Re: M-\ does not work with prefix argument as documented

[Kevin Rodgers]

Dieter Wilhelm wrote:
> Kevin Rodgers <address@hidden> writes:
> > Dieter Wilhelm wrote:
...
> > Neither the Emacs manual nor the delete-horizontal-space's doc string
> > says anything about how prefix args are handled.  But it seems like a
>
> You're right, I interpreted the documentation in this way because I
> wanted M-\ to work also with a prefix argument.  My notion was that a
> numeric argument to an interactive function means implicitly that a
> prefix argument must be working as well.  The next time I'll be
> looking out for the word prefix.
>
> Thank you for installing your patch. (You did install it, didn't you?)

No, I don't have that permission.  But RMS posted a similar patch,
so I assume that was committed.  The main difference was that my patch
included changes to the doc string and the manual.

> > > Could you, please, also make the command symmetric? For example
> > > negative arguments are killing on the left hand side of the point and
> > > positive arguments on the left hand side.  The default would remain as
> > > is.  (There was recently a request--unrelated to M-\--for such a
> > > feature on gnu.emacs.help, killing white space left or right of the
> > > cursor).
>>
> > That would be cool.
>
> 8-), OK, I'll have a go at it.  But probably only after the release.
>
> What I need to know is how could I distinguish in the function whether
> the user gave a prefix argument '1' or whether it was the default
> argument '1'?  The default argument would mean kill on both sides, the
> same numerical value would mean just kill to the right.  Is it
> possible to intercept the argument before the (interactive "*p") form
> is evaluated?  

I think that would be a poor interface: for any command foo that takes
a numeric argument, `M-x foo' and `M-1 M-x foo' ought to mean the same
thing.

But yes, you can examine the value of current-prefix-arg, which will
be nil in the `M-x foo' case.

--
Kevin