Re: [groff] improve a few terminal renderings of special characters

[Ingo Schwarze]

Hi,

this likely isn't quite as bad as you seem to fear.

John Gardner wrote on Wed, Aug 29, 2018 at 11:01:29PM +1000:

> I should ask, what's the most reliable way of running `man -a -w`?
> I thought this was universal to `man` implementations,

It is, see

  http://mandoc.bsd.lv/man/man.options.1.html#w

It has been available since AT&T Unix v7, consequently in BSD since
3BSD, in Eaton man, man-1.6, man-db, ... - so for practical purposes,
everywhere.

> but Solaris told me otherwise.

That is very strange.  Given that Solaris is derived from Sys V Unix
(which had it) and from CSRG BSD (which also had it), i have no idea
how they managed to break such a simple option.

> illumos uses `man -w` to update the `makewhatis` database,

What a terrible idea.  Illumos is derived from Solaris.  So probably,
they mistakenly considered the option free because Solaris already
broke it earlier, and reused it for a different purpose, failing
to do the required research first.  They should probably fix that
mistake.

> and this is part of what killed my first attempt at finishing Roff.js.

I suspect you will get away with using -w if you merely implement
a workaround for Solaris and Solaris-derived systems like Illumos:
if man -w fails, use man -l with a bit of postprocessing instead.

Of course, do not use -l otherwise:  It is wildly non-portable,
and for this purpose, purely a Solaris idiosyncracy:

  http://mandoc.bsd.lv/man/man.options.1.html#l

Yours,
  Ingo