abi.dd

Ddoc

$(SPEC_S Application Binary Interface,

	$(P A D implementation that conforms to the D ABI (Application Binary
	Interface)
	will be able to generate libraries, DLL's, etc., that can interoperate
	with
	D binaries built by other implementations.
	)

$(SECTION3 C ABI,

	$(P The C ABI referred to in this specification means the C Application
	Binary Interface of the target system.
	C and D code should be freely linkable together, in particular, D
	code shall have access to the entire C ABI runtime library.
	)
)

$(SECTION3 Endianness,

	$(P The $(LINK2 http://en.wikipedia.org/wiki/Endianness, endianness)
	(byte order) of the layout of the data
	will conform to the endianness of the target machine.
	The Intel x86 CPUs are $(I little endian) meaning that
	the value 0x0A0B0C0D is stored in memory as:
	$(D 0D 0C 0B 0A).
	)
)

$(SECTION3 Basic Types,

	$(TABLE
	$(TROW bool, 8 bit byte with the values 0 for false and 1 for true)
	$(TROW byte, 8 bit signed value)
	$(TROW ubyte, 8 bit unsigned value)
	$(TROW short, 16 bit signed value)
	$(TROW ushort, 16 bit unsigned value)
	$(TROW int, 32 bit signed value)
	$(TROW uint, 32 bit unsigned value)
	$(TROW long, 64 bit signed value)
	$(TROW ulong, 64 bit unsigned value)
	$(TROW cent, 128 bit signed value)
	$(TROW ucent, 128 bit unsigned value)
	$(TROW float, 32 bit IEEE 754 floating point value)
	$(TROW double, 64 bit IEEE 754 floating point value)
	$(TROW real, implementation defined floating point value$(COMMA) for x86 it is
	 80 bit IEEE 754 extended real)
	)
)

$(SECTION3 Delegates,

	$(P Delegates are $(I fat pointers) with two parts:)

	$(TABLE2 Delegate Layout,
	$(THEAD offset, property, contents)
	$(TROW $(D 0), $(D .ptr), context pointer)
	$(TROW $(I ptrsize), $(D .funcptr), pointer to function)
	)

	$(P The $(I context pointer) can be a class $(I this)
	reference, a struct $(I this) pointer, a pointer to
	a closure (nested functions) or a pointer to an enclosing
	function's stack frame (nested functions).
	)
)

$(SECTION3 Structs,

	$(P Conforms to the target's C ABI struct layout.)
)

$(SECTION3 Classes,

	$(P An object consists of:)

$(TABLE_3COLS Class Object Layout,
$(THEAD size, property, contents)
$(TROW $(I ptrsize), $(D .__vptr), pointer to vtable)
$(TROW $(I ptrsize), $(D .__monitor), monitor)
$(TROW $(D ...), $(D ...), $(ARGS super's non-static fields and super's interface vptrs, from least to most derived))
$(TROW $(D ...), named fields, non-static fields)
$(TROW $(I ptrsize)..., $(NBSP), $(ARGS vptr's for any interfaces implemented by this class in left to right, most to least derived, order))
)

	$(P The vtable consists of:)

$(TABLE2 Virtual Function Pointer Table Layout,
$(THEAD size, contents)
$(TROW $(I ptrsize), pointer to instance of TypeInfo)
$(TROW $(I ptrsize)..., pointers to virtual member functions)
)

	$(P Casting a class object to an interface consists of adding the offset of
	the interface's corresponding vptr to the address of the base of the object.
	Casting an interface ptr back to the class type it came from involves getting
	the correct offset to subtract from it from the object.Interface entry at vtbl[0].
	Adjustor thunks are created and pointers to them stored in the method entries in the vtbl[]
	in order to set the this pointer to the start of the object instance corresponding
	to the implementing method.
	)

	$(P An adjustor thunk looks like:)

$(CCODE
  ADD EAX,offset
  JMP method
)

	$(P The leftmost side of the inheritance graph of the interfaces all share
	their vptrs, this is the single inheritance model.
	Every time the inheritance graph forks (for multiple inheritance) a new vptr is created
	and stored in the class' instance.
	Every time a virtual method is overridden, a new vtbl[] must be created with
	the updated method pointers in it.
	)

	$(P The class definition:)

---------
class XXXX
{
    ....
};
---------

	$(P Generates the following:)

	$(UL
	$(LI An instance of Class called ClassXXXX.)

