fill-paragraph-or-region

fill-paragraph-or-region

[Stefan Monnier]

I see that fill-paragraph-or-region was removed, and instead fill-paragraph
was changed to fill a region if transient-mark-mode is active.

What was the reason for this change?

It seems like a bad idea to me: fill-paragraph is already complex enough:
what's the benefit of conflating the two?


        Stefan

Re: fill-paragraph-or-region

[Juri Linkov]

> I see that fill-paragraph-or-region was removed, and instead fill-paragraph
> was changed to fill a region if transient-mark-mode is active.
>
> What was the reason for this change?
>
> It seems like a bad idea to me: fill-paragraph is already complex enough:
> what's the benefit of conflating the two?

As I understood, Richard wants the same solution to be applied to all
commands that modify their behavior when the region is active.

First, a new condition for checking the active region was added to
indent-for-tab-command.  A similar change was made later in ispell-word.
ispell-word is more complex than fill-paragraph, but still a new condition
fits nicely to the logic of this command.

It would be good to decide now what is the preferable way of doing this
and to apply the same solution consistently everywhere.

As I see, there are two possible solutions:

1. modifying the existing command by adding a condition for the
active region;

2. creating a new command with the old key binding:
fill-paragraph-or-region with M-q, ispell-word-or-region with M-$,
indent-for-tab-command-or-region with TAB.  This command decides what
a pair of existing functions to call depending on the active region.

I personally don't have an opinion which one is better.

-- 
Juri Linkov
http://www.jurta.org/emacs/

Re: fill-paragraph-or-region

[Richard Stallman]

    As I see, there are two possible solutions:

    1. modifying the existing command by adding a condition for the
    active region;

That is the method I've chosen.  I've decided not to change the
traditional command names, such as fill-paragraph, merely because they
now do something special when called interactively while the region is
active in transient mark mode.

Re: fill-paragraph-or-region

[Richard Stallman]

    It seems like a bad idea to me: fill-paragraph is already complex enough:
    what's the benefit of conflating the two?

The benefit is to keep `fill-paragraph' as the name of the command
that M-q runs.  That name is better even though it now has the feature
of doing something different for certain interactive calls.

The complexity added to `fill-paragraph' is a handful of lines.
Not significant.

Re: fill-paragraph-or-region

[Stefan Monnier]

>     It seems like a bad idea to me: fill-paragraph is already complex enough:
>     what's the benefit of conflating the two?

> The benefit is to keep `fill-paragraph' as the name of the command
> that M-q runs.  That name is better even though it now has the feature
> of doing something different for certain interactive calls.

> The complexity added to `fill-paragraph' is a handful of lines.
> Not significant.

OK, actually the consistency with the other commands that obey the region (if
transient-mark-mode is set and the region is active) is a good argument.


        Stefan

Re: fill-paragraph-or-region

[David Reitter]

On 18 Oct 2007, at 17:04, Stefan Monnier wrote:

> I see that fill-paragraph-or-region was removed, and instead fill- 
> paragraph
> was changed to fill a region if transient-mark-mode is active.

Whatever the internal implementation, this is a good idea from the  
user's perspective. We have had M-q bound to `fill-paragraph-or- 
region' for some time now in Aquamacs precisely because of the  
shortcoming that this change fixes, and I haven't heard any  
complaints from users.