movem.l d0-d7/a0-a6,-(SP) inside a subroutine ?

[demolition]

Originally Posted by Mrs Beanbag

Unless you have some specific, strict coding style that handles this sort of thing, i've thought before about what if you want a function to return an object whose size isn't known to the caller and might not be a constant... but i can't think of a neat way to do that anyway without using two stacks.

By having the last item on the stack indicate the size of the object before it?

[Mrs Beanbag]

Originally Posted by demolition

By having the last item on the stack indicate the size of the object before it?

you could do that, but then it would be difficult to access any particular item, you'd have to iterate over all the objects until you got to the right one... my idea was to have a "data stack" and a "pointer stack" and then it would be easy but, wa wa wa,,,

[Leffmann]

That's just regular allocation on the heap, and giving the caller the responsibility of cleaning up.

[Mrs Beanbag]

it's only regular allocation on the heap, if you are calling "Allocate" and "Free", but it is simple to just reserve a big lump of memory for a second stack, afterall the main stack itself is allocated on the heap by the OS when it runs your program.

[Leffmann]

Yes, but allocating on the heap has always been how you solve this, it's not a new problem.

Originally Posted by Mrs Beanbag

Unless you have some specific, strict coding style that handles this sort of thing, i've thought before about what if you want a function to return an object whose size isn't known to the caller and might not be a constant... but i can't think of a neat way to do that anyway without using two stacks.

[Mrs Beanbag]

heap allocation is much slower, though. if you know you are going to construct and destroy in strict LIFO order it could be useful to use a second stack for performance benefits, and don't need to involve the operating system.

[sodapop]

Thanks for your enlightments guys...
While doing the "add.l #60,SP // sub.l #60,SP" worked OK under Devpac, it returned an error code when launching the executable file under the shell or from a bootable floppy disc (from scratch). The code is well executed, but going back to CLI returned a "39718" error, so I will stick with the usual "movem.l d0-d7/a0-a6,-(SP)" OUTSIDE the context saving subroutine...

[Leffmann]

Just return 0 in D0 and you'll get rid of the error message.

[sodapop]

@Leffmann: you're right !