	$(LI A type called StaticClassXXXX which defines all the static members.)

	$(LI An instance of StaticClassXXXX called StaticXXXX for the static members.)
	)
)

$(SECTION3 Interfaces,

	$(P An interface is a pointer to a pointer to a vtbl[].
	The vtbl[0] entry is a pointer to the corresponding
	instance of the object.Interface class.
	The rest of the $(D vtbl[1..$]) entries are pointers to the
	virtual functions implemented by that interface, in the
	order that they were declared.
	)

	$(P A COM interface differs from a regular interface in that
	there is no object.Interface entry in $(D vtbl[0]); the entries
	$(D vtbl[0..$]) are all the virtual function pointers, in the order
	that they were declared.
	This matches the COM object layout used by Windows.
	)
    	$(P A C++ interface differs from a regular interface in that
	it matches the layout of a C++ class using single inheritance
	on the target machine.
	)

)

$(SECTION3 Arrays,

	$(P A dynamic array consists of:)

$(TABLE2 Dynamic Array Layout,
$(THEAD offset, property, contents)
$(TROW $(D 0), $(D .length), array dimension)
$(TROW $(D size_t), $(D .ptr), pointer to array data)
)

	$(P A dynamic array is declared as:)

---------
type[] array;
---------

	$(P whereas a static array is declared as:)

---------
type[dimension] array;
---------

	$(P Thus, a static array always has the dimension statically available as part of the type, and
	so it is implemented like in C. Static array's and Dynamic arrays can be easily converted back
	and forth to each other.
	)
)

$(SECTION3 Associative Arrays,

	$(P Associative arrays consist of a pointer to an opaque, implementation
	defined type.

	The current implementation is contained in and defined by
	 $(DRUNTIMESRC rt/aaA.d).
	)
)

$(SECTION3 Reference Types,

	$(P D has reference types, but they are implicit. For example, classes are always
	referred to by reference; this means that class instances can never reside on the stack
	or be passed as function parameters.
	)
)


$(SECTION3 Name Mangling,

	$(P D accomplishes typesafe linking by $(I mangling) a D identifier
	to include scope and type information.
	)

$(GRAMMAR
$(GNAME MangledName):
    $(B _D) $(GLINK QualifiedName) $(GLINK Type)
    $(B _D) $(GLINK QualifiedName) $(B M) $(GLINK Type)

$(GNAME QualifiedName):
    $(GLINK SymbolName)
    $(GLINK SymbolName) $(I QualifiedName)

$(GNAME SymbolName):
    $(GLINK LName)
    $(GLINK TemplateInstanceName)
)

	$(P The $(B M) means that the symbol is a function that requires
	a $(D this) pointer.)

	$(P Template Instance Names have the types and values of its parameters
	encoded into it:
	)

$(GRAMMAR
$(GNAME TemplateInstanceName):
    $(GLINK Number) $(B __T) $(GLINK LName) $(GLINK TemplateArgs) $(B Z)

$(GNAME TemplateArgs):
    $(GLINK TemplateArg)
    $(GLINK TemplateArg) $(I TemplateArgs)

$(GNAME TemplateArg):
    $(GLINK TemplateArgX)
    $(B H) $(GLINK TemplateArgX)

$(GNAME TemplateArgX):
    $(B T) $(GLINK Type)
    $(B V) $(GLINK Type) $(GLINK Value)
    $(B S) $(GLINK LName)

$(GNAME Value):
    $(B n)
    $(GLINK Number)
    $(B i) $(GLINK Number)
    $(B N) $(GLINK Number)
    $(B e) $(GLINK HexFloat)
    $(B c) $(GLINK HexFloat) $(B c) $(GLINK HexFloat)
    $(GLINK CharWidth) $(GLINK Number) $(B _) $(GLINK HexDigits)
    $(B A) $(GLINK Number) $(GLINK Value)...
    $(B S) $(GLINK Number) $(GLINK Value)...

