>> Step 3: fixed where I forgot to for->while in mdoc_validate.c's
>> check_text() loop; removed DECO_SSPECIAL; cleaned up DECO_NOSPACE;
>> cleaned up print_encode() in html.c. The next will be removing the DECO
>> stuff entirely.
>
> Ok (step 4?), one last patch to show full fidelity between DECO and
> ESCAPE before I rip out DECO stuff. This also comes with a fix of a
> yet-unnoticed bug: \F escapes were being treated as \f escapes in terms
> of recognising styles (bold, italic, etc.). However, in the groff
> manual, \F accepts families (Times, Helvetica, etc.) while \f accepts
> styles (duh). I've actually completely removed the FONTFAMILY notion, as
> we don't really support it (it can be put in later, if necessary, but
> it's unlikely as it can't be reliably displayed across output media).

Hi,

Here's the finished patch: enum roffdeco has been completely removed and 
all components are dipping into mandoc_escape() to get the correct 
subsequences.

This has resulted in some clean-up throughout the code, as all the mess 
of escapes is focussed in one place.  It's also given me a chance to 
clean up other areas in the code that danced around string handling 
(e.g., print_encode() in html.c, term_strlen() and term_word() in 
term.c, etc.).

Oks?  Comments?

Kristaps