:2: error: ‘findstate_t’ was not declared in this scope
src/posix/sdl/i_system.cpp:403:15: error: ‘state’ was not declared in this scope
src/posix/sdl/i_system.cpp:403:37: error: expected primary-expression before ‘)’ token
src/posix/sdl/i_system.cpp:404:33: error: request for member ‘count’ in ‘* fileinfo’, which is of non-class type ‘int’
src/posix/sdl/i_system.cpp:406:39: error: request for member ‘count’ in ‘* fileinfo’, which is of non-class type ‘int’
src/posix/sdl/i_system.cpp:413:2: error: ‘findstate_t’ was not declared in this scope
src/posix/sdl/i_system.cpp:413:15: error: ‘state’ was not declared in this scope
src/posix/sdl/i_system.cpp:413:37: error: expected primary-expression before ‘)’ token
src/posix/sdl/i_system.cpp:425:16: error: ‘findstate_t’ was not declared in this scope
src/posix/sdl/i_system.cpp:425:29: error: expected primary-expression before ‘const’
src/posix/sdl/i_system.cpp:426:1: error: expected ‘,’ or ‘;’ before ‘{’ token
src/posix/sdl/i_main.cpp:160:6: error: ‘gamestate’ was not declared in this scope
src/posix/sdl/i_main.cpp:160:19: error: ‘GS_LEVEL’ was not declared in this scope
src/posix/sdl/i_main.cpp:160:44: error: ‘GS_TITLELEVEL’ was not declared in this scope
src/posix/sdl/i_main.cpp:217:18: error: ‘atterm’ was not declared in this scope
src/posix/sdl/i_main.cpp:242:11: error: ‘I_Quit’ was not declared in this scope
src/gameconfigfile.cpp:126:27: error: ‘SHARE_DIR’ was not declared in this scope
src/scripting/decorate/thingdef_parse.cpp:929:2: error: use of undeclared identifier 'strlwr'; did you mean 'strlen'?
src/scripting/decorate/thingdef_parse.cpp:937:3: error: use of undeclared identifier 'strlwr'; did you mean 'strlen'?
src/scripting/decorate/thingdef_states.cpp:564:2: error: use of undeclared identifier 'strlwr'; did you mean 'strlen'?
src/scripting/vm/jit_runtime.cpp:957:2: error: ‘unique_ptr’ is not a member of ‘std’
src/scripting/vm/jit_runtime.cpp:957:38: error: expected primary-expression before ‘>’ token
src/scripting/vm/jit_runtime.cpp:957:40: error: ‘nativeSymbols’ was not declared in this scope
src/utility/cmdlib.cpp:962:2: error: ‘DIR’ was not declared in this scope
src/utility/cmdlib.cpp:962:7: error: ‘directory’ was not declared in this scope
src/utility/cmdlib.cpp:962:34: error: ‘opendir’ was not declared in this scope
src/utility/cmdlib.cpp:967:33: error: ‘readdir’ was not declared in this scope
src/utility/cmdlib.cpp:969:10: error: invalid use of incomplete type ‘struct ScanDirectory(TArray<FFileList>&, const char*)::dirent’
src/utility/cmdlib.cpp:973:34: error: invalid use of incomplete type ‘struct ScanDirectory(TArray<FFileList>&, const char*)::dirent’
src/utility/cmdlib.cpp:985:20: error: ‘closedir’ was not declared in this scope
src\sound\mididevices\music_opl_mididevice.cpp(112): error C3861: 'I_DebugPrint': identifier not found
src\sound\mididevices\music_opl_mididevice.cpp(206): error C3861: 'I_DebugPrint': identifier not found
src\sound\mididevices\music_opl_mididevice.cpp(234): error C3861: 'I_DebugPrint': identifier not found
src\sound\mididevices\music_opl_mididevice.cpp(244): error C3861: 'I_DebugPrint': identifier not found
The menu is a very 'dirty' header, and forcing it to be pulled in with something entirely unrelated is not good - even though only two files include oalsound.h.
- Allows defining of what actor is replacing another for information.
