this short program crashes the spectrum

[sn3j]

[sn3j]

Spoiler

[sn3j]

Spoiler

[sn3j]

Spoiler

Spoiler

[PeteProdge]

10 NEW

Say what you like about this program, but it gives you a flawless Spectrum emulator on any physical ZX Spectrum.

[equinox]

Say what you like about this program, but it gives you a flawless Spectrum emulator on any physical ZX Spectrum.

I've been trying to do a bit of -- never mind -- dealing with someone else's machine code. And it made me thought about CLEAR. Like, suppose I only wanted to reserve exactly one byte, could i do like CLEAR 65534

JUST SAYING. Arguing with AIs. shut up leaev me alone.

[equinox]

Well, to be precise that command doesn't crash the Spectrum - it resets it. Big difference

This is the kind of pedantry I come here for.
(HAHAH! I ended a sentence with a preposition. fight me)

[SNG]

suppose I only wanted to reserve exactly one byte, could i do like CLEAR 65534

Yes.

The byte reserved will be at 65535.

Enjoy POKEing it.

[AndyC]

Yes.

The byte reserved will be at 65535.

Enjoy POKEing it.

Can you move RAMTOP that high? I thought BASIC stores it's UDGs etc above RAMTOP and so I'd assume it would complain about it (though I've never tried)

[1024MAK]

@AndyC well try it!

Spoiler

A Spectrum has no trouble setting RAMTOP to a higher than normal value by using CLEAR 65534. It doesn't care where the UDG data is, or even if it's just garbage. Just don't try to redefine the UDG if it's not in a protected RAM area.

Mark

[uglifruit]

(As found in my signature) I love the behaviour of a 48k spectrum when CLEAR 23855 is entered.

Wish I'd done it in Dixon's.

[arjun]

10 NEW

A crash is an unexpected behaviour during the normal course of a program. The BASIC command "NEW" however behaves exactly like it should. So I would argue that it's not a crash in the technical sense.

[ParadigmShifter]

You can also write a basic program that does a CLEAR, pokes some assembler above the clear address and then do a new then save out the data if you want I suppose (although probably better just to save the data and do a full reset lol).

NEW just clears all variables and up to RAMTOP or whatever the end of the BASIC reserved area is (and moves the stack to immediately below the CLEAR address I think).

[akeley]

(As found in my signature) I love the behaviour of a 48k spectrum when CLEAR 23855 is entered.

Wish I'd done it in Dixon's.

My old time favourite is RANDOMIZE USR 5050.

Found exactly how you'd imagine: by a bunch of us kids bashing the keyboard to see what happens.

[equinox]

Yes.

The byte reserved will be at 65535.

Enjoy POKEing it.

But couldn't that be part of like USR "U"? Are you telling me that I have to CLEAR space for my UDGs? mind blown

(My first Speccy was a +2 so I don't consider U a real UDG, but.)

Okay, I know the real answer (I think): no sane BASIC program could stretch into the 65000s so you should technically clear that address (so your BASIC can't get up there) but it's just theoretical. 64K should be enough for everybody. (And +2 owners can rejoice in having 16 free bytes up there -- that's mostly why they bought a +2, isn't it?)

[equinox]

A crash is an unexpected behaviour during the normal course of a program. The BASIC command "NEW" however behaves exactly like it should. So I would argue that it's not a crash in the technical sense.

Very smart. Hire this man for a 10% reduction in IT support tickets without actually changing what you do, except saying "ACTUALLY" in the voice that women around the world have come to hate.

[TMD2003]

Okay, I know the real answer (I think): no sane BASIC program could stretch into the 65000s so you should technically clear that address (so your BASIC can't get up there) but it's just theoretical. 64K should be enough for everybody. (And +2 owners can rejoice in having 16 free bytes up there -- that's mostly why they bought a +2, isn't it?)

I use those for high scores that won't be cleared by RUN, or a quick and dirty LDIR routine that only needs 12 bytes.