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.
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.
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.
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.
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.
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.
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 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.
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.
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.
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.
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.
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.
* 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.
Game crashed when any of gl_brightfog, gl_lightadditivesurfaces, gl_notexturefill CVARs is set with no level loaded
There was impossible to reset settings to defaults because of this
* only call hw_CheckViewArea if the result is not known yet.
* check the map up front if it even contains heightsecs. This allows to shortcut the above check entirely for maps without sector transfers and will allow further optimizations.
* the MAPINFO options now get handled in g_mapinfo.cpp and g_level.cpp, just like the rest of them as members of level_info_t and FLevelLocals.
* RecalcVertexHeights has been made a member of vertex_t and been moved to p_sectors.cpp.
* the dumpgeometry CCMD has been moved to p_setup.cpp
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.
- 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.
This is mainly to support UMAPINFO which does not have clusters but has been extended to define separate exit texts for each target map that can be reached from a given map.
Special names 'normal' and 'secret' can be used to define texts specific to the default exits.
New MAPINFO properties:
* exittext = mapname, "text"...
* textmusic = mapname, "musicname", order
* textflat = mapname, flatname
* textpic = mapname, picname
textflat and textpic are like 'flat' and 'pic' for clusters, one defines a tiled background, the other a fullscreen image.
Setting an empty exittext will disable a cluster-based text screen that may apply to the given map.
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.
