Re: awesome feature (yet to be added?)

[Thorsten Jolitz]

Davis Herring <address@hidden> writes:

>> Judging by the uptake.  We are the only two to think this is useful.  :)

No, I would really like to see a generic `invert-comment' function in
Emacs. My use case is outorg.el that converts programming language buffers
(structured the outshine way, i.e. with outcommented Org-mode headers)
temporarily to org-mode for text editing and then back to the
programming major-mode. 

That means all the comment sections of the source file are uncommented
when converting to Org, and outcommented again when converting back to
the original major-mode. I do this linewise, and in fact outorg can't
handle those special multiline comments like in C or Java. 

A robust generic (= major-mode agnostic) `invert-comment' function as
part of Emacs that knows how to handle multiline comments too would come
very handy.

> I thought about implementing `invert-comment' a few years ago, but I
> wasn't sure what to do about code like this:
>
>   /* old_implementation_prepare (); */
>   /* old_implementation_go (); */
>
>   new_implementation_prepare ();
>   /* The new implementation needs a buffer allocated. */
>   char *p = malloc (new_implementation_size ());
>   new_implementation_go (p);

IMO such a function should only act on comments that start at the
beginning of line, indented comments should remain untouched.

> What to do about the actual comment embedded in the "new" code?
>
>   x = foo () + /* bar () + */ baz ();
>   quux (x, x + 1);     /* quux now accounts for bar internally */
>
> What does it mean to toggle commented-ness here?

Again, IMO these comments should remain untouched. E.g. when converting
source files to org-mode, code section are enclosed in source-blocks

#+begin_src lang
  x = foo () + /* bar () + */ baz ();
 quux (x, x + 1);     /* quux now accounts for bar internally */
#+end_src

and the embedded comments clearly belong to the code. 

> Your idea of markers likely works out better than "toggle the comments"
> because of cases like these.

It would indeed be very useful if there is a way to further act on the
region/line that was inverted after the function has done its job. So
returning start/end position, set markers or running a hook or so would
be nice.

There were quite a lot of proposals in this thread. Any chance that this
is actually outdiscussed and included in Emacs any time soon?

-- 
cheers,
Thorsten