Re: How do I highlight word at point?

[Xah <xahlee@gmail.com>]

On Oct 19, 4:58 pm, Nikolaj Schumacher <m...@nschum.de> wrote:
> Xah <xah...@gmail.com> wrote:
> > press a key, expand selection to the current word, press again, expand
> > to the next semantic unit (with respect to the current lang/mode),
> > press again, expand further.
>
> I have a strong interest in such a function, as well.
>
> > However, it doesn't work when the cursor is in a screwed nested
> > position. For example:
>
> > (something here A (and) that)
>
> That case I can fix.  Here's some code I wrote a while ago...
>
> (defun my-mark-sexp (arg &optional incremental)
>   "Mark the sexp surrounding point.
> Subsequent calls mark higher levels of sexps."
>   (interactive (list (prefix-numeric-value current-prefix-arg)
>                      (or (and transient-mark-mode mark-active)
>                          (eq last-command this-command))))
>   (if incremental
>       (progn
>         (up-list (- arg))
>         (forward-sexp)
>         (mark-sexp -1))
>     (if (> arg 1)
>         (my-mark-sexp (1- arg) t)
>       (re-search-forward "\\_>")
>       (mark-sexp -1))))

Super.

is it possible to make it work with selecting a string?

suppose your cursor is on the word “thing” below:

([
«  ‹   "some thing here"  ›   »
{ ht}
「 chinese bracket 『more』  」
])

after second press, it'd be nice to select the string.

> > Ideally, this mark-semantic-unit should just extend a semantic unit,
> > where what's considered a semantic unit depends on the language. But
> > this i imagine would be rather a non-trivial problem. I am not sure
> > emacs's syntax table system is rich enough to be used for this.
>
> As you can see from your own code, character syntax isn't enough.  There
> would have to be a real parser involved to detect where statements start
> and end.  As far as I know even Semantic doesn't parse any function
> bodies, and that's probably the smartest lib we have.  (Luckily lisp is
> that easy to parse.)
>
> Maybe there is an adequate heuristic...

“Semantic”? is that a elisp package?

yes, i think heuristics will do sufficient job. C syntax langs (java,
javascript, C, C++, C#, LSL, and even to large extent perl), has
little tell tail signs good enough for a practical solution. typically
just of this form:

import xyz;

f f (type x, type y) {
...;
...;
}

to start with, we can have the cmd such that:
one press → select current word
2nd press → current string if inside string
press again → current line
press again → current content inside {}
press again → next outer braces
press again → whole function def

i think this is not difficult to implement this to the degree it works
correctly 95% of time.

the nxml and js2 packages contains full parser for xml and javascript.
I think eventually lang modes in emacs would have full lexers.

  Xah
∑ http://xahlee.org/

☄