So I've been looking at using the Break Key intercept. As I understand it, you put #&4C (i.e. the JMP instruction value) at &0287, then the lo-addr of your routine at &0288 and the hi-addr at &0289. Then, after a *FX200,3 (disable Esc; clear on Break) and pressing the Break key, your custom intercept routine should be called.

Except I haven't seen it happen. The routine is not called, and a normal break reset occurs. The second page seems to be cleared to 0.

(BTW, my routine was a simple 'jump to self' infinite loop, just to see would I get an expected hang - but it didn't happen.)

.
What am I missing?

Seems to work OK here.

I did :

P%=&2000
[JMP &2000
?&287=&4C
?&288=0
?&289=&20
*FX200,3

Pressing break clears the screen and goes into an infinite loop as expected.

This was tested using BeebEm3 in Model B mode but I don't see why it shouldn't work on a real machine or in Master 128 mode.

Tell us more about how you tried it and what code you were using.

Jon.

Except of course, when I actually try it in Master 128 mode, it doesn't work.

I guess you don't need the *FX200, 3 as this appears to clear all memory in Master 128 mode, including resetting the break vector and clearing out the break intercept code.

So, just miss out the *FX200,3 and pressing break then runs the intercept code as expected.

Jon.

Jon just beat me to it, yes I tried a quick example that just printed out the letter A. Worked fine, code was loaded at &3000 and was

lda #65
jsr &ffee
rts

The rts needs to be changed if you don't want it to go back into the normal break routine. In this example it prints an A out before the Acorn Boot message.

Then just typed
?&287=&4c
?&288=0
?&289=&30

and pressed break. Ran it in BeebEm but effects should be the same on real Beeb.

I have a sneaking suspicion that you've forget to do those addresses as hex ! i.e. I reckon you've typed

?287=4C

Least that's my guess

Try this little program:


It makes the Break key generate an error, just like the Escape key, which can then be trapped by ON ERROR in a Basic program. Muhahaha, no breaking out of the program!

(unless you press Space)

If there are still problems you could try *FX 247,76 *FX 248,<lsb> *FX 249,<msb> as the memory locations may be in a different place on the Master? Just a thought.

--G

No, it's ok, they're in the same place on a Master too.

But *FX 200,3 is a bad thing used in conjunction with the Break intercept - on a Master, as Jon said, memory's cleared before the Break intercept is tested, so it has no effect at all. On OS 1.20, the Beeb will hang on reset, regardless of what code it's pointed to.

Rich-TW

can you explain please why I can't put code in the low-stack area on a BBC Micro and do the *fx200,3 thing with break intercept? why won't this run?

The soft reset code in the ROM skips the stack on such a reset
Thanks.

It works fine for me in OS 1.20:



It could be that using &100 is not safe as paged ROMs sometimes copy code there to execute errors (via BRK). Maybe DFS or something is doing that. But putting code at &150 seems fine (if a little dangerous!)

hmm, sounds like an idea for my protection system....

BTW, I originally tried this on B-em, and BeebEm using OS1.2 and using OS2.00 (B+), without success. I must recheck where I assembled the code when I get home - I was pretty sure I put it away from the pages that the OS clears with its FX200 loop routine.

.

.

Furthermore, my MAIN intention was to simply enhance the memory-clearing effect on 'Break' for a B Plus machine, since the standard O.S. routine only clears certain pages in main ram. The shadow screen ram and special sideways ram remains untouched by pressing the 'Break' key. Easy to remedy (if I could get the break key vector to work [and with the fx200 stuff], that is!)