From:	"Campbell, Rolf [SKY:1U32:EXCH]" <cp1v45 AT americasm01 DOT nt DOT com>
Newsgroups:	comp.os.msdos.djgpp
Subject:	Re: From Bytes to Int and Char
Date:	Tue, 17 Aug 1999 11:16:22 -0400
Organization:	Nortel Networks
Lines:	31
Message-ID:	<37B97CC6.7E054E7F@americasm01.nt.com>
References:	<rfXs3.4$bZ1 DOT 1603 AT typhoon01 DOT swbell DOT net> <37B466D7 DOT 958F09E5 AT americasm01 DOT nt DOT com> <37B4DFD1 DOT 4D66 AT surfsouth DOT com> <37B57EB0 DOT 492A93AE AT unb DOT ca> <37ba82f5 DOT 7514501 AT news DOT clara DOT net> <37B82AAA DOT E7223AB1 AT americasm01 DOT nt DOT com> <37bd890c DOT 14677884 AT news DOT clara DOT net>
NNTP-Posting-Host:	bmerhc00.ca.nortel.com
Mime-Version:	1.0
X-Mailer:	Mozilla 4.7 [en] (X11; I; HP-UX B.10.20 9000/712)
X-Accept-Language:	en
To:	djgpp AT delorie DOT com
DJ-Gateway:	from newsgroup comp.os.msdos.djgpp
Reply-To:	djgpp AT delorie DOT com

George Ryot wrote:

> "Campbell, Rolf [SKY:1U32:EXCH]" <cp1v45 AT americasm01 DOT nt DOT com> wrote:
>
> >     I don't understand how your code is any less of a 'hack' than mine.  Both are not
> > endian safe.  Both will work consistantly on each individual machine (I'm not saying
> > that you could exchange data files and have them work).  How is my code 'unsafe'?
>
> Casting between pointers to different types is nearly always unsafe.

    It depends what you mean by 'unsafe'.  If you look at code, I think you would find
that most pointer casts work fine and that a small minority cause unsafe code.


> The code I posted will solve the original problem (as I understand it)
> and will work the same on *all* implementations except attention is
> required to unsigned values.

[your original code]:
 unsigned short sz, tz;
    sz = (file_buffer[5]  << 8) & file_buffer[6];
    tz = (file_buffer[10] << 8) & file_buffer[11];

Your code does solve the original problem.  And you know what... I totally missed the fact
that you were forcing big-endian-ness.  Oh well, I live: I learn...

--
     -Rolf Campbell (39)3-6318

- Raw text -