Re: Paragraph indentation suppression

[Simon Law]

On Mon, Apr 07, 2003 at 07:11:25PM -0400, Karl Berry wrote:
> Well, two minor comments:
> 
>     +\let\parorig=\par
>     +%
>     +\def\yesword{yes}
>     +\def\noword{no}
>     +\def\firstparagraphindent{\parsearg\dofirstparagraphindent}
>     +\def\dofirstparagraphindent#1{%
>     +  \def\temp{#1}%
>     +  \ifx\temp\yesword
>     +    \def\par{\parorig}%
> 
> 1) I think it would be better to \let\parorig=\par to get the runtime
>    definition (so to speak), within \dofirstparagraphindent, instead of
>    using the definition when texinfo.tex is read.  A number of things
>    redefine \par.
> 
> 2) Might as well say \let\par=\parorig when restoring, not \def, to save
>    a level of expansion.

        Both of those seem to be quite rational.

> Anyway, as for the general approach, I am frightened :).  I wonder if it
> would be more robust to use \hangindent=-\parindent (\hangindent gets
> automatically reset to 0 after each paragraph, so it would by definition
> only affect one paragraph) or \everypar={\kern-\parindent\everypar={}}.

        I think that it doesn't matter much.  (\hangindent=-\parident
may be optimal.)  The biggest problem, I would say, is to ensure that
\par occurs once and only once after every section.  This way, it
becomes rather simple to do the Right Thing on the first paragraph, no
matter what.  Unfortunately, I don't quite know how to od this.  I shall
have to think harder.

> Some research is in order, I think, to see what other solutions have
> been found.  I know you're just trying to solve an immediate problem for
> an upcoming printing, and your solution is fine for that (assuming the
> gcc manual prints with it :), but before we put something like this into
> the default texinfo.tex, I'd like to take a little more time with it.

        Oh, it prints fine.  For a while, I was just putting @noindents
all over the place before I got lazy and looked for another way.

Simon