[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: manpages, hyphens, minus signs, HTML, and lintian

On Sat, Jul 02, 2016 at 09:06:57AM +0100, Colin Watson wrote:
> On Fri, Jul 01, 2016 at 10:57:45PM -0700, Josh Triplett wrote:
> > On Fri, 15 May 2015 22:15:14 +0200 Jakub Wilk <jwilk@debian.org> wrote:
> > > Yeah, these days even upstream groff renders both - and \- as
> > 
> > However, this doesn't appear true with current groff when rendering to
> > HTML.  "man -H" (or "man -Thtml") passes "-" through as "-", but renders
> > "\-" as "&minus;", which browsers typically render as U+2122.
> I think that's simply a bug in groff and should be reported as such.
> (To fix it, we could for example adjust \- specifically when rendering
> man/mdoc output to HTML.)

Reported as a bug; thanks.

> > As far as I can tell, manpages should never use "\-" at all unless they
> > actually want a mathematical minus sign (or in the one line in the NAME
> > section between the program name and description, as whatis and apropos
> > apparently require that).  (For manpages that want an em-dash, use the
> > four-character sequence "\(em".)
> > 
> > Any time a manpage wants "the dash corresponding to the key '-' on the
> > keyboard", which includes any text the user would type on the keyboard
> > using that key such as an --option or command-name, the manpage should
> > just use "-".
> That's specifically contrary to upstream's consistent typographical
> advice over the years, and your suggested advice has its own problems
> such as inappropriate line-breaks due to hyphenation.  I don't mind that
> we've dropped the lintian check nowadays, but please don't reverse it.

I only made that suggestion on the assumption that groff's HTML
rendering behavior was working as designed and intended; if it's a bug,
then please disregard that.

- Josh Triplett

Reply to: