[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Ciderpress for OS/X?

In article 
<a6e69e75-0399-4b14-aaa5-604fa9b8380b@g10g2000yqh.googlegroups.com>,
 datajerk <datajerk@gmail.com> wrote:

[...]
> Loader.system.  Thanks for the tip.

This is my favorite part:

        ; That's what it's all about !
        lda     #<MLI
        ldx     #>MLI
        sta     HIMEM
        stx     HIMEM + 1

> If you change:
> 
> time_t starttime = time(NULL);
> time_t endtime = time(NULL);
> col80();
> 
> to:
> 
> col80();
> time_t starttime = time(NULL);
> time_t endtime = time(NULL);
> 
> Then you'll see my problem.  Once col80 is called, the time card 
> cannot be read.  -1 is returned.  basic.system and 
> loader.system--same results.

I compared assembler listings, generated with the -l option:

<http://www.cc65.org/doc/cl65-2.html>

In my example, pusheax (exported from lpush.s) saves the initialized 
time before calling col80. In your code, decsp8 allocates the two stack 
variables, then col80 is called before the times are initialized. 
col80() probably needs to preserve some registers, but I don't know the 
runtime well enough to suggest how.

An expedient approach might be to check the machine ID and initialize 
the 80-column firmware in the loader, before the runtime gets started. 
You wouldn't need to update the Basic global vectors; and you can exit 
your code via $3D0, which is set up to invoke the ProDOS bye routine.

Alternatively, you could initialize the firmware in a Basic STARTUP 
program that BRUNs the compiled object. Of course, that's what the 
loader was designed to avoid. :-)

-- 
John B. Matthews
trashgod at gmail dot com
<http://sites.google.com/site/drjohnbmatthews>