- If multiple arachnotrons, a modder can attribute them as being a replacer of Arachnotron itself, allowing A_BossDeath and GetReplacee to work with it.
This was always used with 'consoleplayer' which really is the only thing making sense here. But this is a part of the global state which should be avoided in play code.
In particular, this makes no real sense in case of secondary maps where it should always return false.
This is supposed to be come the place where all pure play code should be placed, but for that all CVARs and CCMDs and other things that do not directly handle play data should be taken out to make code reviewing easier. These now get collected in two separate files, g_cvars.cpp and g_dumpinfo.cpp respectively.
The sole ZScript property in here has also been moved - to thingdef_properties.cpp.
This made reviewing the code for accessing the global state hard, because the doomdef.h contains mainly constants, this particular item was the only thing in there that represents actual engine state.
The Map loader may not access any global state at all - everything it can touch must be exchangable.
Furthermore, if we want to sandbox each level, there may be no direct access to any kind of global state whatsoever from the play code.
It used the current console player's camera, not the actual camera being used for rendering. Although this is the same most of the time, let's better do it right.
This also removes a few leftover references to the player array elsewhwere in the hardware renderer
All these required access to the sector's Level reference.
The remaining references to the global 'level' variable are all in deprecated functions which is ok.
LevelLocals on the left side of.a function call will now always be remapped to 'Level', which will either remap to the same-named instance variable or the global deprecated one.
In a few degenerate cases where there is a conflicting local variable named 'level' it may error out but that is unavoidable here but this is very unlikely.
- moved parts of the render setup out of the separate render functions.
Things like particle and polyobject linking were duplicated several times for rendering different things in different renderers.
These things only need to be set up once before the renderer is started so it makes a lot more sense to consolidate them into one place outside the actual rendering code.
This had two different flags that were checked totally inconsistently, and one was not even saved.
Moved everything into a few subfunctions so that these checks do not have to be scattered all over the code.
There is no need to do this deep inside the renderer where it required code duplication and made it problematic to execute on multiple levels.
This is now being done before and after the top level call into the renderer in d_main.cpp.
This also serializes the interpolator itself to avoid problems with the Serialize functions adding the interpolations into the list which can only work with a single global instance.
The shader timer may be taken from the primary level for the entire scene, because it will always be the same for all levels in a set.
The camera textures need to be prepared for all levels.
Since currently there is only one level, this will obvciously only run once on that level for the time being.
This is mainly used for CCMDs and CVARs which either print some diagnostics or change some user-settable configuration.
Depending on serialization order is not a good idea here, so now it's no longer stored as a parent in the main level script but explicitly checked for when looking for a variable.
This is what should be audible. To prevent other levels from playing sound, all entry points check whether the sound playing entity belongs to the current UI level.
src/c_dispatch.cpp:143:5: warning: delete called on 'FDelayedCommand' that is abstract but has non-virtual destructor [-Wdelete-non-virtual-dtor]
src/tarray.h:582:5: warning: delete called on 'FDelayedCommand' that is abstract but has non-virtual destructor [-Wdelete-non-virtual-dtor]
src/tarray.h:574:5: warning: delete called on 'FDelayedCommand' that is abstract but has non-virtual destructor [-Wdelete-non-virtual-dtor]
This was done to ensure that this code only runs when the thinker itself is fully set up.
With a constructor there is no control about such things, if some common initialization needs to be done it has to be in the base constructor, but that makes the entire approach chosen here to ensure proper linking into the thinker chains impossible.
ZDoom originally did it that way, which resulted in a very inflexible system and required some awful hacks to let the serializer work with it - the corresponding bSerialOverride flag is now gone.
The only thinker class still having a constructor is DFraggleThinker, because it contains non-serializable data that needs to be initialized in a piece of code that always runs, regardless of whether the object is created explicitly or from a savegame.
Doing this intermingled with the thinkers is highly unsafe because there are absolutely no guarantees about order of execution.
Effectively it ran these commands right in the middle of the playsim which could cause all sorts of synchronization issues, because CCMDs are part of the UI, not the playsim.
- pass a const string to AddCommandString.
This function manipulated the input buffer, leading to all sorts of code contortions to make sure that the passed parameter is clean for that.
