From: david.brown@hesbynett.no
On 29/11/2025 20:36, Waldek Hebisch wrote:
> David Brown wrote:
>> _BitInt's are not arrays, they are scalars - they are integer types.
>> There is no concept of a type "_BitInt" - they always have compile-time
>> fixed sizes, such as "_BitInt(12)". So the idea of passing around
>> generic _BitInt's makes no more sense than passing around any other kind
>> of generic integer types. (Of course you can have an array of _BitInt's
>> of any given size.)
>
> There are languages which pass generic types, but C is not one
> of them. So idea of passing around generic _BitInt's makes sense,
> but this is not included in C.
>
By "pass around generic types", do you mean that you can write a
function that handles different types in the same parameter spot, and
have the function do different things for different passed-in types? C
has at least three ways to handle such things :
1. Variadic functions and the macros. "printf" is a prime
example of a function that can take parameters of many different types.
2. Passing arbitrary data via "void *" pointers. People do it all the time.
3. _Generic macros. These are not functions, but can pass off control
to different type-specific functions depending on the type of the arguments.
You can use any of these methods with _BitInts, but they are all missing
something significant in comparison to the support you would (and I
expect will) get in C++ with a template-based std::bit_int, or in
higher level languages like Python (if it had an equivalent of a set of
_BitInt types).
I therefore expect to be able to do more interesting things with _BitInt
types in C++ than in C, and do so in much simpler ways than equivalents
in C.
My point was that to do "something with generic _BitInt's" you have to
have to deal with a range of types, not one type with a range of sizes
or bit-counts.
--- SoupGate-Win32 v1.05
* Origin: you cannot sedate... all the things you hate (1:229/2)
|
