On Fri, Apr 21, 2017 at 12:38 PM, Lee Olivares <l...@braains.net> wrote:

> Definitely saw that link and once I got over the hump of actually
> transferring the CO file using TEENY or SXM.BA it got me further than I
> was for sure.
>
> So I guess I understand how loadm works, once the CO file exists in file
> storage RAM, loadm copies into work RAM and spits out the load points.
>
> Once it's copied to work RAM you have to reserve that memory by CLEARing
> to the start (Top) address to actually execute it.
>
>
Actually, first CLEAR and then LOADM. LOADM'ing first is technically an
error. But, the nice thing about that error is it gives you a clue what you
need to clear.


> Cool, got it, now how do you "unload" something? Just load on top and do
> an appropriate clear?
>
>
Generally you don't need to do anything. If you need the memory back from
the CLEAR for something else (for other files, or a different CO program),
you do a new CLEAR command.

If you usually keep, say TEENY loaded, you can go one step further: CLEAR
and LOADM it once, then delete the original file and replace it with a tiny
trigger file instead. The reason for that is to not use the memory twice
(once for the RAM file, and once for the loaded instance).


> Final clarification, I notice some apps in warm.ba have clear256,XXXXX,
> and others have clear0,XXXXX, and I suppose the runm lines without loadm
> are for ROM based apps like xtel?
>
>
The first parameter to CLEAR is to reserve string memory for BASIC
programs. It's unrelated to running CO files. There's one other kind of
memory allocation, file control buffers. There's a different command for
that MAXFILES.

If you don't reserve enough string space some BASIC programs could fail or
run really slow due to hitting time consuming garbage collection more
frequently.


> PS, Is it accurate that XTEL is only available in the Booster Pak?
>
>
Yes. It would be nice to extract it and make it compatible with REX. Maybe
include the flow control feature and unicode translation demonstrated by
HTERM.

-- John.

Reply via email to