bug#50731: `progress-reporter-update' docstring and `backward-sexp' interaction

[Eli Zaretskii]

> From: Stefan Kangas <stefan@marxist.se>
> Date: Thu, 23 Sep 2021 00:12:08 -0700
> Cc: 50731@debbugs.gnu.org
> 
> > Once again, why does checkdoc want you to have 2 spaces there? what
> > rule is it that ends up requiring that?
> 
> Sorry, I didn't understand you were asking about checkdoc.

Because you said:

> This in turn leads to checkdoc flagging this as a mistake, and asks you
> to put two spaces after dot.

I read this as the issue you'd like to solve: to prevent checkdoc from
flagging this as a mistake.  Was I misinterpreting your report?

> The code
> there is really simple: find the next dot (that is also at a word
> boundary), run `backward-sexp', use a regexp to see if this is an
> abbreviation.
> 
> So it lands on the "r" in "reporter---i.e." and says, no, this is not an
> abbreviation I know about

Isn't _that_ a bug, right there?  Why does checkdoc use this strange
strategy for identifying abbreviations?  A punctuation character that
precedes an abbreviation shouldn't interfere with recognizing the
abbreviation, IMO.  And using sexp movements in text that is not
really a sexp is perhaps "the original sin" here?

IOW, I see no reason whatsoever to believe backward-sexp is the
problem here.  The problem is in checkdoc, and IMO should be solved
there.