This function will now create a copy of the passed parameter which it can manipulate without complicating its calling code.
# Conflicts:
# src/c_dispatch.cpp
This doesn't really write out any info for the pointer, if the level does not match it just errors out.
This is both for quick detection of badly used level data and for automatic restoring of the pointer from the serializer's working level.
This also removed the temporary workarounds in DAutomap and DLevelScript to restore these pointers when a savegame is loaded.
This was a relatively cheap change but removes a significant batch of references to the global variable, only making the entry points to the ACS interpreter relevant.
UI always runs on the primary level, so this does not need the ability to operate on multiple levels. Additionally, this can later be set to null when running play code so that scope violations result in an abort.
This also changes the action special interface to pass a Level parameter to the separate functions and makes a few other minor adjustments to the polyobject code.
This should later be done for everything else as well, but the map loader should really be free of global dependencies ASAP.
Also replace TThinkerIterator<AActor> with FThinkerIterator globally because this only adds pointless type checks - with all actor subclasses being scripted this class has become redundant.
This entered the code path which warned about ambiguous use of variables in action functions and as a result ran afoul of subsequent error checks.
Since ZScript has no global scope resolution operator, this needs to ignore all non-static class symbols and try to look up any of these as global identifiers.
This caused bad calculations with CMF_OFFSETPITCH. Note that to compensate for the fix, the SphericalCoords function had to have its own inversion of the value removed so that it calculates the same result as before.
Apparently some people have to pass positive numbers in here to get a negative pitch, e.g. 350.0 instead of -10.0...
This prevents clamping of such out-of-range values that would otherwise constitute valid pitches with the wraparound in place.
Using global variables for this is bad, and it didn't even catch all cases. Now a node build is only considered successful if everything is set up successfully.
# Conflicts:
# src/maploader/maploader.cpp
src/scripting/decorate/thingdef_parse.cpp:80:11: error: no viable conversion from 'const FName' to 'FString'
src/scripting/zscript/zcc_compile.cpp:1359:26: error: use of undeclared identifier 'Name_globalfreeze'; did you mean 'NAME_globalfreeze'?
# Conflicts:
# src/scripting/zscript/zcc_compile.cpp
They are intentionally omitted from both MAPINFO and compatibility settings.
This removes the last place where it still went through the map-modified versions of the compatflags.
# Conflicts:
# src/s_sound.cpp
This was yet another piece of code that lived or died with the assumption that there can only be one level, stored in global variables.
# Conflicts:
# src/p_saveg.cpp
- It was calling the fallback aiming in the wrong place when it should have been outside the speed check.
- Credit to _mental_ for the base code, but no gotos involved.
There are several places where a temporary change of light mode is needed, all these made this change in the global level struct. Now the change is only local to the active draw info.
If we ever want to refactor the global level data these must not reference the 'level' variable.
The main parts of the map loader cannot use this information, because it can only be created after running the node builder, so it got its own set of index functions instead.
Visual C++ will never statically initialize a class instance where a member field has a default value set, so the DEFINE_ACTION_FUNCTION variants without a direct native call need to be handled differently. The easiest way to do this is to leave out the nullptr default and omit the value in the initializer list. For trailing fields this will always get them nulled.
Also tried to sort them by renderer use but that turned out to be without any performance effect, even though the struct is a bit better aligned now and several bytes shorter.
I have to wonder why it had to use such a complicated implementation that provided no advantages whatsoever.
The new code is just 1/5th of the old one's size and much closer to Hexen's original implementation which also was a simple array but with no means to resize the queue.
This involves passing the level explicitly to many functions. What was done here may seem a bit excessive but at least it covers everything.
Most importantly, the global ActiveThinker pointer has been moved into FLevelLocals and is now getting tracked properly by the level without using dangerous assumptions about how the game organizes its data.
- upload only the dynamic AABB subtree using glBufferSubData
- change internal raytracing stack limit from 16 to 32
- update shadowmap AABB tree after R_SetupFrame for proper frame interpolation
Unlike the other classes, the places where variables from this class were accessed were quite scattered so there isn't much scriptified code. Instead, most of these places are now using the script variable access methods.