$(GNAME HexFloat):
    $(B NAN)
    $(B INF)
    $(B NINF)
    $(B N) $(GLINK HexDigits) $(B P) $(GLINK Exponent)
    $(GLINK HexDigits) $(B P) $(GLINK Exponent)

$(GNAME Exponent):
    $(B N) $(GLINK Number)
    $(GLINK Number)

$(GNAME HexDigits):
    $(GLINK HexDigit)
    $(GLINK HexDigit) $(GLINK HexDigits)

$(GNAME HexDigit):
    $(GLINK Digit)
    $(B A)
    $(B B)
    $(B C)
    $(B D)
    $(B E)
    $(B F)

$(GNAME CharWidth):
    $(B a)
    $(B w)
    $(B d)
)

$(DL
	$(DT $(B n))
	$(DD is for $(B null) arguments.)

	$(DT $(GLINK Number))
	$(DD is for positive numeric literals (including
	character literals).)

	$(DT $(B N) $(GLINK Number))
	$(DD is for negative numeric literals.)

	$(DT $(B e) $(GLINK HexFloat))
	$(DD is for real and imaginary floating point literals.)

	$(DT $(B c) $(GLINK HexFloat) $(B c) $(GLINK HexFloat))
	$(DD is for complex floating point literals.)

	$(DT $(GLINK CharWidth) $(GLINK Number) $(D _) $(GLINK HexDigits))
	$(DD $(GLINK CharWidth) is whether the characters
	are 1 byte ($(B a)), 2 bytes ($(B w)) or 4 bytes ($(B d)) in size.
	$(GLINK Number) is the number of characters in the string.
	The $(GLINK HexDigits) are the hex data for the string.
	)

	$(DT $(B A) $(GLINK Number) $(GLINK Value)...)
	$(DD An array or asssociative array literal.
	$(GLINK Number) is the length of the array.
	$(GLINK Value) is repeated $(GLINK Number) times for a normal array,
	and 2 * $(GLINK Number) times for an associative array.
	)

	$(DT $(B S) $(GLINK Number) $(GLINK Value)...)
	$(DD A struct literal. $(GLINK Value) is repeated $(GLINK Number) times.
	)

)

$(GRAMMAR
$(GNAME Name):
    $(GLINK Namestart)
    $(GLINK Namestart) $(GLINK Namechars)

$(GNAME Namestart):
    $(B _)
    $(I Alpha)

$(GNAME Namechar):
    $(GLINK Namestart)
    $(GLINK Digit)

$(GNAME Namechars):
    $(GLINK Namechar)
    $(GLINK Namechar) $(I Namechars)
)

	$(P A $(GLINK Name) is a standard D identifier.)

$(GRAMMAR
$(GNAME LName):
    $(GLINK Number) $(GLINK Name)

$(GNAME Number):
    $(GLINK Digit)
    $(GLINK Digit) $(I Number)

$(GNAME Digit):
    $(B 0)
    $(B 1)
    $(B 2)
    $(B 3)
    $(B 4)
    $(B 5)
    $(B 6)
    $(B 7)
    $(B 8)
    $(B 9)
)

	$(P An $(GLINK LName) is a name preceded by a $(GLINK Number) giving
	the number of characters in the $(GLINK Name).
	)
)

$(SECTION3 Type Mangling,

	$(P Types are mangled using a simple linear scheme:)

$(GRAMMAR
$(GNAME Type):
    $(GLINK Shared)
    $(GLINK Const)
    $(GLINK Immutable)
    $(GLINK Wild)
    $(GLINK TypeArray)
    $(GLINK TypeStaticArray)
    $(GLINK TypeAssocArray)
    $(GLINK TypePointer)
    $(GLINK TypeFunction)
    $(GLINK TypeIdent)
    $(GLINK TypeClass)
    $(GLINK TypeStruct)
    $(GLINK TypeEnum)
    $(GLINK TypeTypedef)
    $(GLINK TypeDelegate)
    $(GLINK TypeVoid)
    $(GLINK TypeByte)
    $(GLINK TypeUbyte)
    $(GLINK TypeShort)
    $(GLINK TypeUshort)
    $(GLINK TypeInt)
    $(GLINK TypeUint)
    $(GLINK TypeLong)
    $(GLINK TypeUlong)
    $(GLINK TypeFloat)
    $(GLINK TypeDouble)
    $(GLINK TypeReal)
    $(GLINK TypeIfloat)
    $(GLINK TypeIdouble)
    $(GLINK TypeIreal)
    $(GLINK TypeCfloat)
    $(GLINK TypeCdouble)
    $(GLINK TypeCreal)
    $(GLINK TypeBool)
    $(GLINK TypeChar)
    $(GLINK TypeWchar)
    $(GLINK TypeDchar)
    $(GLINK TypeNull)
    $(GLINK TypeTuple)
    $(GLINK TypeVector)
    $(GLINK Internal)

