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.
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 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.
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.
The instruction one free instruction byte so it's now using that to extend its argument's register range to 65535.
For param this is needed because it passes strings by reference and creating an implicit temporary copy for string constants does not work here.
This setup has been a constant source of problems so now I reviewed all uses of FName to make sure that everything that needs to be initialized is done manually.
This also merges the player_t constructor into the class definition as default values.
Due to how function calls get resolved by the code generator there were some incorrect flags on the self pointer which triggered an assert and caused incorrect code generation.
Fortunately this was a mostly contained special case for which a workaround was possible.
Due to the special nature of this expression the code generator got stuck in 'address' mode and passed the address of the variable instead of its value.
Note that this is merely a hotfix. Properly handling this to allow universal use of state scopes will require more work in cases where a scoped state is being accessed through a non-self pointer.
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.
Let's use inline checkers in PType instead of constantly having to do clumsy IsKindOf checks etc. Once complete this also means that the types can be taken out of the class hierarchy, freeing up some common names.
Combining these two groups of data has been the cause of many hard to detect errors because it allowed liberal casting between types that are used for completely different things.
Having these two types related can cause problems with detection in the compiler because for some parts they need quite different handling.
Common handling for the fields has been moved into PSymbolTable but overall redundancy was quite minor as both types share surprisingly little functionality.
- 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.
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.
- 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.
It used the expression's value type, but needs to use the variable's, which can be different when the assignment is synthesized from a builtin function.
- keep string registers which are being used as function parameters allocated until after the function call returns. This is for allowing to pass strings by reference which would avoid some costly constructor/destructor loops in the call instruction.
src/gl/scene/gl_clipper.h:150:23: warning: comparison of integers of different signs: 'int' and 'unsigned int' [-Wsign-compare]
src/gl/dynlights/gl_aabbtree.cpp:137:24: warning: using integer absolute value function 'abs' when argument is of floating point type [-Wabsolute-value]
src/gl/dynlights/gl_aabbtree.cpp:137:34: warning: using integer absolute value function 'abs' when argument is of floating point type [-Wabsolute-value]
src/gl/dynlights/gl_aabbtree.cpp:137:44: warning: using integer absolute value function 'abs' when argument is of floating point type [-Wabsolute-value]
src/gl/dynlights/gl_aabbtree.cpp:139:6: warning: using integer absolute value function 'abs' when argument is of floating point type [-Wabsolute-value]
src/gl/dynlights/gl_aabbtree.cpp:139:30: warning: using integer absolute value function 'abs' when argument is of floating point type [-Wabsolute-value]
src/gl/dynlights/gl_aabbtree.cpp:139:54: warning: using integer absolute value function 'abs' when argument is of floating point type [-Wabsolute-value]
src/gl/dynlights/gl_aabbtree.cpp:142:6: warning: using integer absolute value function 'abs' when argument is of floating point type [-Wabsolute-value]
src/gl/dynlights/gl_aabbtree.cpp:143:3: warning: using integer absolute value function 'abs' when argument is of floating point type [-Wabsolute-value]
src/gl/dynlights/gl_aabbtree.cpp:144:3: warning: using integer absolute value function 'abs' when argument is of floating point type [-Wabsolute-value]
src/gl/dynlights/gl_aabbtree.cpp:167:6: warning: using integer absolute value function 'abs' when argument is of floating point type [-Wabsolute-value]
src/gl/dynlights/gl_shadowmap.cpp:163:31: warning: '&&' within '||' [-Wlogical-op-parentheses]
src/p_saveg.cpp:367:16: warning: comparison of integers of different signs: 'unsigned int' and 'int' [-Wsign-compare]
src/p_saveg.cpp:402:60: warning: comparison of integers of different signs: 'int' and 'unsigned int' [-Wsign-compare]
src/p_setup.cpp:1553:39: warning: format specifies type 'ptrdiff_t' (aka 'long') but the argument has type 'int' [-Wformat]
src/scripting/zscript/zcc_compile.cpp:293:74: warning: field 'AST' will be initialized after field 'mVersion' [-Wreorder]
src/swrenderer/drawers/r_thread.cpp:113:21: warning: comparison of integers of different signs: 'int' and 'size_t' (aka 'unsigned long') [-Wsign-compare]
- allow treatment as one-character string constants as character constants. This became necessary because name constants already use single quotes and are much harder to repurpose due to a higher degree of ambiguity.
- fixed: protected methods in structs were not usable.
The code for constants was not optimal because the first operand of the instructions cannot be constant. This was solved by swapping it with the second choice which will always be non-constant.
The code for local variables did not allocate a new destination register and would overwrite the first parameter's variable.
It now uses a dedicated opcode instead of piggybacking on OP_CALL and it passes data that is closer to the VM. Symbols should be avoided at this level.
It also will skip the scope instruction if the code generator detects that both calling function and the self pointer type have the same scope, this assumes that subclasses cannot flip between UI and Play.
They are not needed for OP_NEW_K which can evaluate the class relations at compile time and for OP_NEW the calling function can also be checked at compile time, passing only the scope value itself.
