bug#15573: 24.3.50; `comment-end' in c mode

bug#15573: 24.3.50; `comment-end' in c mode

[Drew Adams]

I don't work in C anymore, but I recently noticed that `comment-end' for
C mode seems to be " */", not "*/".  That presumably makes sense for the
use of `comment-end' as "String to insert to end a new comment.", per
the doc string.

But `comment-end' can also be used to recognize the end of a comment,
and in that case I would think that "*/", not " */", would be
appropriate for C.

Case in point: I have some code that does this:

(while (and (< start end) (setq cbeg  (comment-search-forward end 'NOERROR)))
  (setq cend  (if (string= "" comment-end)
                  (min (1+ (line-end-position)) (point-max))
                (search-forward comment-end end 'NOERROR)))
  (when (and cbeg cend)...))

But that fails to DTRT with a comment such as this, because `comment-end'
is not found:

/*
Some comment.
*/

So my question is whether `comment-end' should not perhaps be "*/" for C
code.

I do realize that `comment-skip' is available.  Perhaps I should just
use that somehow in my code, above?  Advice welcome.

I also see that there is `comment-padding', and that "the strings used
as comment ends are built from `comment-end' and `comment-padding'" (doc
string of `comment-region').  Perhaps `comment-end' should be "*/" and
`comment-padding' as " " should be used with it to do what is done now
using " */" as `comment-end' (and "" as `comment-padding')?

Presumably, since lots of people have been using this for a long time, I
just have something to learn wrt my code, above.  Again, please let me
know.  Thx.

In GNU Emacs 24.3.50.1 (i686-pc-mingw32)
 of 2013-09-30 on LEG570
Windowing system distributor `Microsoft Corp.', version 6.1.7601
Configured using:
 `configure --enable-checking 'CFLAGS=-O0 -g3' CPPFLAGS=-DGLYPH_DEBUG=1'

bug#15573: 24.3.50; `comment-end' in c mode

[Stefan Monnier]

> But `comment-end' can also be used to recognize the end of a comment,

No: that's what comment-end-skip is for.


        Stefan

bug#15573: 24.3.50; `comment-end' in c mode

[Drew Adams]

> > There seems to be no good regexp variable to use, to search for the
> > end of a comment (e.g. in C mode, but therefore generally).
> 
> There can't be.  The syntax of the end of a comment depends on the
> comment-starter that was used.  So don't look for a regexp.  Look for
> a function instead, e.g. comment-forward.

Actually, that's what I have been using, `comment-search-forward' and
`comment-forward'.

> C also accepts //...\n

Right.  Forgot about that.

I still don't see why `comment-end' should not match "*/" without
a SPC preceding it.

bug#15573: 24.3.50; `comment-end' in c mode

[Stefan Monnier]

> I still don't see why `comment-end' should not match "*/" without

comment-end doesn't *match* anything.  It's a string that is meant to be
inserted in a buffer.  It's not a regular expression, "substring
pattern", or something to which things are "matched".

It's very much correct for the user to set comment-end to "-- TOTO  */"
if she so wishes.


        Stefan