Path: utzoo!attcan!uunet!ncrlnk!ncrcae!hubcap!gatech!rutgers!cmcl2!adm!smoke!gwyn
From: gwyn@smoke.BRL.MIL (Doug Gwyn )
Newsgroups: comp.lang.c
Subject: Re: address of structure == address of first member? (long)
Message-ID: <8954@smoke.BRL.MIL>
Date: 22 Nov 88 09:29:51 GMT
References: <2172@iscuva.ISCS.COM>
Reply-To: gwyn@brl.arpa (Doug Gwyn (VLD/VMB) <gwyn>)
Organization: Ballistic Research Lab (BRL), APG, MD.
Lines: 22

In article <2172@iscuva.ISCS.COM> carlp@iscuva.ISCS.COM (Carl Paukstis) writes:
>What other portability problems will I run into with this code?
1.  Use of prototypes constrains you to a modern compiler.
2.  You implicitly declare exit() as returning int, which is incorrect.
3.  sizeof results in a size_t but mybsearch() wants an int.  Although the
    type is automatically converted in the scope of the prototype, if you
    tried to use a really large struct the size might not fit.
4.  You really ought to #include <string.h> to declare strcmp().

>86	        code = strcmp (key, *(char **)((char *)table + (m * size)));
>Line 86 bothers me.  Is it permissible to cast a void* into a char* and do
>arithmetic on it like this?  Line 89 would have the same problem (if any).

Line 86 is okay, but you really don't need to cast to a (char**) then
dereference to get the (char*) key.  That's extra work that amounts to
a no-op.

>On machines (environments?) with non-byte-sized chars, is there a better
>way to do the necessary arithmetic?  Or what else have I missed?

That seems about as good as any, if you don't want to tailor individual
versions to specific systems.