Do you mean that the D language does not support generation of 16-bit code, or that the existing D compilers do not generate 16-bit code? The older versions of Microsoft Visual C++ (Visual C++ V1.52) supported 16-bit code generation with far and near pointers, meaning that at least with non-standard extensions such as the aforementioned near and far pointers, C++ could support writing 16-bit programs. Is it concievable a compiler could be written for different memory architectures than flat 32/64 bit architectures? Could it be written for, say, a 600Mhz microcontroller that had 16-bit unmapped, unprotected segmented architecture that supported 8 bit segment addresses with 16 bit offsets and 16MB of main memory? Let's say that this architecture had near (16-bit addresses) and far pointers (24-bit addresses)... Would a compiler for such an architecture be possible?

On Thu, 29 Apr 2004 09:29:56
-0700, Walter wrote:

> 
> "LightEater" <LightEater_member@pathlink.com> wrote in message news:c6qlmd$omq$1@digitaldaemon.com...
>> I'm trying to set the cursor position using inline asm, i came up with the following nifty function, the program compiles, however when i call the
> function
>> i get an "Error: Access Violation". Any ideas on what might be wrong?
>>
>> void GotoXY (ubyte x, ubyte y)
>> {
>> asm {
>> mov AH, 2; /* cursor position */
>> mov DH, y;
>> mov DL, x;
>> mov BH, 0; /* video page #0 */
>> int 0x10;
>> }
>> }
>> }
> Generating interrupts only works when running a 16 bit program. D only supports 32 bit programming. Generating an interrupt from a Win32 program will cause an access violation. To set the cursor position in a 32 bit Win32 program, check out the Windows console API.

"Brian Bober" <netdemonz@yahoo.com> wrote in message news:cit1nf$2q3$1@digitaldaemon.com...
> Do you mean that the D language does not support generation of 16-bit code, or that the existing D compilers do not generate 16-bit code?

Yes and yes.

> The
> older versions of Microsoft Visual C++ (Visual C++ V1.52) supported 16-bit
> code generation with far and near pointers, meaning that at least with
> non-standard extensions such as the aforementioned near and far pointers,
> C++ could support writing 16-bit programs.

Digital Mars C++ also supports writing 16 bit programs, and does so with the current version.

> Is it concievable a compiler
> could be written for different memory architectures than flat 32/64 bit
> architectures? Could it be written for, say, a 600Mhz microcontroller that
> had 16-bit unmapped, unprotected segmented architecture that supported 8
> bit segment addresses with 16 bit offsets and 16MB of main memory? Let's
> say that this architecture had near (16-bit addresses) and far pointers
> (24-bit addresses)... Would a compiler for such an architecture be
> possible?

Yes, it's just a lot of work to do. I don't see sufficient demand for those architectures to justify it. But if some company was to offer to fund its development, any reasonable platform could be supported.