$(GNAME Shared):
    $(B O) $(GLINK Type)

$(GNAME Const):
    $(B x) $(GLINK Type)

$(GNAME Immutable):
    $(B y) $(GLINK Type)

$(GNAME Wild):
    $(B Ng) $(GLINK Type)

$(GNAME TypeArray):
    $(B A) $(GLINK Type)

$(GNAME TypeStaticArray):
    $(B G) $(GLINK Number) $(GLINK Type)

$(GNAME TypeAssocArray):
    $(B H) $(GLINK Type) $(GLINK Type)

$(GNAME TypePointer):
    $(B P) $(GLINK Type)

$(GNAME TypeVector):
    $(B Nh) $(GLINK Type)

$(GNAME TypeFunction):
    $(GLINK CallConvention) $(GLINK FuncAttrs) $(GLINK Parameters) $(GLINK ParamClose) $(GLINK Type)

$(GNAME CallConvention):
    $(B F)       $(GREEN // D)
    $(B U)       $(GREEN // C)
    $(B W)       $(GREEN // Windows)
    $(B V)       $(GREEN // Pascal)
    $(B R)       $(GREEN // C++)

$(GNAME FuncAttrs):
    $(GLINK FuncAttr)
    $(GLINK FuncAttr) $(I FuncAttrs)

$(GNAME FuncAttr):
    $(I empty)
    $(GLINK FuncAttrPure)
    $(GLINK FuncAttrNothrow)
    $(GLINK FuncAttrProperty)
    $(GLINK FuncAttrRef)
    $(GLINK FuncAttrTrusted)
    $(GLINK FuncAttrSafe)
    $(GLINK FuncAttrNogc)

$(GNAME FuncAttrPure):
    $(B Na)

$(GNAME FuncAttrNothrow):
    $(B Nb)

$(GNAME FuncAttrRef):
    $(B Nc)

$(GNAME FuncAttrProperty):
    $(B Nd)

$(GNAME FuncAttrTrusted):
    $(B Ne)

$(GNAME FuncAttrSafe):
    $(B Nf)

$(GNAME FuncAttrNogc):
    $(B Ni)

$(GNAME Parameters):
    $(GLINK Parameter)
    $(GLINK Parameter) $(I Parameters)

$(GNAME Parameter):
    $(GLINK Parameter2)
    $(B M) $(GLINK Parameter2)     $(GREEN // scope)

$(GNAME Parameter2):
    $(GLINK Type)
    $(B J) $(GLINK Type)     $(GREEN // out)
    $(B K) $(GLINK Type)     $(GREEN // ref)
    $(B L) $(GLINK Type)     $(GREEN // lazy)

$(GNAME ParamClose)
    $(B X)     $(GREEN // variadic T t...$(RPAREN) style)
    $(B Y)     $(GREEN // variadic T t,...$(RPAREN) style)
    $(B Z)     $(GREEN // not variadic)

$(GNAME TypeIdent):
    $(B I) $(GLINK QualifiedName)

$(GNAME TypeClass):
    $(B C) $(GLINK QualifiedName)

$(GNAME TypeStruct):
    $(B S) $(GLINK QualifiedName)

$(GNAME TypeEnum):
    $(B E) $(GLINK QualifiedName)

$(GNAME TypeTypedef):
    $(B T) $(GLINK QualifiedName)

$(GNAME TypeDelegate):
    $(B D) $(GLINK TypeFunction)

$(GNAME TypeVoid):
    $(B v)

$(GNAME TypeByte):
    $(B g)

$(GNAME TypeUbyte):
    $(B h)

$(GNAME TypeShort):
    $(B s)

$(GNAME TypeUshort):
    $(B t)

$(GNAME TypeInt):
    $(B i)

$(GNAME TypeUint):
    $(B k)

$(GNAME TypeLong):
    $(B l)

$(GNAME TypeUlong):
    $(B m)

$(GNAME TypeFloat):
    $(B f)

$(GNAME TypeDouble):
    $(B d)

$(GNAME TypeReal):
    $(B e)

