Re: Matching programming language identifiers, not "words"

[Stefan Monnier]

> ! a word or symbol constituent character follows.

No, my point was that a "word constituent" is also a "symbol constituent",
because a "symbol constituent" has either word or symbol syntax.

This distinction between "symbol syntax" and "symbol constituent" as
currently used in the manual is good.  We just need to make it more clear by
explaining it.

My sample text was wrong in that instead of 

   A symbol-constituent character is a character whose syntax is either
   @vsamp{w} or @samp{_}.  @samp{\_<} matches at the beginning of ...

we should say something like:

   A symbol-constituent character is a character whose syntax is either
   word (@samp{w}) or symbol (@samp{_}).  @samp{\_<} matches at the
   beginning of ...

We can drop the (@samp{w}) and (@samp{_}) if they are too
elisp-oriented.


        Stefan