GitLab

The new way of detecting RWP is superior, and we don't have time to rework _rwport to use it.
This may come back in the future (if required), but I honestly don't see a need for it.

I'll admit that some of these suggestions are of dubious importance, but at least we shut up the tester :)

Again, this is a relatively untested BS scheme, with only one test ROM available.

This is currently mostly untested, since this scheme isn't widespread and has only a few test ROMS.

That's an *awful* lot of work to simply pass info to each class :)

- abstracted RWP code into Cart base class (only F8SC converted for now)
- RWP now works by analysing RAM accesses before and after each instruction, catching all occurrences

Fixes for compiling without debugger support.

fixed #398

updated change log

Previously these values were from whatever was on the databus, but new testing
indicates that they are semi-random (ie, random, but with some pattern).

The new code bases the values in part on the md5sum of the ROM (for randomization),
but also pre-generates a set number of values (for a pattern).

Currently only F8SC scheme is implemented.  The remainder will be done once
we test this one.

The new code introduces a generic 'EmulationWarning' exception that can
be thrown for any reason.  The RWP functionality has been ported to use
this, with the main benefit that if we ever have to add new exceptional
cases, we no longer need to modify M6502 class; it will pick up on the
new exceptions and pass them to the EmulationWorker as necessary.

Functionally, the RWP stuff works the same as before, just with a
different implementation behind the scenes.