$(GNAME TypeIfloat):
    $(B o)

$(GNAME TypeIdouble):
    $(B p)

$(GNAME TypeIreal):
    $(B j)

$(GNAME TypeCfloat):
    $(B q)

$(GNAME TypeCdouble):
    $(B r)

$(GNAME TypeCreal):
    $(B c)

$(GNAME TypeBool):
    $(B b)

$(GNAME TypeChar):
    $(B a)

$(GNAME TypeWchar):
    $(B u)

$(GNAME TypeDchar):
    $(B w)

$(GNAME TypeNull):
    $(B n)

$(GNAME TypeTuple):
    $(B B) $(GLINK Number) $(GLINK Parameters)

$(GNAME Internal):
    $(B Z)
)
)

$(SECTION3 Function Calling Conventions,

	$(P The $(D extern (C)) and $(D extern (D)) calling convention matches the C
	calling convention
	used by the supported C compiler on the host system.
	Except that the extern (D) calling convention for Windows x86 is described here.
	)

$(SECTION4 Register Conventions,

	$(UL

	$(LI EAX, ECX, EDX are scratch registers and can be destroyed
	by a function.)

	$(LI EBX, ESI, EDI, EBP must be preserved across function calls.)

	$(LI EFLAGS is assumed destroyed across function calls, except
	for the direction flag which must be forward.)

	$(LI The FPU stack must be empty when calling a function.)

	$(LI The FPU control word must be preserved across function calls.)

	$(LI Floating point return values are returned on the FPU stack.
	These must be cleaned off by the caller, even if they are not used.)

	)
)

$(SECTION4 Return Value,

	$(UL

	$(LI The types bool, byte, ubyte, short, ushort, int, uint,
	pointer, Object, and interfaces
	are returned in EAX.)

	$(LI long and ulong
	are returned in EDX,EAX, where EDX gets the most significant
	half.)

	$(LI float, double, real, ifloat, idouble, ireal are returned
	in ST0.)

	$(LI cfloat, cdouble, creal are returned in ST1,ST0 where ST1
	is the real part and ST0 is the imaginary part.)

	$(LI Dynamic arrays are returned with the pointer in EDX
	and the length in EAX.)

	$(LI Associative arrays are returned in EAX.)

	$(LI References are returned as pointers in EAX.)

	$(LI Delegates are returned with the pointer to the function
	in EDX and the context pointer in EAX.)

	$(LI 1, 2 and 4 byte structs and static arrays are returned in EAX.)

	$(LI 8 byte structs and static arrays are returned in EDX,EAX, where
	EDX gets the most significant half.)

	$(LI For other sized structs and static arrays,
	the return value is stored through a hidden pointer passed as
	an argument to the function.)

	$(LI Constructors return the this pointer in EAX.)

	)
)

$(SECTION4 Parameters,

	$(P The parameters to the non-variadic function:)

---
foo(a1, a2, ..., an);
---

	$(P are passed as follows:)

	$(TABLE
	$(TR $(TD a1))
	$(TR $(TD a2))
	$(TR $(TD ...))
	$(TR $(TD an))
	$(TR $(TD hidden))
	$(TR $(TD this))
	)

	$(P where $(I hidden) is present if needed to return a struct
	value, and $(I this) is present if needed as the this pointer
	for a member function or the context pointer for a nested
	function.)

	$(P The last parameter is passed in EAX rather than being pushed
	on the stack if the following conditions are met:)

	$(UL
	$(LI It fits in EAX.)
	$(LI It is not a 3 byte struct.)
	$(LI It is not a floating point type.)
	)

	$(P Parameters are always pushed as multiples of 4 bytes,
	rounding upwards,
	so the stack is always aligned on 4 byte boundaries.
	They are pushed most significant first.
	$(B out) and $(B ref) are passed as pointers.
	Static arrays are passed as pointers to their first element.
	On Windows, a real is pushed as a 10 byte quantity,
	a creal is pushed as a 20 byte quantity.
	On Linux, a real is pushed as a 12 byte quantity,
	a creal is pushed as two 12 byte quantities.
	The extra two bytes of pad occupy the $(SINGLEQUOTE most significant) position.
	)

	$(P The callee cleans the stack.)

