On Tue, Jul 12, 2022 at 05:15:50PM +0200, Omar Polo wrote:
> Omar Polo <op@omarpolo.com> wrote:
> > Stefan Sperling <stsp@stsp.name> wrote:
> > > On Mon, Jul 11, 2022 at 09:12:48PM +0200, Omar Polo wrote:
> > > > while talking on irc this issue pop up: there can be some strange
> > > > interactions between the commands in the histedit script.
> > > > 
> > > > while none of this is really surprising, preventing `mesg' in some
> > > > situations where the outcome is non-sensical.  In particular, a fold or
> > > > drop followed by a mesg.  (in the first case the commit would be first
> > > > reworded and then deleted, in the second case it would be reworded and
> > > > then the new text discarded by fold.)
> > > > 
> > > > so, the only valid options are pick -> mesg and edit -> mesg (which is
> > > > used by the regress test at least.)  It would be nice to prevent mesg ->
> > > > mesg too maybe.
> > > 
> > > Indeed, two consecutive mesg lines should also be a syntax error.
> > 
> > on a second thought, here's a slightly different version that also
> > catches the consecutive mesg (and adds a regress for it ;)
> > 
> > I tried also to add a sentence to the manpage, but suggestion for a
> > better phrasing are welcome (or i can drop it and let you fix the manpage)
> 
> second thought of a second thought, i like this one more:

Very nice indeed. OK.
I like the idea of storing the previous command code instead of a pointer.