This was the last remaining subclass of AActor, meaning that class Actor can now be opened for user-side extensions.
This was the only code using the ViewBob member variable.
This also moves the range check for this variable to its application, because a badly behaved mod can just as easily change it at run time instead of just setting an absurdly large value in the class definition.
Since these can be changed on the placed light actor they have to be read from there, so this is now a pointer in FDynamicLight, just like the other properties that can be user-changed.
Also did some cleanup on the interface so that external code doesn't need to dereference the lightflags pointer but can use utility functions for all flags.
backtrace functions are not present in all libc implementations. Cmake has
module to add external libraries into build if needed so use it to fix build on
Unix systems without backtrace in libc.
ReadString allocates a buffer, so saving it in a local variable and then forgetting it will not free the buffer afterward.
(This should probably be refactored to use some safer methods to read the string than this old-school method...)
This should be less of a drag on the playsim than having each light a separate actor. A quick check with ZDCMP2 showed that the light processing time was reduced to 1/3rd from 0.5 ms to 0.17 ms per tic.
It's also one native actor class less.
Since the SpawnedThings array is still available when polyobjects are spawned it makes no sense to create an expensive linked list in P_SpawnMapThing.
This can be done far better by scanning through the array again and collect all matching items in a second array.
This thing made sense when GZDoom and ZDoom were separate projects to avoid having to change some core files for the added options.
Now, with only 3 ones remaining, one for FraggleScript and two for Extradata the overhead here is just too high. The 3 variables can just be moved to level_info_t without carrying along this much baggage.
Since actors are being spawned before the renderer gets set up this needs to fully initialize the list before spawning the actors, then take it down again for creating the vertex buffer and then recreate it.
Add 'useowncoloradd_{top,mid,bottom}' sidedef properties to the UDMF
spec
Only use side's additive colors if 'useowncoloradd_(top|mid|bottom)' is
set.
Rename UseOwnColors flag to UseOwnSpecialColors
Add UseOwnAdditiveColor flag to side_t::part
Add EnableAdditiveColor to side_t
Add Side.EnableAdditiveColor to ZScript API
Sector.SetAdditiveColor actually called Sector.SetSpecialColor
Add use boolean property, used to determine whether or not to override the sector's additive wall colour with the side's additive colour.
The new specification is more flexible, and allows assigning additive
colors to individual parts of a sector (walls, sprites, flats) and even
individual parts of a side (top, middle, bottom)
Add AdditiveColors arrays to sector_t and side_t::part
Initialize AdditiveColors arrays to 0
Export AdditiveColors to ZScript
Save AdditiveColors in saved game files
Use colors from AdditiveColors arrays when setting the additive color
for the render state
Add code to parse the new UDMF additive color properties
Remove additive color slot from sector color/part enum
Add SetAdditiveColor to sector_t and side_t
Add GetAdditiveColor to side_t
Export new methods and additive color arrays to ZScript
Rename ColorAdd to AddColor
Add AddColor to FRenderState
Tweak SpecialColors array in ZScript to include the additive color
Add uAddColor to the shader compiler
Add uAddColor to the texel
- TriggerPainChance(Name mod, bool forcedPain)
- One exception: PainThrehold is only checked in ReactToDamage, since this function does not require checking damage amount.
- Split off all reactive functionality (pain, infighting, etc) into its own function, ReactToDamage.
- Refactored all DamageMobj's damage <= 0 values.
- Any unconditional cancellations now return -1. ReactToDamage will not be called if values < 0.
- All pain/wound/target changing allowances return 0.
NOFRICTION disables all friction effects on the thing it's set on
(including the speed cap from water/crouching), and NOFRICTIONBOUNCE
disables the "bounce off walls on an icy floor" effect on the thing
it's set on.
This did no longer sort sprites in the same position reliably since the feature to render sprites which only partially are inside a sector was added.
With this, sprites in the same position are no longer guaranteed to be added to the render list in sequence.
Fixed by adding an 'order' field to AActor which gets incremented with each spawned actor and reset when a new level is started.