	$(P The parameters to the variadic function:)

---
void foo(int p1, int p2, int[] p3...)
foo(a1, a2, ..., an);
---

	$(P are passed as follows:)

	$(TABLE
	$(TR $(TD p1))
	$(TR $(TD p2))
	$(TR $(TD a3))
	$(TR $(TD hidden))
	$(TR $(TD this))
	)

	$(P The variadic part is converted to a dynamic array and the
	rest is the same as for non-variadic functions.)

	$(P The parameters to the variadic function:)

---
void foo(int p1, int p2, ...)
foo(a1, a2, a3, ..., an);
---

	$(P are passed as follows:)

	$(TABLE
	$(TR $(TD an))
	$(TR $(TD ...))
	$(TR $(TD a3))
	$(TR $(TD a2))
	$(TR $(TD a1))
	$(TR $(TD $(D _)arguments))
	$(TR $(TD hidden))
	$(TR $(TD this))
	)

	$(P The caller is expected to clean the stack.
	$(D _argptr) is not
	passed, it is computed by the callee.)
)

)

$(SECTION3 Exception Handling,

    $(SECTION4 Windows,

	$(P Conforms to the Microsoft Windows Structured Exception Handling
	conventions.
	)
    )

    $(SECTION4 Linux$(COMMA) FreeBSD and OS X,

	$(P Uses static address range/handler tables.
	It is not compatible with the ELF/Mach-O exception handling tables.
	The stack is walked assuming it uses the EBP/RBP stack frame
	convention. The EBP/RBP convention must be used for every
	function that has an associated EH (Exception Handler) table.
	)

	$(P For each function that has exception handlers,
	an EH table entry is generated.
	)

	$(TABLE2 EH Table Entry,
	$(THEAD field, description)
	$(TROW $(D void*), pointer to start of function)
	$(TROW $(D DHandlerTable*), pointer to corresponding EH data)
	$(TROW $(D uint), size in bytes of the function)
	)

	$(P The EH table entries are placed into the following special
	segments, which are concatenated by the linker.
	)

	$(TABLE2 EH Table Segment,
	$(THEAD Operating System, Segment Name)
	$(TROW Windows, $(D FI))
	$(TROW Linux, $(D .deh_eh))
	$(TROW FreeBSD, $(D .deh_eh))
	$(TROW OS X, $(ARGS $(D __deh_eh), $(D __DATA)))
	)
	$(BR)

	$(P The rest of the EH data can be placed anywhere,
	it is immutable.)

	$(TABLE2 DHandlerTable,
	$(THEAD field, description)
	$(TROW $(D void*), pointer to start of function)
	$(TROW $(D uint), offset of ESP/RSP from EBP/RBP)
	$(TROW $(D uint), offset from start of function to return code)
	$(TROW $(D uint), number of entries in $(D DHandlerInfo[]))
	$(TROW $(D DHandlerInfo[]), array of handler information)
	)
	$(BR)

	$(TABLE2 DHandlerInfo,
	$(THEAD field, description)
	$(TROW $(D uint), offset from function address to start of guarded section)
	$(TROW $(D uint), offset of end of guarded section)
	$(TROW $(D int), previous table index)
	$(TROW $(D uint), if != 0 offset to DCatchInfo data from start of table)
	$(TROW $(D void*), $(ARGS if not null, pointer to finally code to execute))
	)
	$(BR)

	$(TABLE2 DCatchInfo,
	$(THEAD field, description)
	$(TROW $(D uint), number of entries in $(D DCatchBlock[]))
	$(TROW $(D DCatchBlock[]), array of catch information)
	)
	$(BR)

	$(TABLE2 DCatchBlock,
	$(THEAD field, description)
	$(TROW $(D ClassInfo), catch type)
	$(TROW $(D uint), offset from EBP/RBP to catch variable)
	$(TR $(D void*), catch handler code)
	)
    )
)

$(SECTION3 Garbage Collection,

	$(P The interface to this is found in $(D phobos/internal/gc).)
)

$(SECTION3 Runtime Helper Functions,

	$(P These are found in $(D phobos/internal).)
)

$(SECTION3 Module Initialization and Termination,

	$(P All the static constructors for a module are aggregated into a
	single function, and a pointer to that function is inserted
	into the ctor member of the ModuleInfo instance for that
	module.
	)

	$(P All the static denstructors for a module are aggregated into a
	single function, and a pointer to that function is inserted
	into the dtor member of the ModuleInfo instance for that
	module.
	)
)

$(SECTION3 Unit Testing,

	$(P All the unit tests for a module are aggregated into a
	single function, and a pointer to that function is inserted
	into the unitTest member of the ModuleInfo instance for that
	module.
	)
)

$(SECTION2 Symbolic Debugging,

	$(P D has types that are not represented in existing C or C++ debuggers.
	These are dynamic arrays, associative arrays, and delegates.
	Representing these types as structs causes problems because function
	calling conventions for structs are often different than that for
	these types, which causes C/C++ debuggers to misrepresent things.
	For these debuggers, they are represented as a C type which
	does match the calling conventions for the type.
	The $(B dmd) compiler will generate only C symbolic type info with the
	$(B -gc) compiler switch.
	)

	$(TABLE2 Types for C Debuggers,
	$(THEAD D type, C representation)
	$(TROW dynamic array, $(D unsigned long long))
	$(TROW associative array, $(D void*))
	$(TROW delegate, $(D long long))
	$(TROW $(D dchar), $(D unsigned long))
	)

	$(P For debuggers that can be modified to accept new types, the
	following extensions help them fully support the types.
	)

$(SECTION3 $(LNAME2 codeview, Codeview Debugger Extensions),

	$(P The D $(B dchar) type is represented by the special
	primitive type 0x78.)

	$(P D makes use of the Codeview OEM generic type record
	indicated by $(D LF_OEM) (0x0015). The format is:)

	$(TABLE2 Codeview OEM Extensions for D,
	$(TROW field size, 2, 2, 2, 2, 2, 2)
	$(THEAD D Type, Leaf Index, OEM Identifier, recOEM, num indices,
	type index, type index)
	$(TROW dynamic array, $(D LF_OEM), $(I OEM), 1, 2, @$(I index), @$(I element))
	$(TROW associative array, $(D LF_OEM), $(I OEM), 2, 2, @$(I key), @$(I element))
	$(TROW delegate, $(D LF_OEM), $(I OEM), 3, 2, @$(I this), @$(I function)))

	$(P Where:)

	$(TABLE_2COLS ,
	$(TROW $(ARGS $(I OEM)), $(ARGS 0x42))
	$(TROW $(ARGS $(I index)), $(ARGS type index of array index))
	$(TROW $(ARGS $(I key)), $(ARGS type index of key))
	$(TROW $(ARGS $(I element)), $(ARGS type index of array element))
	$(TROW $(I this), $(ARGS type index of context pointer))
	$(TROW $(I function), $(ARGS type index of function)))

	$(P These extensions can be pretty-printed
	by $(LINK2 http://www.digitalmars.com/ctg/obj2asm.html, obj2asm).

	$(P The $(LINK2 http://ddbg.mainia.de/releases.html, Ddbg) debugger
	supports them.)
	)
)

$(SECTION3 $(LNAME2 dwarf, Dwarf Debugger Extensions),
	$(P The following leaf types are added:)

	$(TABLE2 Dwarf Extensions for D,
	$(THEAD D type, ID, Value, Format)
	$(TROW $(ARGS dynamic array)
	, $(ARGS $(D DW_TAG_darray_type))
	, $(ARGS 0x41)
	, $(ARGS $(D DW_AT_type) is element type))
	$(TROW
	$(ARGS associative array)
	, $(ARGS $(D DW_TAG_aarray_type))
	, $(ARGS 0x42)
	, $(ARGS $(D DW_AT_type) is element type, $(D DW_AT_containing_type) key type))
	$(TROW $(ARGS delegate)
	, $(ARGS $(D DW_TAG_delegate_type))
	, $(ARGS 0x43)
	, $(ARGS $(D DW_AT_type) is function type, $(D DW_AT_containing_type) is $(SINGLEQUOTE this) type)))

	$(P These extensions can be pretty-printed
	by $(LINK2 http://www.digitalmars.com/ctg/dumpobj.html, dumpobj).

	$(P The $(LINK2 https://zerobugs.codeplex.com, ZeroBUGS)
	debugger supports them.)
	)

	$(P Note that these Dwarf extensions have been removed as they conflict with
	recent gcc additions.
	)
)
)
)

Macros:
	TITLE=Application Binary Interface
	WIKI=ABI
	CATEGORY_SPEC=$0