Re: [Sdcc-user] Possible changes to inline assembler syntax

[Maarten B. <sou...@ds...>]

Philipp Klaus Krause schreef op 2022-07-15 16:45:
>> 
>>> Since people like form 1, I'd now think we should keep it, but drop
>>> #pragma sdcc_hash. That looks like a better way forward and also
>>> solves the issues.
>> 
>> When this pragma is dropped does that mean it is no longer possible to 
>> use the # for an immediate operand?
> 
> You can still use a plain # (it already works now without using the 
> pragma):
> 
> void h(void)
> {
> __asm
> 	add	a, #3
> __endasm;
> }
> 
>> Or will it be limited to use outside of a macro? I wouldn't even know 
>> how to escape the #.
> 
> The C standard requires "Each # preprocessing token in the replacement
> list for a function-like macro shall be followed by a parameter as the
> next preprocessing token in the replacement list." (see section
> 6.10.3.2 of the current C2X standard N2912). Which makes sense, since
> in the standard, # is meant for stringizing macro arguments. The
> #pragma sdcc_hash is used to overide that.
> 
> lib/ds400/ds400.rom.c currently has the following lines:
> 
> // This macro is invalid for the standard C preprocessor, since it
> // includes a hash character in the expansion, hence the SDCC specific
> // pragma.
> #pragma sdcc_hash +
> #define ROMCALL(x) \
>         mov     R6_B3, #(x & 0xff)              \
>         mov     R7_B3, #((x >> 8) & 0xff)       \
>         lcall   __romcall
> 
> And then uses this macro from within __asm / __endasm. This will have
> to be changed to e.g.:
> 
> #define HASH #
> #define ROMCALL(x) \
>         mov     R6_B3, HASH(x & 0xff)              \
>         mov     R7_B3, HASH((x >> 8) & 0xff)       \
>         lcall   __romcall
> 
> Philipp

Ok, so the proper way to 'escape' the # inside a macro definition is by 
using another macro.
That is fine by me. We should probably put this in the manual even 
though it is standard compliant.

Maarten