The software renderer will also need a variation of this fix but its data no longer has access to the defining actor when being sorted, so a bit more work is needed here.
Changed to use 32 bit and also fixed the random number call which was using the byte value variant of the access operator, effectively limiting the number of choices to 256.
Since this deletes the resolved elements one by one and needs to start at the front to ensure consistency, it is better to reverse the order so that the deletions take place at the end of the list which requires a lot less data movement.
On Total Chaos this slowed down texture setup to the point where the mod was basically unlaunchable.
The mod which prompted me to add this is "The Chosen" which is a Dehacked-based TC and repurposes many original actors for something entirely different.
The stock lights are not usable for this and would make it impossible to add a GAMEINFO lump to it because then there is no way to disable loading of lights in the startup screen.
Making callstack resolving operational will require a little bit more effort
src/scripting/vm/jit_runtime.cpp:900:31: error: use of undeclared identifier 'frames'
src/scripting/vm/jit_runtime.cpp:903:23: error: use of undeclared identifier 'cnt'
The ZScript DrawChar function was incomplete and FFont::GetChar did not always return the proper texture.
To make things clearer the OriginalPic is now only used in the few cases where substitution takes place and nothing else.
Since unfortunately this cannot be set as a general default, let's at least make it as easy as possible to disable that panning+scaling madness without having to edit the texture data.
Since the entire font setup is very much incapable of handling this during rendering, short of a complete rewrite, it was necessary to put the relevant code into the places which process the characters for drawing so that it can disable the translation table (which needs to be passed as raw data to the draw functions) and keep track of both the translatable and the original variant of the character graphics.
This particular case incorrectly factored in the sidedef's scaling factor for how to calculate the offset.
Fortunately this is a very rare case - a quick check yielded no maps depending on it.
Should any map surface that depends on this bug a compatibility option may be needed but it doesn't seem likely that this may be the case.
This particular case incorrectly factored in the sidedef's scaling factor for how to calculate the offset.
Fortunately this is a very rare case - a quick check yielded no maps depending on it.
Should any map surface that depends on this bug a compatibility option may be needed but it doesn't seem likely that this may be the case.
Between creation and resolving the texture manager will resort the textures and after that the old ids are no longer valid. The textures themselves have the correct one, so that should be used.
This was leaking memory with being handled like a regular image texture and also would prevent further changes to the in-game texture handling because the savegame picture was imposing some limitations on FPNGTexture's implementation
This is essentially a stripped down version of FHardwareTexture, which can exist on the API independent size, and which stores pointers to hardware textures instead of OpenGL texture handles.
For ad-hoc Dehacked state functions no ArgFlags are created, in this case they can just be assumed to not be relevant here, because none of these function produces reference arguments.
Initially, settings controller flag was false by default
It was not touched during construction and destruction of player_t instances though
Now, with all members initialized in class definition, this flag must be saved and restored manually
https://forum.zdoom.org/viewtopic.php?t=62830
FDummyTexture had a big problem: Whenever it was accessed by accident it crashed the app because it wasn't fully implemented.
What it should do is return empty pixels of the given size, and an unextended FImageTexture is doing just that.
This will mostly ensure that each patch used for composition is only loaded once and automatically unloaded once no longer needed.
So far only for paletted rendering, but the same logic can be used for true color as well.
* split up FMultiPatchTexture into a builder class and the actual image source.
* since images can now be referenced by multiple textures the old redirection mechanism has been removed. It can be done better and less intrusive now. Simple single patch textures already directly reference the underlying patch image now.
* allocate all image source related data from a memory arena. Since this is all static this makes it a lot easier to free this in bulk.
In ZDoom this would affect everything using a patch that got used in a front sky layer, even if the texture was totally unrelated. It is only owed to the low usability of such patches for other purposes that this hasn't caused problems.
Previously it tried to copy all patches of composite sub-images directly onto the main image.
This caused massive complications throughout the entire true color texture code and made any attempt of caching the source data for composition next to impossible because the entire composition process operated on the raw data read from the texture and not some cacheable image. While this may cause more pixel data to be processed, this will be easily offset by being able to reuse patches for multiple textures, once a caching system is in place, which even for the IWADs happens quite frequently.
