Memory leak

[adrian4096]

Something takes up a lot of memory (4.2GiB)!
I suspect it has something with JSUB & JRET, but am not sure.

Why are JSUB & RSUB tracked in java anyway?

[jetomit]

Something takes up a lot of memory (4.2GiB)!

Do you have (minimal) example SIC/XE code to reproduce this?

Why are JSUB & RSUB tracked in java anyway?

To enable the "step out" function of the debugger, which jumps out of the current subroutine. The VM keeps track of all JSUBs, but it is a single address per jump.

[adrian4096]

Do you have (minimal) example SIC/XE code to reproduce this?

Haven't tested, but the following should be problematic

example
    JSUB example

A more realistic example of this being problematic can be found at https://ass.si/git/SIC/ass1/src/branch/master/rec.asm#L184-L197 where i had to manually color some subroutines with J calling convention and some with JSUB calling convention.

step out

I'd expect that to run until RSUB instruction, regardless of the implicit stack.

[jetomit]

I'd expect that to run until RSUB instruction, regardless of the implicit stack.

So would I. But looking at the relevant code

https://github.com/jurem/SicTools/blob/master/src/sic/sim/Executor.java#L172-L176

it's probably because Executor doesn't have simple access to the instruction being executed. I'd be happy to check out a PR with some alternative solution.