[Groff] Re: new grohtml patches

[Gaius Mulley]

Werner LEMBERG <address@hidden> writes:

> Very nice!  I'll apply them in the next few days (hopefully).  Some
> comments, based on pic.ms.  It looks like those buglets can be fixed
> easily:
> 
> . pic-6.html: Suddenly all paragraphs have
> 
>     <font color="#000000"> .... </font>
> 
>   (after the image with colour).  I think there should be some code
>   to suppress this.

thanks for spotting this, certainly a bug.

> . The captions of figure 6-5 and figure 6-6 (and many others) are
>   centered by grohtml itself.  I think this shouldn't be done -- the
>   html browser will do it by itself.  Additionally, each centered line
>   is envelopped by <p> ... </p> which adds ugly vertical space.

yes there is quite a bit of ugly vertical space.

> 
>   Maybe the line length should be set to an `infinite' value?

I don't think the infinite value trick will work as tabs and
indentation operate on percentages of the line width.

> 
> . Sometimes, indentation is lost.  For example, look at pic-16.html.
>   The line
> 
>     for variable = expr1 to expr2 [by [*]expr3] do X body X
> 
>   should be indented.  I'm not sure whether this is a problem in
>   grohtml or missing support in the ms macros.
> 
> . On the last page, pic-23.html, I see
> 
>     [ prev | next | top ]
> 
>   which should be
> 
>     [ prev | top ]
> 
>   instead.  A similar problem can be seen in pic-1.html.

sure bugs..

> . groff.css isn't mentioned Makefile.in.  It should probably be
>   handled similarly to grnexmpl.g, this is, copied from $(srcdir) to
>   the current directory if srcdir != builddir.

ok thanks, will do.

> A completely different buglet can be seen after 
> 
>   groff -e -man -Thtml groff_diff.n > groff_diff.html
> 
> Have a look at the documentation of the `Dp' drawing command.  The
> small images representing dx1, dy1, etc. aren't properly aligned
> vertically.  We probably have to modify eqn so that it makes troff
> emit the `depth' (in the TeX sense) of an equation box as a tag -- the
> change to eqn is trivial since it already emits \x with the proper
> depth of that box.

will examine this...

Thanks,
Gaius