NetNews Usenet Archive 1992 #30/NN_1992_30.iso/spool/comp/lang/c/18259

Text File | 1992-12-12 | 2.5 KB | 57 lines

Path: sparky!uunet!zaphod.mps.ohio-state.edu!swrinde!cs.utexas.edu!qt.cs.utexas.edu!yale.edu!spool.mu.edu!agate!dog.ee.lbl.gov!horse.ee.lbl.gov!torek
From: [e-mail] (Chris Torek)
Newsgroups: comp.lang.c
Subject: Re: Calling convention: C vs PASCAL
Date: 12 Dec 1992 18:34:11 GMT
Organization: Lawrence Berkeley Laboratory, Berkeley CA
Lines: 45
Message-ID: <[e-mail]>
References: <[e-mail]> <[e-mail]>
NNTP-Posting-Host: 128.3.112.15

In article <[e-mail]> [e-mail]
(Steve Clamage) writes:
>Neither the C nor Pascal language specifications require the parameter
>ordering mentioned, but they are common implementation techniques.

(For that matter, these days it is common to pass the first two-to-six
parameters in registers, except on the 80x86 where the registers are
far too precious to be used this way.)

>C allows a function to be called with more parameters than were declared
>for it.

In ANSI C, this is only true of `, ...' functions, such as printf().
Moreover, these must be declared in advance as such.  This gives IBM PC
C compilers license to use the variant of `return' that adds to the
stack pointer, except when calling or returning from one of these
`special' functions.

In particular, this means that:

    /* incorrect version of `hello world' */
    int main(void) { printf("Hello world\n"); return 0; }

may well malfunction on an IBM PC.  To correct it, you *must* add at
least one of the following two lines:

    #include <stdio.h>
    int printf(char *, ...);

It is then guaranteed to work on any ANSI-conformant hosted system,
including all IBM PC C compilers.  Without the `advance notice' that
printf() is a `special' function, it need not work properly.

Of course, if one happened to go to your nearby mall and buy an IBM PC
C compiler, and it failed to compile and run the incorrect version, one
would be tempted to advise others against buying this particular
compiler---not necessarily believing that it is `wrong' or `buggy', but
perhaps only because one knows just how much old, nonconformant code
exists.  In other words, market forces pretty much guarantee that *your*
IBM PC C compiler by default uses the so-called `slow method' (which,
as Steve pointed out, need not actually be slower: there is no reason
that one *must* clear the stack after every function call).
--
In-Real-Life: Chris Torek, Lawrence Berkeley Lab CSE/EE ([phone])
Berkeley, CA        Domain:    [e-mail]