From:	Nathan <ucgansj AT ucl DOT ac DOT uk>
Newsgroups:	comp.os.msdos.djgpp
Subject:	DJGPP: seekg 8/16bit problem!!!
Date:	Wed, 12 Jan 2000 09:45:35 +0000
Organization:	University College London
Lines:	40
Message-ID:	<387C4D3F.9624B646@ucl.ac.uk>
NNTP-Posting-Host:	psaurus.anat.ucl.ac.uk
Mime-Version:	1.0
X-Trace:	uns-a.ucl.ac.uk 947669940 8552 128.40.53.42 (12 Jan 2000 09:39:00 GMT)
X-Complaints-To:	usenet AT ucl DOT ac DOT uk
NNTP-Posting-Date:	12 Jan 2000 09:39:00 GMT
X-Mailer:	Mozilla 4.7 [en] (Win98; I)
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

Hello fellow DJGPP users,
	I've spent the last two weeks banging my head against a brick wall
trying write a very simply piece of C++ code, my first ever, to strip
the headers from medical image files (DICOM) and bunch them together.
The code is attached. 

Now, it appears to work perfectly with 8bit images, but screws up the
grey distribution given a 16bit image!!!... looks like isobars of 5  or
6 greys. What looked a very simply problem, i.e. skip and then copy,
is now becoming a nightmare. 

By a process of elimination I've located the problem to seekg, which is
used to skip the header length given matrix size etc. I have attached
the code if anybody wants to take a look. My question is whether this
problem is caused by the compiler DJGPP or some other limitation of
seekg under DJGPP?.

I would be grateful for any advice or fixes to my code,

many thanks

nathan
	
-- 


~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Nathan Jeffery PhD
Evolutionary Anatomy Unit
Dept. of Anatomy & Developmental Biology
Rockefeller Building, University Street
University College London
London WC1E 6JJ
UK

tel: 0171-209-6215
fax: 0171-209-0346
email: ucgansj AT ucl DOT ac DOT uk
http://evolution.anat.ucl.ac.uk/people/nathan/
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

- Raw text -