Re: question about linker

On 13/12/2024 01:59, James Kuyper wrote:

On 12/1/24 06:34, David Brown wrote:

On 30/11/2024 00:44, Keith Thompson wrote:

...

David apparently has a different definition of "totally different types"
than you do.  Since the standard doesn't define that phrase, I suggest
not wasting time arguing about it.
>

>
"int", "void" and "double" are totally different types in my view.
"int", "pointer to int", "array of int", "function returning int" all
have a relation that means I would not describe them as /totally/
different types - though I would obviously still call them /different/
types.
>
The syntax of C allows one declaration statement to declare multiple
identifiers of types related in this way - it does not allow declaration
of types of /totally/ different types.

 There's a rule I sometimes find useful, when trying to choose a precise
definition for a poorly defined term: figure out what statements you'd
like to say using the term, then define it in such a way as to guarantee
that those statements are correct.
 In C, a declaration may contain an init-declarator-list, preceded by
declaration-specifiers and optionally by an attribute-specifer-sequence
(6.7p1). Each of the declarators in the list share the
declaration-specifiers and the attribute-specifier-sequence (6.7p7). Any
syntax that's part of a declarator applies to that declarator's identifier.
 Therefore, your statement suggests that two types should be considered
"totally different types" if they are incompatible in either the
declaration-specifiers or the attribute-specifier-sequence. With that
definition, 6.7p7 in the standard would guarantee the truth of your
statement above.
 Does that definition sound suitable?

That definition sounds correct, yes, but also completely useless.  It leads directly to a tautology - you can declare two things in the same declaration in C if you are allowed to declare them in the same declaration in C.  I was simply trying to give such "things" can't have /totally/ different types, but need types that are in a certain way related.  I didn't go into the details of how they had to be related, because it should have been obvious that those details are precisely what the C standard says in the section about declarations.
It turns out that my comments there did not really help much - most C programmers already understand this stuff, and the poster who says he finds it all confusing likes to view every explanation as more confusing.
So I think this is best dropped.