Files
ss/include/s
robojumper f8e61c7de8 More d/lyt (#52)
* More Lyt (maybe some wrong type associations)

* more meter

* Rename

* More splits and renames

* Fix SizedString conversion operators

* Renames

* dLytBattery_c almost OK

* d_lyt_battery OK

* diff clean

* Continue doing splits while I'm at it

* two more splits
2024-10-06 16:34:31 -04:00
..
2024-06-01 23:57:46 +02:00
2024-09-12 16:36:34 -04:00
2024-06-29 12:36:03 +02:00
2024-06-01 23:57:46 +02:00
2024-06-01 23:57:46 +02:00
2024-06-01 23:57:46 +02:00
2024-06-01 23:57:46 +02:00
2024-08-27 10:48:50 -04:00
2024-06-01 23:57:46 +02:00
2024-10-06 16:34:31 -04:00
2024-09-12 16:36:34 -04:00
2024-06-01 23:57:46 +02:00
2024-06-01 23:57:46 +02:00

The state system in this library was ported from https://github.com/NSMBW-Community/NSMBW-Decomp/tree/master/include/dol/sLib
with differences/modifications outlined below:

## No inline destructors

The NSMBW code uses inlined destructors. In SS, these destructors are all part of the s_StateID.cpp(?) TU,
and pretty much every generic instantiation of the state manager (which for a large part are in RELs)
will call back into those destructors in the main DOL. As a result, the destructors were made not inline
and explicitly instantiated in s_StateID.cpp

## s_StateIDChk.hpp

In order for vtable ordering to match, some classes had to be extracted to a new header.

## sStateMethodIf_c

S_StateMethod_c's destructors calls back into another destructors. sStateMethodIf_c follows the example of
the other abstract interface classes and provides this dtor.

## operator== / != return BOOL instead of bool

We're observing a lot of word-to-bool casts in code after these
operators are invoked, and while there are ways to force the conversion,
this seems the most reasonable.