Removing the now unneeded arguments from the implementation also makes things a lot easier to handle.
* it's no longer the main texture objects managing the pixel buffer but FSoftwareTexture.
* create proper spans for true color textures. The paletted spans only match if the image does not have any translucent pixels.
* create proper warp textures instead of working off the paletted variants.
As a side effect, caching of pixel buffers for texture composition is temporarily disabled, as it management of texture redirections. These things will be reimplemented once things progress further. The existing methods here had their share of serious issues that should be fixed.
This was done to make reviewing easier, again because it is virtually impossible to search for the operators in the code.
Going through this revealed quite a few places where texture animations were on but shouldn't and even more places that did not check PASLVERS, although they were preparing some paletted rendering.
This class has only meaning for software-based warping so it doesn't have to be a part of the FTexture hierarchy.
Making it a subclass of FSoftwareTexture is fully sufficient.
As a bonus this already fixes several bugs caused by the botched texture scaling implementation the original texture manager came with.
System cursors are currently disabled because they rely on functionality that needs to be moved to different classes.
- disallow bool as a return value for direct native calls because it only sets the lowest 8 bits of the return register.
- changed return type for several functions from bool to int where the return type was the only thing blocking use as direct native call.
CMakeFiles/zdoom.dir/r_utility.cpp.o: In function `R_SetupFrame(FRenderViewpoint&, FViewWindow&, AActor*)':
/home/travis/build/coelckers/gzdoom/src/r_utility.cpp:832: undefined reference to `AActor::GetCameraHeight() const'
CMakeFiles/zdoom.dir/g_shared/a_action.cpp.o: In function `A_Unblock(AActor*, bool)':
/home/travis/build/coelckers/gzdoom/src/g_shared/a_action.cpp:64: undefined reference to `AActor::GetDropItems() const'
CMakeFiles/zdoom.dir/scripting/vmthunks_actors.cpp.o: In function `AF_AActor_Distance2DSquared(VMValue*, int, VMReturn*, int)':
vmthunks_actors.cpp:(.text+0x6ab): undefined reference to `AActor::PosRelative(AActor const*) const'
CMakeFiles/zdoom.dir/scripting/vmthunks_actors.cpp.o: In function `AF_AActor_Distance3DSquared(VMValue*, int, VMReturn*, int)':
vmthunks_actors.cpp:(.text+0x75f): undefined reference to `AActor::PosRelative(AActor const*) const'
CMakeFiles/zdoom.dir/scripting/vmthunks_actors.cpp.o: In function `AF_AActor_Vec3To(VMValue*, int, VMReturn*, int)':
vmthunks_actors.cpp:(.text+0x850): undefined reference to `AActor::PosRelative(AActor const*) const'
CMakeFiles/zdoom.dir/scripting/vmthunks_actors.cpp.o: In function `AF_AActor_Vec2To(VMValue*, int, VMReturn*, int)':
vmthunks_actors.cpp:(.text+0x930): undefined reference to `AActor::PosRelative(AActor const*) const'
CMakeFiles/zdoom.dir/scripting/vmthunks_actors.cpp.o: In function `AF_AActor_DistanceBySpeed(VMValue*, int, VMReturn*, int)':
vmthunks_actors.cpp:(.text+0xa00): undefined reference to `AActor::PosRelative(AActor const*) const'
CMakeFiles/zdoom.dir/scripting/vmthunks_actors.cpp.o:vmthunks_actors.cpp:(.text+0xad3): more undefined references to `AActor::PosRelative(AActor const*) const' follow
CMakeFiles/zdoom.dir/scripting/vmthunks_actors.cpp.o: In function `AF_AActor_PosRelative(VMValue*, int, VMReturn*, int)':
vmthunks_actors.cpp:(.text+0x1200): undefined reference to `AActor::PosRelative(sector_t*) const'
CMakeFiles/zdoom.dir/scripting/vmthunks_actors.cpp.o: In function `AF_AActor_ClearInterpolation(VMValue*, int, VMReturn*, int)':
vmthunks_actors.cpp:(.text+0x127d): undefined reference to `AActor::ClearInterpolation()'
vmthunks_actors.cpp:(.text+0x1298): undefined reference to `AActor::ClearInterpolation()'
CMakeFiles/zdoom.dir/scripting/vmthunks_actors.cpp.o: In function `AF_AActor_GetBobOffset(VMValue*, int, VMReturn*, int)':
vmthunks_actors.cpp:(.text+0x135a): undefined reference to `AActor::GetBobOffset(double) const'
CMakeFiles/zdoom.dir/scripting/vmthunks_actors.cpp.o: In function `AF_AActor_Distance2D(VMValue*, int, VMReturn*, int)':
vmthunks_actors.cpp:(.text+0x222b): undefined reference to `AActor::PosRelative(AActor const*) const'
* use I_Error instead of I_FatalError to abort. I_FatalError is only for things that are not recoverable and should not be handled outside of error cleanup and rethrowing.
