From: johne AT parallax DOT co DOT uk (John Eccleston)
Newsgroups: comp.os.msdos.djgpp
Subject: Re: Padding Question
Date: Wed, 11 Dec 96 09:07:28 GMT
Organization: Parallax Solutions Ltd
Lines: 57
Message-ID: <58ltk9$i1n@red.parallax.co.uk>
References: <32A810D9 DOT 366E9B4E AT m-net DOT arbornet DOT org>
NNTP-Posting-Host: parsnip.parallax.co.uk
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp

In article <32A810D9 DOT 366E9B4E AT m-net DOT arbornet DOT org>, Rohit Jalan <rohitj AT m-net DOT arbornet DOT org> wrote:
>Hi !!
>
>Is it possible to stop gcc from padding structures ?
>
>Thnx. 
>Rohit
><rohitj AT hotmail DOT com>

Hi All,

When using DEC C/C++ on Alpha work stations we could get
around padding problems by rearranging structures, for
example:

struct fred
{
  short id;
  char  type[3];
  int   count;
};

(The byte map would be ssppcccpiiii where s=short, p=pad, c=char, i=int)

would be rewritten as:

struct fred
{
  int   count;
  short id;
  char  type[3];
};

(The byte map would be iiiisscccppp where s=short, p=pad, c=char, i=int)

and then all the padding would move to the end of the structure
and to all intents and purposes be ignored.  Is this the case with
gcc/gxx?

Admittedly, if this does work in gcc/gxx then it is only of use if you
have control over the data formats being used, difficult if interfacing
to 3rd party software/hardware.

Thanks in advance

John

________________________________________________________________
Parallax Solutions Ltd.   Tel.: 01203 514522
Stonecourt,               Fax.: 01203 514401
Siskin Drive,             Web : http://www.parallax.co.uk/~johne
Coventry CV3 4FJ          Mail: johne AT parallax DOT co DOT uk
________________________________________________________________
       Kaizen - Japanese, Lit: Never ending improvement
________________________________________________________________