afscian@violet.waterloo.edu (Anthony Scian) (06/19/89)
This question arose from comp.lang.c:
Is this legal (ANSI)?
int foo( int, int y ) { ... } (1)
I think this should be legal:
int foo( int, int y );
and combined with this:
int foo( int x, int ) {return(x+y);}
COULD be equivalent to:
int foo( int x, int y ) {return(x+y);}
The grammar allows this but it is not clear to me what (1) means
in isolation.
Can identifiers in prototypes be different than the identifiers
in the function definition?
What about?
int test( int, int z );
int test( int x, int y ) {return(x+y);}
Should the compiler keep the identifiers and override them when
the function is defined (declared)?
//// Anthony Scian afscian@violet.uwaterloo.ca afscian@violet.waterloo.edu ////
"I can't believe the news today, I can't close my eyes and make it go away" -U2dfp@cbnewsl.ATT.COM (david.f.prosser) (06/20/89)
In article <14587@watdragon.waterloo.edu> afscian@violet.waterloo.edu (Anthony Scian) writes: >This question arose from comp.lang.c: > >Is this legal (ANSI)? >int foo( int, int y ) { ... } (1) No, but it would have been nice if it were valid, in my opinion. > >I think this should be legal: >int foo( int, int y ); Yes, it is valid. > >and combined with this: >int foo( int x, int ) {return(x+y);} > >COULD be equivalent to: >int foo( int x, int y ) {return(x+y);} No. The scope of prototype parameters in a declaration ends at the close paren. > >The grammar allows this but it is not clear to me what (1) means >in isolation. Nothing, since it is invalid. An (obvious, at least to me) interpretation is to have an int parameter whose value is not accessible in the function. It would be a more precise (and linguistic) method of noting that a parameter which is needed for other reasons is unused in a function. One example is while a program is being developed: the stub functions could be implemented without drawing comments about unused parameters. Another is when a number of functions are called indirectly. All these functions must have identical interfaces for portability, but some of the functions may well not need the value of certain parameters. >Can identifiers in prototypes be different than the identifiers >in the function definition? Yes, since the scope of the names are completely disjoint. >What about? >int test( int, int z ); >int test( int x, int y ) {return(x+y);} Valid. > >Should the compiler keep the identifiers and override them when >the function is defined (declared)? No. I see no benefits gained by producing warning messages about differently named parameters between a declaration (probably in a header file) and the definition. In fact, the only use for naming parameters in a declaration is to provide a little documentation that might help to distinguish between the parameters: div_t div(int numerator, int denominator); > >//// Anthony Scian afscian@violet.uwaterloo.ca afscian@violet.waterloo.edu //// >"I can't believe the news today, I can't close my eyes and make it go away" -U2 Dave Prosser ...not an official X3J11 answer...