One superficial observation: I've noticed you tend not to stack your closing parens. For example, you have:
(def txt-block (text)
(flat ; some shorthand features require splicing
(with (anon? nil ext? nil tag "p" content "")
(map
(re-split "\n\n" text)
txt-html-block
)
)
)
)
Instead of:
(def txt-block (text)
(flat ; some shorthand features require splicing
(with (anon? nil ext? nil tag "p" content "")
(map
(re-split "\n\n" text)
txt-html-block))))
Of course, you're welcome to format your code however you like. However, you'll notice most people here (and in other lisp communities, I think) would choose the latter. I like it better because it shortens code and most editors make it easy to check who a paren is paired with if you're ever in doubt.
I've seen the latter style in other people's code, and I do indeed use it for shorter functions with small amounts of indentation. I use BBEdit, which does have a balance option, (Command-B) but I still like to be able to see when something's out of line at a glance with longer functions. I don't mind either style, though.
This is the beginnings of an implementation of Textile in Arc. The main textile routine is empty; txt-block doesn't work because re-split, despite being in re.arc, doesn't seem to work (at least for me), so the two main working things are txt-html-block, which needs a couple variables to be defined in order to work, and txt-span and friends, which perform the standard span-like substitutions, without attributes as yet. There are probably gotchas in both routines.