* only catch CRecoverableError in JitCompile. Everything else should fall through to the outermost catch block.
* Do not I_FatalError out after handling the exception locally. Just print an error and return null, indicating failure.
Temple of the Lizardmen 3 has segs lumps in every level that seem to use a different data format and are completely unusable, up to triggering undefined behavior.
* The functions had no prototype and caused crashes.
* even after creating a prototype it didn't work because CreateAnonymousFunction was set up incorrectly for the case where a known return type was given.
- fixed: P_PoisonDamage's check for Buddha2 and the Buddha powerup were inverted.
- fixed: P_PoisonDamage checked the modified damage instead of the raw, allowing amplifiers to boost the damage beyond telefrag and circumventing regular buddha.
This was the last bit of code standing in the way of making AInventory a fully scripted class.
All that's left to sort out is some variable accesses - the vast majority of them in SBARINFO.
- moved the ALTHUDCF parser PClass::StaticInit, so that it gets done right after creating the actor definitions.
All left to do is not to reallocate the AltHud object for each frame but store it in a better suited place.
Instead of overriding the Massacre method it is preferable to clear the flags causing the bad behavior, most notably ISMONSTER.
# Conflicts:
# src/g_inventory/a_pickups.cpp
# src/g_inventory/a_pickups.h
Now a child type can decide for itself how to treat 'amount'.
The scripting interfaces to this function in ACS and FraggleScript have been consolidated and also scriptified.
- offloaded key list generation for alternative HUD to non-UI parts.
This change also revealed a problem with handling empty sprites in the key list so this got fixed, too.
- changed PARAM_STRING to use the passed string by reference instead of by value. The 3 instances where passing by value was needed now use PARAM_STRING_VAL.
* Attempts to fix errors introduced by merging of asmjit branch and weapons scripting branch
* This didn't work. The others should work though.
* fix scriptutil.cpp:(.text+0x18d): undefined reference to `PClass::FindActor(FName)'
* Attempt to fix compilation on macOS
On Windows none of this is needed, because we can generate a proper unwind frame for the JITed functions, but even on Linux, it would require manual additions to each single piece of native code that ever gets called from inside a JIT compiled function.
This is an utterly prohibitive proposition because it makes direct native calls a virtual impossibility
So, in order to get the thrown error properly presented both I_Error and ThrowAbortException will now forward to I_FatalError if it is called from inside a JIT context.
This was by far the largest block of native virtuals, and they were only native to be able to allow checking if the event was implemented for the current handler. This can easily be done by looking at the byte code, just like VMCall also does but in turn it removes more than half of the existing native virtuals from the interface.
Since this function creates dynamic copies for 3D floors that need to be split it requires the vertex buffer index to be set up.
In older versions this did not produce errors because there was a fallback render path that was less efficient.
Now with that fallback removed this resulted in temporary 3D floors being created without valid vertex data.
It has been like this initially but was changed when ZDoom gained an overly complicated polymorphic class descriptor object that required a lot of support code. All these complications have long been removed but these methods remained. Since they prevent a class from being moved to the script side entirely they had to be removed.
This was the last major blocker to make Weapon a purely scripted class, the only remaining native method is Serialize which is of no concern for the coming work.
