Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!uupsi!sunic!liuida!isy!lysator.liu.se!zap
From: zap@lysator.liu.se (Zap Andersson)
Newsgroups: comp.sys.amiga.programmer
Subject: Re: Is this valid C?
Message-ID: <635@lysator.liu.se>
Date: 16 May 91 05:35:48 GMT
References: <53200@nigel.ee.udel.edu> <38763@ditka.Chicago.COM>
Sender: news@isy.liu.se (Lord of the News)
Organization: Lysator Computer Club, Linkoping University, Sweden
Lines: 64

comeau@ditka.Chicago.COM (Greg Comeau) writes:

>In article <53200@nigel.ee.udel.edu> lawsonse@vttcf.cc.vt.edu (Shannon Lawson) writes:
>>> In article <616@lysator.liu.se> zap@lysator.liu.se (Zap Andersson) writes:
>>> >struct foop *p; struct foop { int a; char *b };
>>> > >main() { p->a = 39;  /* Will complier barf here? */ }
>>> > There will be a no barf since at the ->a, foop is no longer an incomplete
>>> > type.  It is not a problem.
>>I disagree.  p is a pointer to foop, and has not been allocated a foop to which
>>it may point.  Depending on the compiler, p points to some random location, so
>>you are happily sticking the value 39 into some nebulous area.  Although your
>>syntax is correct (and the compiler may not gag), you should see a problem
>>at runtime when you try to reference p.

>Please do not compare apples to oranges.

>The original posting dealt strictly wiht the accessibility  of p/a and
>hence it was answered in that context.

Yes, and I don't know how it got there since I wrote it - and I had TWO
pices of code in it. In ONE of them I explicitly wrote malloc(sizeof(foo))
just to SHOW people that that wasn't the point. I seem to have forgottn
that in the last example - sorry - but that was NOT the point at ALL, ok?

The point was: If you say:

struct kevlar_oysters *ipp; /* Wich is valid EVEN IF kevlar_oysters are un-
                               known as this time: ipp is a pointer to an
                               'unknown' structure */

and AFTER that you do:

struct kevlar_oysters { int foo; float bar; };

Now the structure TYPE itself is known - however, is a compiler
sufficiently intellegent to know that it's previously marked as 'pointer to
unknown structure' is suddenly a KNOWN structore? I dont know, I think it
is compiler dependant. So if I NOW do

ipp = calloc(39,sizeof(struct kevlar_oysters)); /* To keep you happy :-) */
if (ipp) /* To keep you EVEN HAPPIER B-] */
   ipp->bar = 39;

The POINT is, will the COMPILER barf here, period. I dont care if the
program kills memory - I won't RUN it, I will just use it as a nice
exercise for my compiler (up - and stretch - and down - and relax)

>Re the runtime issues (not the original compile time issues), what you 
>describe is most certainly true given the way the code snipet above is
>composed.

Agreement is utter and total. The code is never run - just compiled. Hep!

>- Greg
>-- 
>	 Comeau Computing, 91-34 120th Street, Richmond Hill, NY, 11418
>                          Producers of Comeau C++ 2.1
>          Here:attmail.com!csanta!comeau / BIX:comeau / CIS:72331,3421
>                     Voice:718-945-0009 / Fax:718-441-2310
--
* * * * * * * * * * * * * * * * *          (This rent for space)
* My signature is smaller than  * Be warned! The letter 'Z' is Copyright 1991
* yours!  - zap@lysator.liu.se  * by Zap Inc. So are the colors Red, Green and
* * * * * * * * * * * * * * * * * Greenish-yellow (Blue was taken by IBM)