Commit graph

12493 commits

Author SHA1 Message Date
Christoph Oelckers
0d7b7d6ab1 - merged PStruct and PNativeStruct.
There were exactly 4 places in the code that checked for the difference, this is better done with a flag.
2017-04-12 15:12:41 +02:00
Christoph Oelckers
63eb3e331e - un-const-ify some functions. 2017-04-12 14:40:29 +02:00
Christoph Oelckers
abc4481431 - set PointedType to null for PClassPointer to simplify the functions for this class. 2017-04-12 13:11:59 +02:00
Christoph Oelckers
6308251084 - added a FindSymbol wrapper to PClass so that other code does not need to access the symbol table directly.
- added an isActor utility function to codegen.cpp to simplify a few constructs.
2017-04-12 13:08:41 +02:00
Christoph Oelckers
b17ab7a133 - put PObjectPointer to some use.
- fixed: There was no check for compatible class pointers.
2017-04-12 11:40:40 +02:00
Christoph Oelckers
bc486904cd - split PPointer into PPointer and PObjectPointer.
A pointer to an object has quite different semantics so let's do this with proper virtual inheritance. This should allow to simplify a lot of pointer related checks in the compiler.
2017-04-12 10:52:54 +02:00
Christoph Oelckers
80801d11b1 - removed some redundant static_casts. 2017-04-12 10:29:04 +02:00
Christoph Oelckers
9c9b2ccf6d - replaced more dyn_casts and checks for RUNTIME_CLASS(PClassActor)
It is preferable to use IsDescendantOf wherever possible if we ever want to be able to separate PClass from PType.
2017-04-12 10:20:58 +02:00
Christoph Oelckers
5350721ec5 - removed PClass::DeriveData because it is no longer needed.
- fixed: IsVisibleToPlayer and optimized it a bit more.
- moved SourceLumpName to PClass, so that it can also be used for non-actors (there's a lot of non-Actor classes already.)
- separated the serializer for PClassPointer from PPointer. Even though not doable, a pointer to a class type is something entirely different than a class pointer with a restriction so each should handle its own case.
2017-04-12 09:55:27 +02:00
Christoph Oelckers
0e0eca0e0f - replaced some dyn_casts with calls to FindActor.
- fixed type checks in A_Morph.
2017-04-12 01:29:51 +02:00
Christoph Oelckers
e4d2380775 - moved all remaining fields from PClassActor to FActorInfo.
- added a few access functions for FActorInfo variables.

