Now that I am no longer responsible for actually implementing Pod
thingies, I feel free (existentially!) to make sage pronouncements and
saucy suggestions. And here's one idea that recently occurred to me:
I'm a not-very-sophisticated Emacs user, so I beg a thousand
cyberpardons if my understanding of the problem space is wonky. But:
Last I looked at emacs's Perl support, the only way to browse Pod from
within emacs was either just to run a shell command to perldoc, or to
call the smarter function cperl-perldoc. But cperl-perldoc (from
cperl-mode.el, of course) simply passes off to calling pod2man (either
directly or via perldoc and Man-getpage-in-background) with TERM set to
"dumb". Of course, this produces just plaintext, with no fontification
and with no hyperlinking of the the L<...> things. That makes me sad.
I figure there's two different ways that someone with better emacs
skills that I could make this work:
* Have perldoc spit out the raw *roff code and then coax some part of
woman.el into doing the formatting. (woman.el parses and formats *roff
code right in emacs, instead of having *roff do it.)
* Or completely avoid the whole detour thru nroff, and instead have
perldoc use a different formatting class (via a -M switch) that would
dumpthe parsed Pod as a big Lisp form. This could be easily produced by
copying the short and sweet Pod::Simple::XMLOutStream and tweaking it to
spit out Lisp instead (freely scavanging from as_Lisp_form in
HTML::Element). Emacs could then just eval that Lisp form, and then
accordingly produce a buffer with the text bits having appropriate
formatting and also hyperlinks that work.
I figure an experienced elisp superstah could do either of these on some
well-caffeinated afternoon. If there are any within the sound of my
voice: you know you want to do this!
--
Sean M. Burke http://search.cpan.org/~sburke/
