Can someone please explain to me why people in this thread are trying to parse a context-free language using regular expressions?
Parsing is done in Scheme, not with regexes, they're only used for substitions of inline elements which (with the exceptions of spoilers) are not composable.
That minimalist typographically sound markup bold italics is utter nonsense was designed almost a decade ago (without spoilers). First implemented in C, then in Scheme, then in Erlang. The Erlang textboard was 95% finished when w4ch's /prog/ got captchas but the progrider's admin quickly hosted a tablecat alternative and I eventually gave up on writing that alternative textboard software (it was announced in a shelter thread at tablecat, if anyone remembers that). I then ported the Erlang's parser which was a really nice one-pass parser for binary strings, in MIT Scheme. In the very limited time frame I had to write SchemeBBS, I had to make some trade-offs to "finish" the project in time. Regex weren't supposed to be used at all and it's very plausible that I'll rewrite the whole parser one day. I find it to be too slow