Re: yank-match.el -- yank matches for a regexp from kill-ring

[Karl Fogel <kfogel@red-bean.com>]

Juri Linkov <juri@jurta.org> writes:
> Modifying `yank-pop' in simple.el to take into account the initial
> command `yank-match' has also benefits of using the recent and more
> correct code to do this.  For instance, your command uses `delete-region'
> to replace the previously-yanked text, but `yank-pop' has more advanced
> code that calls (funcall (or yank-undo-function 'delete-region) ...)
>
> Also your command uses `insert' to insert the previously-yanked text,
> but `yank-pop' uses `insert-for-yank' to strip text properties from the
> inserted text according to `yank-excluded-properties'.  I guess your patch
> was written when there was no such functions in Emacs?

Probably -- the code is a bit over ten years old now.

> I think it would be safer to reuse code in `yank-pop' and just find
> places in it that differ from `yank-match' functionality, like
> `current-kill'.  Maybe you could just add a new match-string argument
> to `current-kill' that will rotate the yanking point by N matching
> places?

All these are good ideas, yup.  I grok your general message here:
*really* incorporate 'yank-match' into the existing yank code, don't
just stitch it in in the quickest way possible :-).  I will make sure
to DTRT.  (I wish we were in bzr already, it would make this slightly
easier... But only slightly, not a big deal either way.)

Thanks for the help!

-Karl