On Tue, 29 Jun 2004, Jonadab the Unsightly One wrote: > If we have $foo.undo(), then we will want a multi-step undo to go with > it, probably $foo.undo($n), with $n able to be negative for redo. Are
Definitely! I didn't add that to the point that it wuld have been obvious, and I wanted to keep the message as simple as possible. Not sure about negative values... > we prepared to give the mouse that cookie? (This is not intended as a > rhetorical question; I suspect people will stake out both positions.) Well, I've slightly changed my mind myself too. As we have C<redo>, it may make more sense and better fit in the natural-language-like scheme of Perl to have a builtin undo() for blocks, with the obvious caveat that not everything that was done a block can be undo()ne, e.g. wrt removed files et similia... > I heard a rumour we were getting continuations (a la Scheme). They Even if it is not expressed properly, I think that the idea hinted at above is more similar to this one. > wouldn't be tied to a specific variable like what you propose, but > they would allow the state of the entire process to be rolled back to > an earlier point, or something along those lines. Of course, the Yep, for sure this makes much more sense! Michele -- > (I don't know why I can't resist jumping in on this thread. I wonder if > it's a mass hypnosis worm ...) It must be the usenet equivalent of a spectacular car crash, where passersby feel compelled to stop and stare. - Keith Keller in clpmisc, "Re: Clarifications"
