This a follow up to #82 (comment). I recently implemented AsciiMath support. It has just been released.
However, I now realize the asciimath inline macro is incorrectly handled. Basically, it is handled in exact the same way the latexmath inline macro is. That means the macro is not processing AsciiMath code, it's actually processing LaTeX directly. In other words, currently the asciimath inline macro is functionally equivalent to the latexmath inline macro. This is an error.
I'm working on a patch, but I may need some help on it. This issue was cause by my inability to comprehend how inline stem blocks are currently handled. I think I've now got things figured out, but I'd like someone to review the patch in detail to make sure the implementation is not broken again.
This is a follow up to #84: