Stefan Sperling <stsp@stsp.name> wrote:
> On Fri, Jul 14, 2023 at 11:44:08PM +1000, Mark Jamsek wrote:
> > I've made a few changes that I think make it a lot nicer, and have
> > added some non-trivial and somewhat contrived tests too.
> > 
> > I nixed the hand-rolled overflow code so we use strtonum(3), and made
> > the keyword parsing strict. That lax keyword comparison was by design
> > initially but I forgot to remove it; I was using -cBASED in testing :)
> > 
> > Something I just noticed that I like about the ':' delimiters is ^W
> > removes back to the ':', which is handy. And the +/-N works great with
> > tog's count modifier. You can open tog, see which commit it is you
> > want with Nj, and use the same N count (or the headline index-1) in the
> > 'got * -c:head:-N' command! This is going to save a lot of copypasting
> > and misremembering commit ids. It's pretty neat :)
> 
> Ok by me.
> 
> I am certain we will be making more tweaks to this over time but
> this seems like a good starting point for further work to happen
> in the tree.
> Thanks for addressing the problem with ambiguous syntax. That was
> the biggest issue from my point of view.

Thanks! And no worries wrt syntax. I think it's much better like this :)

Yes, I want to make it consistent and expand support (where it makes
sense) to all commands that take a commit id or reference.

Then I want to add more keywords like temporal expressions.

But I felt like this was a good start; it already provides a convenience
I think I'll use daily, and wider use will help inform future work.


-- 
Mark Jamsek <https://bsdbox.org>
GPG: F2FF 13DE 6A06 C471 CA80  E6E2 2930 DC66 86EE CF68