Re: mwm(1) is a mess UGHGH!

[Robert Elz <kre%munnari.OZ.AU@localhost>]

    Date:        Thu, 15 Jun 2023 09:59:42 +0000
    From:        Todd Gruhn <tgruhn2%gmail.com@localhost>
    Message-ID:  <CA+9Akf8y4QRzKhzn3dYsNoqKLreMQAFXMqxL2SBLwcFsBq4r5Q%mail.gmail.com@localhost>

  | my perlman2 line is this:
  |
  |      if ($section)
  |        {
  |          $section =~ s/[()]//g; # remove parens
  |          return "man -s $section $man 2> /dev/null | rman |"
  |        }
  |
  |
  | Is it better to do it using this ??
  |
  |   tbl /usr/pkg/man/man1/mwm.1 | groff -Tascii -mandoc

No.   It would be much more complex.   First the '1's in that are the
$section from the perl version (but that is trivial to handle).

More importantly, tbl is not appropriate if the man page in question has
already been processed by tbl (which it seemed as if that one might
have been).   The same is true for eqn, refer, pic, ...    It is fine (but
perhaps a bit slow) to use those preprocessors on man pages that don't
need them (they just become a sluggish "cat" in that case), and it is
necessary (when using *roff as the formatter) when using man pages that
do require them, but they can't be used twice - so you cannot use tbl
if the man page has already been processed by tbl, and while it would
be possible to create a tool which would try to work that out (and
similarly for each of the other preprocessors) it isn't a trivial task
(you cannot just do it with a simple grep).

I also have no idea what the "rman" in the perl pipeline is, or what
its purpose might be.

Maybe someone else can help with some of this - but for some (fortunately
not very many these days) man pages, you just have to do some work to
discover the correct way to format them.

kre