– Improved the key code emulation for Pause/Break key, and provided the alternative key code as well (accessible from the right-click menu).
– When the button assignment changes due to SHIFT function or active profile switch, there has been a feature to ignore the already pressed input « for the specified time ». This feature was enhanced to be also able to ignore the input « until it’s pressed again ».
– Added « Tab » key in the right-click assignment menu.
– Bug fix: When the same button’s definition is triggered from multiple buttons, (e.g. Button[20] is called from both LT and RT), it wasn’t released at the expected timing.
– Bug fix: Fixed the occasional display issue when « Options » tab was selected.
– Added some more unaligned read/write edge case handling. This addresses a problem in Theme Park, possibly among other things.
– Changed some undocumented RISC instruction behavior to match some of my local hardware tests. One of these changes is known to correct an issue with object directions in Defender 2000.
– Added more Jaguar GD functionality. Filesystem functions are stubbed out, but I wasn’t sure if anyone really needs/wants this fully implemented, so let me know if it’s important to you.
– Fixed a few « undefined » Blitter cases. (not known to affect any existing titles)
– Linux ARM64 builds are now part of the automated build set. Special thanks to cubanismo for guiding me through the minefield of cross-compiling for ancient Linux targets.
– Significant improvements to the network rollback functionality, propagated back from BigInstinct.
– All kinds of new platform/framework functionality, also propagated back from BigInstinct. I didn’t feel like going through hundreds of changes to make a coherent summary, but one of the more notable additions is a byuu-style run-ahead option.
– Exposed a bunch of sound spatialization functionality to the scripting API.
– Exposed MOD and MP3 playback to the scripting API.
– Added a timing fix for Val d’Isere Skiing and Snowboarding.
– Various developer build changes to more accurately reflect hardware and/or detect errors which would break on non-developer builds. (this includes enabling M68K address exceptions)
– Lots of « anti-emulator » code will be failing in developer builds at this point, and some related functionality may be propagating to non-developer builds soon; relying on inaccuracies to prevent your software from running in BigPEmu is unwise. If you really want to prevent your software from running in BigPEmu, contact me about better/future-proof methods.