This stuff is now kept locally in the bot code so that it doesn't infest the rest of the engine.
And please don't read the new botsupp.txt file as some new means to configure bots! This was merely done to get this data out of the way.
The bots are still broken beyond repair and virtually unusable, even if proper data is provided for all weapons.
In both cases, having this flag on will render the monster-backing-off-check for melee attacks ineffective because it would misinterpret these weapons as close range only - which they aren't. Even for the PhoenixRod the range is longer than what gets checked here.
As a consequence, the bot's check for missile shooting melee weapons has also become pointless because no such weapon is defined anymore.
src/doomerrors.h:74:14: error: exception specification of overriding function is more lax than base version
src/posix/sdl/i_main.cpp:272:28: error: 'class std::exception' has no member named 'GetMessage'
Since it forwards directly to FindState and has no script bindings there is no need to keep it, it'd only complicate the full scriptification of the weapon class if it stuck around.
* it was never saved in savegames, leaving the state of dead bodies undefined
* it shouldn't be subjected to pointer substitution because all it contains is old dead bodies, not live ones.
Since the only thing it gets used for is swapping out PlayerPawns it can safely skip all global variables that never point to a live player, which allowed to remove quite a bit of code here that stood in the way of scriptifying more content
Null pointers must be allowed and non-object pointers which are not null must be explicitly checked for because the code could crash on them when performing a static_cast on an incorrect type.
For the varargs functions that used the Type field to validate their parameters, now a hidden additional argument is passed which contains a byte array with the type info for the current call's arguments. Since this is static per call location it can be better prepared once when the code is being compiled instead of being put in a runtime created array for each invocation. Everything else uses the per-function instance of the same data.
The only thing that still needed the type field with a VMValue is the defaults array, so this uses a different struct type now to store its data.
Currently used for loading parameters into registers.
For checking parameters of native functions some more work is needed to get the info to the function. Currently it doesn't receive the function descriptor.
src/scripting/vm/jit_call.cpp:164:38: warning: offset of on non-standard-layout type 'VMScriptFunction' [-Winvalid-offsetof]
src/scripting/vm/jit_load.cpp:87:50: warning: offset of on non-standard-layout type 'DObject' [-Winvalid-offsetof]
src/scripting/vm/jit_load.cpp:96:50: warning: offset of on non-standard-layout type 'DObject' [-Winvalid-offsetof]
src/scripting/vm/jit_load.cpp:257:53: warning: offset of on non-standard-layout type 'DObject' [-Winvalid-offsetof]
The amount of support code for this minor optimization was quite large and this stood in the way of streamlining the VM's calling convention, so it was preferable to remove it before moving on.
This reuses the FTexCoordInfo class the hardware renderer had been using to calculate wall texture offsetting.
The software renderers still need this sorted out to bring them in line with the rest of the code, though, but they do not have this code sufficiently well organized to make this a straightforward task.
With a proper count value available this can be done properly. The only relevant targets are the jumps immediately succeeding the IJMP instructions, nothing else.
This removes the last non-vararg cases where a native VM function checks 'numparam'. As of this commit all function calls will pass the complete list of arguments.
This isn't used for the 3 action function calls because it requires an array allocation which would be a bit too costly for something as frequently called as action functions.
They will need a different approach.
This allows retaining the functionality, even if for the JIT compiler's benefit all default arguments are pushed onto the stack instead of reading them from the defaults array.
There were some issues here:
* a check for mismatching count is too strict because it is legal to omit return values
* it failed to detect returning multiple values in a single expression.
This can happen if a state that's retrieved with FindState gets used with a different actor type and can lead to hard to trace problems if not checked.
This will now both exclude floor caps when only ceiling elements are used and everything outside the bounding box of active portal lines.
Hopefully this is enough to fix the issues with portal caps but of course it is not foolproof if someone just makes the right setup.
The main reason is to unify the portal hierarchy again. The split into a hardware independent and a hardware dependent part turned out to be unnecessary and complicated matters.
Another issue was that the new stencil setup code was having a few subtle problems, so this recreates the original ones with indirect API calls.
