Re: PET Kernal 'load' routine & CC65 question

From: vtgearhead <snhirsch_at_gmail.com>
Date: Wed, 14 Oct 2020 07:52:43 -0500 (CDT)
Message-ID: <1602679963436-0.post_at_n4.nabble.com>
André Fachat wrote
> Am 14. Oktober 2020 02:11:40 schrieb vtgearhead &lt;

> snhirsch_at_

> &gt;:
> 
>> I'm working on adding overlay support to a CC65 application (CBM
>> Commander).
>> After a lot of hair pulling, I have the configuration setup and can build
>> main program and overlay code.  However, there is some confusion about
>> use
>> of 'cbm_load()' on a PET.  The CC65 sources fail to import the Kernal
>> LOAD
>> entry point ($FFD5) and the way it's being excluded makes me wonder if
>> this
>> is deliberate: All models but PET have it available.
>>
>> Since the entry point is there,
> 
> No it isn't. At least not what you'd expect.
> 
> Follow the source, you'll see that it calls BASIC to actually parse the
> parameters.
> 
> Full separation of BASIC and kernal has only been achieved in the VIC20
> and
> later.

Ah - thanks for the clarification, Andre'!  This had been driving me nuts. 
I'll implement Plan B and implement the load by leveraging cbm_read().  To
the cc65 maintainers: Wouldn't hurt to document this explicitly.  Code that
attempts to use cbm_load actually compiles without error - only to fail with
a linker error that might be cryptic to some folks.




--
Sent from: http://cbm-hackers.2304266.n4.nabble.com/
Received on 2020-10-14 15:00:04

Archive generated by hypermail 2.3.0.