[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: return values

From: Andreas Roehler
Subject: Re: return values
Date: Fri, 13 Nov 2009 11:56:40 +0100
User-agent: Thunderbird (X11/20081227)

Alan Mackenzie wrote:
> Hi, Andreas!
> On Tue, Nov 10, 2009 at 11:33:39AM +0100, Andreas Roehler wrote:
>> Hi,
>> IMHO it might be helpful, if classes of functions have a regular return
>> value.
>> For example functions which move point should return point.
> Here, it's probably more accurate to say "would have been a good idea",
> since some users of the current functions will use the current return
> values, even if only relying on the fact that they're always nil.
>> `goto-char' does it presently, beginning/end-of-defun not.
> Indeed.
>> Also I suggest mentioning the return value in docstring.
> I am in favour of mentioning the return value of EVERY function, even
> when (or _especially_ when :-) it's only a sentence like "The return
> value is of no significance".


Hi Alan,

what about not to have just one return value
hard-coded, but let it being choosed via optional

Let's consider the case of `forward-comment' employed
inside a `delete-comments-in-region' function.

First `forward-comment' must move. If a comment-section
is reached, function must determine its borders in
order to delete.

Now this second operations basically is redundant, as
`forward-comment' determined borders already, otherwise
it couldn't know about nested comments.

A forward-comment is already able to deliver beg, end
--separately or as a list-- and the comment itself as a

So instead of executing two consecutive functions to
determine, forward-comment could take a flag:

For example

b - return beginning
e - return end
l - return beginning and end as a list
s - return found item as a string




reply via email to

[Prev in Thread] Current Thread [Next in Thread]