Note that I've refactored all the constructors to prevent this issue from happening again.
It already happened with 3E+/3E, and now DPC/F8. Also, FA2/FA looked suspicious.
Finally decided to stop fighting C++, and just use size_t everywhere an
array is used.
Even though this is a 64-bit value, it seems that C++/STL expects it
when dealing with arrays, so I guess we will too.
Converted many C-style arrays to C++ std::array. In the process, cleaned up a lot of magic constants.
Replaced most C-style memset/memcpy operations with C++ equivalents: std::fill/std::copy.
std::copy in particular fixes a few potential bugs, since it operates on the datatype size,
and not just on raw bytes.
- Removed System::PAGE_SHIFT from all classes (except where it's needed inside System); it is an implementation detail of that class that shouldn't be used outside the class, and was making the various calls to setPageAccess very messy
- Changed uInt32 to uInt16, and loop variable to 'addr' for all calls to setPageAccess; this is for consistency, and allows the reader to see how all the code is similar in different classes
- Updated comments in various places (mostly TIA and RIOT) to more clearly show how page accessing works.
- remove peek and poke from classes where the addressing is set up such that they aren't needed
- where possible, move from using 'myCurrentBank << 12' everywhere, to precomputing the value (perhaps a small optimization, but still valid)
This has been in patch form for about a year, and I'm finally
getting around to applying it. Basically, we separate the Cart
class into 3 classes:
- Cart: stuff specific to cartridge emulation
- BSTypes: various bankswitch types, both as enum and strings
- CartDetector: actual bankswitch auto-detection code
Also made more use of smart pointers, and added const in a few
more places.
I promise to get back to the phosphor stuff soon. This is just
to clear my head a little.
Introduced Cartridge::initializeRAM() method, to eliminate duplicate
blocks of code in every bankswitch scheme that uses extra RAM.
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@3316 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
std:: functions into BSPF namespace at all. So I removed them, and have the
calls map directly to the std:: versions.
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@3304 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
To anyone reading this, Merry (belated) Christmas and Happy New Year!
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@3239 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
of these aren't actually bugs per-se, but are to follow good programming
practices.
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@3234 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
a class more generic, but when it's never used, all it does it slow
things down.
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@3016 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
or those which implement their own unique scheme no longer have to
implement bank and bankCount; this base class versions will be used
instead.
Removed the debugger 'bank' command, since it didn't work for all cart
types, and is obsolete now that the UI allows to change banks for
all cart types.
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@2960 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
*always* pass in 0,0,0 for the various bases, it is redundant to pass that
data in on each invocation. So now the c'tor does it automatically.
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@2959 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
cart RAM tab to the debugger. This is tailored to each respective
cart bankswitch type, allowing much more information to be shown than
you'd see in the normal RAM area.
git-svn-id: svn://svn.code.sf.net/p/stella/code/trunk@2921 8b62c5a3-ac7e-4cc8-8f21-d9a121418aba
Stephen Anthony