[Groff] Bug#67545: groff: HTML driver doesn't handle overstriking well

[Colin Watson]

Hi,

I have the following Debian bug report on grohtml from some time ago.

On Fri, 21 Jul 2000 at 16:55:05 +0200, Martin Schulze wrote:
> While the following works fine:
> 
> >  1. man -Thtml klogd > klogd.html
> >     lynx -dump klogd.html > klogd.txt
> >     rm klogd.html
> 
> it does not work with s/klogd/perlipc/g.  From a first glance the
> output of 'man perlipc looks similar to the one from klogd.  Thus
> I'd expect that HTML output looks similar as well, but that's not
> the case:
> 
> <a href="#N NA AM ME E">N NA AM ME E</a><br>
> <a href="#D DE ES SC CR RI IP PT TI IO ON N">D DE ES SC CR RI IP PT TI IO ON 
> N</a><br>
> <a href="#S Si ig gn na al ls s">S Si ig gn na al ls s</a><br>
> 
> Please take a look at it.

(You can see the full log at http://bugs.debian.org/67545, but there
isn't really much of interest after that.)

pod2man uses various constructions that cause 'pre-grohtml troff -mandoc
-Thtml' to produce output like this:

24NtN
cAtA
cMtM
cEtE

This causes other devices to overstrike, but post-grohtml doesn't
implement this correctly. I have a tentative patch which defines a new
element 'overstrike' of the style struct, sets it in
html_printer::set_continuation() if sbuf_end_hpos == env->hpos + w +
sbuf_kern, and overrides the font if it's set, but I have the feeling
I'm going about it the wrong way - sbufs containing successive
overstruck characters need to be merged and rendered better than
<b>N</b><b>A</b><b>M</b><b>E</b>!

Any thoughts? Sorry if this has already been reported and I missed it.
I've attached perlipc.1.gz by way of reference.

Thanks,

-- 
Colin Watson                                  address@hidden

perlipc.1.gz
Description: Binary data