Interfacing to C
D is designed to fit comfortably with a C compiler for the target system. D makes up for not having its own VM by relying on the target environment's C runtime library. It would be senseless to attempt to port to D or write D wrappers for the vast array of C APIs available. How much easier it is to just call them directly.
This is done by matching the C compiler's data types, layouts, and function call/return sequences.
Calling C Functions
C functions can be called directly from D. There is no need for wrapper functions, argument swizzling, and the C functions do not need to be put into a separate DLL.
The C function must be declared and given a calling convention, most likely the "C" calling convention, for example:
extern (C) int strcmp(const char* string1, const char* string2);and then it can be called within D code in the obvious way:
import std.string; int myDfunction(char[] s) { return strcmp(std.string.toStringz(s), "foo"); }
There are several things going on here:
- D understands how C function names are "mangled" and the correct C function call/return sequence.
- C functions cannot be overloaded with another C function with the same name.
- There are no __cdecl, __far, __stdcall, __declspec, or other such C extended type modifiers in D. These are handled by linkage attributes, such as extern (C).
- There is no volatile type modifier in D. To declare a C function that uses volatile, just drop the keyword from the declaration.
- Strings are not 0 terminated in D. See "Data Type Compatibility" for more information about this. However, string literals in D are 0 terminated.
C code can correspondingly call D functions, if the D functions use an attribute that is compatible with the C compiler, most likely the extern (C):
// myfunc() can be called from any C function extern (C) { void myfunc(int a, int b) { ... } }
Storage Allocation
C code explicitly manages memory with calls to malloc() and free(). D allocates memory using the D garbage collector, so no explicit free's are necessary.
D can still explicitly allocate memory using core.stdc.stdlib.malloc() and core.stdc.stdlib.free(), these are useful for connecting to C functions that expect malloc'd buffers, etc.
If pointers to D garbage collector allocated memory are passed to C functions, it's critical to ensure that that memory will not be collected by the garbage collector before the C function is done with it. This is accomplished by:
- Making a copy of the data using core.stdc.stdlib.malloc() and passing the copy instead.
- Leaving a pointer to it on the stack (as a parameter or automatic variable), as the garbage collector will scan the stack.
- Leaving a pointer to it in the static data segment, as the garbage collector will scan the static data segment.
- Registering the pointer with the garbage collector with the std.gc.addRoot() or std.gc.addRange() calls.
An interior pointer to the allocated memory block is sufficient to let the GC know the object is in use; i.e. it is not necessary to maintain a pointer to the beginning of the allocated memory.
The garbage collector does not scan the stacks of threads not created by the D Thread interface. Nor does it scan the data segments of other DLL's, etc.
Data Type Compatibility
D | C | |
---|---|---|
32 bit | 64 bit | |
void | void | |
byte | signed char | |
ubyte | unsigned char | |
char | char (chars are unsigned in D) | |
wchar | wchar_t (when sizeof(wchar_t) is 2) | |
dchar | wchar_t (when sizeof(wchar_t) is 4) | |
short | short | |
ushort | unsigned short | |
int | int | |
uint | unsigned | |
ulong | unsigned long long | unsigned long |
core.stdc.config.c_long | long | long |
core.stdc.config.c_ulong | unsigned long | unsigned long |
long | long long | long (or long long) |
ulong | unsigned long long | unsigned long (or unsigned long long) |
float | float | |
double | double | |
real | long double | |
cdouble | double _Complex | |
creal | long double _Complex | |
struct | struct | |
union | union | |
enum | enum | |
class | no equivalent | |
type * | type * | |
type[dim] | type[dim] | |
type[dim]* | type(*)[dim] | |
type[] | no equivalent | |
type1[type2] | no equivalent | |
type function(params) | type(*)(params) | |
type delegate(params) | no equivalent | |
size_t | size_t | |
ptrdiff_t | ptrdiff_t |
These equivalents hold for most C compilers. The C standard does not pin down the sizes of the types, so some care is needed.
Passing D Array Arguments to C Functions
In C, arrays are passed to functions as pointers even if the function prototype says its an array. In D, static arrays are passed by value, not by reference. Thus, the function prototype must be adjusted to match what C expects.
D type | C type |
---|---|
T* | T[] |
ref T[dim] | T[dim] |
For example:
void foo(int a[3]) { ... } // C code
extern (C) { void foo(ref int[3] a); // D prototype }
Calling printf()
This mostly means checking that the printf format specifier matches the corresponding D data type. Although printf is designed to handle 0 terminated strings, not D dynamic arrays of chars, it turns out that since D dynamic arrays are a length followed by a pointer to the data, the %.*s format works:
void foo(char[] string) { printf("my string is: %.*s\n", string.length, string.ptr); }
The printf format string literal in the example doesn't end with '\0'. This is because string literals, when they are not part of an initializer to a larger data structure, have a '\0' character helpfully stored after the end of them.
An improved D function for formatted output is std.stdio.writef().
Structs and Unions
D structs and unions are analogous to C's.
C code often adjusts the alignment and packing of struct members with a command line switch or with various implementation specific #pragma's. D supports explicit alignment attributes that correspond to the C compiler's rules. Check what alignment the C code is using, and explicitly set it for the D struct declaration.
D does not support bit fields. If needed, they can be emulated with shift and mask operations, or use the std.bitmanip.bitfields library type. htod will convert bit fields to inline functions that do the right shift and masks.
D does not support declaring variables of anonymous struct types. In such a case you can define a named struct in D and make it private:
union Info // C code { struct { char *name; } file; };
union Info // D code { private struct File { char* name; } File file; }
Callbacks
D can easily call C callbacks (function pointers), and C can call callbacks provided by D code if the callback is an extern(C) function, or some other linkage that both sides have agreed to (e.g. extern(Windows)).
Here's an example of C code providing a callback to D code:
void someFunc(void *arg) { printf("Called someFunc!\n"); } // C code typedef void (*Callback)(void *); extern "C" Callback getCallback(void) { return someFunc; }
extern(C) alias Callback = int function(int, int); // D code extern(C) Callback getCallback(); void main() { Callback cb = getCallback(); cb(); // invokes the callback }
And an example of D code providing a callback to C code:
extern "C" void printer(int (*callback)(int, int)) // C code { printf("calling callback with 2 and 4 returns: %d\n", callback(2, 4)); }
extern(C) alias Callback = int function(int, int); // D code extern(C) void printer(Callback callback); extern(C) int sum(int x, int y) { return x + y; } void main() { printer(&sum); }
For more info about callbacks read the closures section.
Using Existing C Libraries
Since D can call C code directly, it can also call any C library functions, giving D access to the smorgasbord of existing C libraries. To do so, however, one needs to write a D interface (.di) file, which is a translation of the C .h header file for the C library into D.
For popular C libraries, the first place to look for the corresponding D interface file is the Deimos Project. If it isn't there already, and you write one, please contribute it to the Deimos Project.
Accessing C Globals
C globals can be accessed directly from D. C globals have the C naming convention, and so must be in an extern (C) block. Use the extern storage class to indicate that the global is allocated in the C code, not the D code. C globals default to being in global, not thread local, storage. To reference global storage from D, use the __gshared storage class.
extern (C) extern __gshared int x;