RTC testrom fails about gambatte-speedrun

[entrpntr]

There was previously this preliminary MBC3 RTC testrom which we had failures on, not sure if the two test different things.

Reason for going half speed in DMG/SGB can be traced to this change in time.cpp. The initstate code sets the relevant bit in initial DMG ioamhram, hence why and'ing with CGB was necessary. The "best" solution would be to refactor double speed out to be its own thing to avoid checks like that throughout the code; alternatively we could just wire an isCgb() check into time.cpp.

[CasualPokePlayer]

There was previously this preliminary MBC3 RTC testrom which we had failures on, not sure if the two test different things.

Oh yeah I remember that, looking at the failures on the prelim ROM, they appear to be TestRegisterOverflows (one subtest failed, not sure which), TestWrittenValueIsNotReadableUntilLatched, TestOutOfBoundsWriteAndIncrement, and TestRegisterSizes. Some of the fails seem to be tests within the ROM I linked (TestOutOfBoundsWriteAndIncrement and TestRegisterSizes have equivalent tests), but the ROM I linked just doesn't do the tests anyways due to other tests failing with it (in particular, all register write tests just fail... which seems horribly wrong as it supposedly means GSR is treating all RTC registers as read-only). Somewhat odd as the prelim test ROM has some indication GSR should be treating them as R/W?