Re: rethinking @def*

[pertusus]

On Fri, Jul 29, 2022 at 05:24:53PM +0100, Gavin Smith wrote:
> 
> In texinfo.tex, I have only done the cancelling of the special definitions
> of [ and ] inside @r and @t.  There, the special definition entailed the
> use of a roman font.  In LaTeX \EmbracOn is slightly different in that
> it uses slanted glyphs for brackets, but this could be for both typewriter
> and roman typefaces.
> 
> In texinfo.tex, it was necessary to cancel the special definition in
> @t to allow a typewriter font to be used.  It's also the case that the
> default font for [ in texinfo.tex is roman, not typewriter.  I'm not
> saying that it has to be consistent between TeX and LaTeX, though.
> 
> I imagine there would be very little need for slanted brackets in the
> def arguments.  However, to permit this, I suggest that \EmbracOff{} could
> be limited to the output for @r only.  Does that sound okay?

Yes, sounds ok.  As a side note, in Texinfo TeX, with @r or @{@slanted{
the bracket is thinner than in the default case, or with @code or @slanted.
I do not know if it on purpose.  With @r{@code{ and @r{@t{, the font
seems to be typewriter which seems good to me.

I implemented it.  The differences between TeX and LaTeX are reduced to
* more typewriter fonts in LaTeX
* in @deftype*, @slanted{[} leads to a slanted [ in LaTeX.

This looks good to me for now, and maybe the TeX output will change
in after @def* are considered as code, and @deftype* are not slanted
anymore.

To have upright [ in \textsl{\texttt, I had to do somewhat complex code.
I'll try to contact embrac maintainer to report what I had to do.

-- 
Pat