Re: [cc65] Linker config for assembler users (was: ca65 for stand-alone asm projects)

From: <silverdr1wfmh.org.pl>
Date: 2010-11-13 20:56:34
On 2010-11-13, at 19:09, Ullrich von Bassewitz wrote:

> 
> On Wed, Nov 03, 2010 at 11:57:34AM +0100, Oliver Schmidt wrote:
>> a) There could be a new (per target) secondary linker config making
>> only minimal assumptions on the program.
> 
> Using the new linker features, I'm currently experimenting with such a linker
> config for assembler programmers (for the C64 as a first take).

!! :-)

> When doing so,
> a few questions came up:
> 
> 1. What is a reasonable default start address for a config targeted to
>   assembler programmers? $801? $C000? $4000?

It really depends, but I always used either $0801 (for the beginning - with BASIC stub) for standalone prgs or $C000 for things that were supposed to interact with whatever else. Other start addresses were rarely used.

> 2. How about the usable RAM area? Should it be limited to $A000 by default?
>   My current version uses $FFFF, which means that any errors will go
>   undetected.

For default it could be lower IMHO so that errors can be detected. If the prg grows big, it will probably warrant a customised approach anyway.

> 3. Is it desirable to have an optional BASIC SYS header, or is this used so
>   rarely that people are better served with a less complex linker config?

I believe it is used a lot. Except mentioned modules, almost everything I wrote (and saw people writing) was started with SYS from BASIC. But IMHO it should be (as you wrote) optional.

> 4. How about the SYS token, is it the same on all existing 65xx Commodore
>   machines? In other words: Is a BASIC SYS header module portable across
>   all CBM machines? I do think so, but I'm not 100% sure.

I recall checking it some (long) time ago, and to my best knowledge it was the same in all BASIC versions. We can check it pretty easily these days though (thanks VICE guyz! :-)

> 5. Segments: What I'm currently doing is to make all standard segments
>   optional. Apart from the missing error detection, is there anything that
>   speaks against this approach?

How it was? Things should be made as simple as possible. Just not more ;-)

-- 
SD!----------------------------------------------------------------------
To unsubscribe from the list send mail to majordomo@musoftware.de with
the string "unsubscribe cc65" in the body(!) of the mail.
Received on Sat Nov 13 20:56:42 2010

This archive was generated by hypermail 2.1.8 : 2010-11-13 20:56:44 CET