From:	"Charles Wood" <c DOT dot DOT r DOT dot DOT wood AT worldnet DOT att DOT net DOT null>
Newsgroups:	comp.os.msdos.djgpp
Subject:	Re: Porting Watcom 11.0 to djgpp
Date:	Fri, 7 Jan 2000 15:45:26 -0800
Organization:	AT&T WorldNet Services
Lines:	59
Message-ID:	<855qoe$au9$1@bgtnsc01.worldnet.att.net>
References:	<853fat$5q4$1 AT bgtnsc03 DOT worldnet DOT att DOT net> <3875D12C DOT F6931CFF AT is DOT elta DOT co DOT il> <8559hr$49a$1 AT bgtnsc02 DOT worldnet DOT att DOT net> <855ckh$ptl$1 AT nets3 DOT rz DOT RWTH-Aachen DOT DE>
NNTP-Posting-Host:	12.72.133.227
X-Trace:	bgtnsc01.worldnet.att.net 947285582 11209 12.72.133.227 (7 Jan 2000 22:53:02 GMT)
X-Complaints-To:	abuse AT worldnet DOT att DOT net
NNTP-Posting-Date:	7 Jan 2000 22:53:02 GMT
X-Newsreader:	Microsoft Outlook Express 4.72.2106.4
X-MimeOLE:	Produced By Microsoft MimeOLE V4.72.2106.4
To:	djgpp AT delorie DOT com
DJ-Gateway:	from newsgroup comp.os.msdos.djgpp
Reply-To:	djgpp AT delorie DOT com

I'm only porting to see if my current compiler really has a bug.  I thought
it shouldn't be a *BIG* deal, whatever like 20 hours for 50k lines of code.

I wrapped all compiler specific stuff nicely, so it's easy to change later.

But the lack of the interrupt keyword (which I thought was standard
(duh...)) really trashes tons of code.

The fact that it can't assemble the tons of asm code that I haven't touched
for a year is terrible.  Write it once, forget about it.   Not true here.

Not to mention the quantity of inline assembler I use in the hardware
sections.

This is a "hardware" project.   There is no way to isolate it from the
hardware.

The majority of the code is quite standard, but the minority is where the
numerous time is involved and detailed superdelicate hardware related code.

I've dealt with compiler bugs before by moving stuff around, doing it
differently, no big deal.   I've actually been programming for quite a long
time 20+ years.   5 years have been professionally, (ie paycheck).  I have a
good idea of when it's me and when it's a compiler problem.

Well, I hope Watcom has some success seeing the bug, it's only about 16
lines of code that have to be examined.

As for porting to DJGPP, I don't think it fits in my timeline.

--
Charles Wood
REMOVEMEc DOT r DOT wood AT worldnet DOT att DOT net


Hans-Bernhard Broeker wrote in message
<855ckh$ptl$1 AT nets3 DOT rz DOT RWTH-Aachen DOT DE>...
>Charles Wood <c dot r dot wood @ worldnet.att.net.null> wrote:
>> This is *NOT* fun, nor exciting, nor even moderately interesting to do.
>
>Well, for as much as we know, you're always free to not do it, then :-^)
>
>> Porting does encompass any other the things that make me Like to be a
>> programmer.
>
>Porting is work. Not a nice or funny hacking exercise, but tedious and
>unrewarding, *hard* work, most of the time.
>
>In the case at hand, the heart of the problem is that the code you're
>working on is, by its very nature, strictly unportable. It relies on
>compiler features like the (nonstandard) 'interrupt' keywords. If that
>interrupt stuff is needed, that's a clear sign that this code is
>*very* close to the hardware and O/S, where all portability becomes an
>illusion --- even portability from one compiler to the other.
>--
>Hans-Bernhard Broeker (broeker AT physik DOT rwth-aachen DOT de)
>Even if all the snow were burnt, ashes would remain.

- Raw text -