From:	Sugree Phatanapherom <ioisgp AT std DOT cpc DOT ku DOT ac DOT th>
Newsgroups:	comp.lang.asm.x86,alt.lang.asm,comp.os.msdos.djgpp
Subject:	NASM 0.93
Date:	Sun, 2 Feb 1997 14:17:24 -0700
Organization:	Kasetsart University
Lines:	51
Message-ID:	<Pine.GSO.3.95.970202141530.29827B-100000@std.cpc.ku.ac.th>
NNTP-Posting-Host:	ioisgp AT std DOT cpc DOT ku DOT ac DOT th
Mime-Version:	1.0
To:	djgpp AT delorie DOT com
DJ-Gateway:	from newsgroup comp.os.msdos.djgpp

On 17 Jan 1997, Simon Tatham wrote:

> Yes, it is. The pre-built DOS executables in the NASM 0.90 and 0.91
> archives were compiled without nearly enough stack, because I wasn't
> thinking very hard when I did it.
> 
> 0.92 should be appearing on Simtel mirrors in a few days' time, and
> won't exhibit this problem any more. If you need it before then, you
> can get 0.92 immediately from ftp://yoda.trin.cam.ac.uk/pub/simon.

I have never used 0.92 because I just get 0.93 already :) Old problem, 
however, still occured n 0.93 as 0.91. Expression had been computed fine
if there were 3 terms of numeric or variable.

	times 512-$ db 0
or
	times (512-$) db 0

It doesn't work correctly, compiler told non-constant was in parameter.

buffer:
	times 512-$+buffer db 0

Above command worked fine, no more error.
So I had to fake reference label to use for 'times' command.

[org 0]

start:

	<code>

	times 512-$+start db 0

Just declare 'start' label, in fact 'start' value to zero. zero should be
able to replace to 'start'. No, 3rd term must be LABEL only.

	times 512-$+0 db 0

If you replaced 'start' to '0', error message would occur on your monitor.

I hope all problem is fixed soon.

--
Sugree Phatanapherom                     Everything you do today may
ioisgp AT std DOT cpc DOT ku DOT ac DOT th                  take effect tomorrow.
http://203.155.125.134/~sugreep/
http://std.cpc.ku.ac.th/~ioisgp/
http://www.geocities.com/Tokyo/4744/

- Raw text -