With PClassActor now empty the class descriptors can finally be converted back to static data outside the class hierarchy, like they were before the scripting merge, and untangle the game data from VM internals.
2017-04-12 00:07:41 +02:00
Christoph Oelckers
854053a14f - use TArrays instead of TMaps to store damage factors and pain chances.
For these fields maps have no advantage. Linearly searching a small array with up to 10 entries is nearly always faster than generating a hash for finding the entry in the map.
2017-04-11 23:29:37 +02:00
Christoph Oelckers
4afe2d4218 - moved OwnedStates and NumOwnedStates out of PClassActor. 2017-04-11 22:44:35 +02:00
Christoph Oelckers
05240ccbe5 - fixed redundant reallocation of constructable meta fields.
- some optimization of access to OwnedStates in old DECORATE.
- consolidate all places that print a state name into a subfunction.
- allocate states from the ClassDataAllocator memory arena. States do not need to be freed separately from the rest of the static class data.
2017-04-11 21:48:41 +02:00
Christoph Oelckers
311ce2362a - started moving stuff out of PClassActor into meta data.
This reinstates the old FActorInfo as part of the meta data a class can have so that the class descriptor itself can be freed from any data not directly relevant for managing the class's type information.
2017-04-11 19:37:56 +02:00
Christoph Oelckers
6a3ddaa8fa - moved Restricted/ForbiddenToPlayerClass fully to the script side.
This required some fixes for allowing to read from metadata arrays.
2017-04-11 15:11:13 +02:00
Christoph Oelckers
45691e91c9 - removed PClassActor::PointerSubstitution.
This is not needed anymore because classes do not need to be replaced. The only reason this was implemented was the original design with the class descriptors taking on all the metadata themselves.
2017-04-11 14:03:49 +02:00
Christoph Oelckers
1febf277af - fixed crash on bad PNGs.
The cast to a signed long could create negative numbers which failed the sanity check and caused a stack corruption.
2017-04-11 10:29:52 +02:00
Magnus Norddahl
f049e6145b - remove all usage of std::make_unique to keep things C++11 compliant 2017-04-11 01:25:44 +02:00
Magnus Norddahl
bdf02ea67c - move masked reads 2017-04-11 01:11:30 +02:00
Christoph Oelckers
7a6e70e777 - reenabled the part of OP_THROW that throws a VMAbortException. This part is still needed for a runtime variable access check. 2017-04-11 00:11:36 +02:00
alexey.lysiuk
4c1e03ebff Fixed alignment issue with GCC/Clang i386 build 2017-04-10 22:20:34 +03:00
alexey.lysiuk
d5edd3db36 Commented out remains of VM exception handling opcodes
This fixes build with GCC/Clang
2017-04-10 22:18:43 +03:00
Christoph Oelckers
040c70714c - commented out the exception handling opcodes in the VM.
They are currently not used and not planned to be used - and they are rather costly by increasing stack size per call quite significantly.
2017-04-10 20:41:55 +02:00
Christoph Oelckers
9ae97502be - removed the last remnants of the ATAGs. 2017-04-10 17:08:52 +02:00
Christoph Oelckers
60dd58e7d2 - most ATAG stuff is gone, except for the static storage space in the VMFunction. 2017-04-10 16:06:18 +02:00
Christoph Oelckers
5464336035 - removed ATAGs from the function builder. 2017-04-10 15:48:27 +02:00
Christoph Oelckers
e551ef52f8 - removed atag parameter from GetConstantAddress. 2017-04-10 15:39:04 +02:00
Christoph Oelckers
ef77cbd295 - stop using the ATAGs for checking pointer types in asserts.
This is an incredibly costly way to do a debug check as it infests the entire VM design from top to bottom. These tags are basically useless for anything else but validating object pointers being passed to native functions (i.e. mismatches between definition and declaration) and that simply does not justify a feature that costs execution time in non-debug builds and added memory overhead everywhere.
Note that this commit does not remove the tags, it only discontinues their use.
2017-04-10 15:18:31 +02:00
alexey.lysiuk
175e784b67 Fixed applying of powerup effects after (un)morph
https://mantis.zdoom.org/view.php?id=556
2017-04-10 15:46:30 +03:00
alexey.lysiuk
287a0933bd Fixed lookup of inventory and ammo types in FraggleScript
https://forum.zdoom.org/viewtopic.php?t=55950
2017-04-10 11:58:53 +03:00
alexey.lysiuk
b0a0c62af1 Fixed state selection for weapon secondary fire
https://mantis.zdoom.org/view.php?id=552
2017-04-09 13:04:32 +03:00
Christoph Oelckers
5935dc706d - prevent infinite recursion in PlayerPawn.GetObituary when the inflictor is the same as the origin for the kill. 2017-04-09 00:06:23 +02:00
Christoph Oelckers
e1698c6d41 - split P_PlayerThink into several smaller functions.
This is for a future script export where overriding the actions separately may be interesting.
2017-04-08 15:45:11 +02:00
Christoph Oelckers
18d6fb7f49 - fixed: USEACTORPITCH in the model code was still using inverted pitch. 2017-04-08 10:24:12 +02:00
Magnus Norddahl
9f180b29b9 - Remove GL_MAJOR_VERSION check for OpenGL ES as apparently not even this can be done right by driver writers 2017-04-08 00:19:39 +02:00
Rachael Alexanderson
3a36ffee35 - vid_glswfb is now true by default in Linux 2017-04-07 09:53:59 -04:00
Magnus Norddahl
73effc70bb - The legacy OpenGL path resets the GLSL version, which in turn makes #version 0 crash the AMD driver. Old stuff that uses the legacy path will not get glswfb acceleration. 2017-04-07 15:45:08 +02:00
Magnus Norddahl
35c35fc26f - Check if glCreateProgram or glCreateShader fails, because apparently AMD found a way to make exactly that happen 2017-04-07 15:27:04 +02:00
Rachael Alexanderson
b1e1f25218 - detect gl es shaders and request them 2017-04-06 23:34:50 -04:00
Magnus Norddahl
85a9984807 - Add OpenGL ES 3 support to GL renderer 2017-04-06 23:34:42 -04:00
Magnus Norddahl
b7482e10e2 - Detect dinosaur OpenGL and refuse to use it 2017-04-06 23:34:37 -04:00
Magnus Norddahl
e690de25a8 - Use SDL_GL_GetProcAddress on platforms where the SDL is being used. Since SDL initialized OpenGL for us, it is the only reliable way of retrieving proc addresses.
- Check if ogl_LoadFunctions failed and make OpenGLSWFrameBuffer gracefully recover from that
2017-04-06 23:33:56 -04:00
Christoph Oelckers
74faacd218 - fixed: FxVMFunctionCall::GetDirectFunction did not perform any checks on the function's self pointer and failed to report a mismatch as an error.
- also fixed two places in the code where the above caused some incorrect definitions not to be detected.
2017-04-06 20:52:38 +02:00
alexey.lysiuk
ca4888eb3d Added CVar.GetCVar() function to do player-dependent console variables lookup 2017-04-06 10:51:36 +03:00
alexey.lysiuk
26e4b74261 Restored initial behavior of CVar.FindCVar()
https://mantis.zdoom.org/view.php?id=537
2017-04-06 10:50:40 +03:00
alexey.lysiuk
ac5e516148 Removed obsolete forward declaration 2017-04-06 10:39:53 +03:00
Magnus Norddahl
1312e9fb72 - fix linker error on platforms with no SSE support 2017-04-05 22:36:57 +02:00
alexey.lysiuk
a27ab73f33 Fixed broken dialogue customization in USDF parser
https://mantis.zdoom.org/view.php?id=539
2017-04-05 13:43:49 +03:00
alexey.lysiuk
dd30d2a045 DEHACKED lumps from IWAD now have precedence over embedded lumps and separate files
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
2017-04-04 12:20:12 +03:00