[PATCH evolve_ext V2] fold: take an explicit list of revisions (BC)

Sean Farley sean.michael.farley at gmail.com
Thu Jun 26 13:55:59 CDT 2014


Greg Ward writes:

> On 24 June 2014, Augie Fackler said:
>> I agree with Greg here - the magic default of 'hg fold $REV' -> fold
>> together $REV::. feels a little spooky. This feels like the right case
>> for an alias to me.
>
> OR if it's as important a use case as Pierre-Yves claims, add an
> option. E.g. make
>
>   hg fold [-r] REV...
>
> (as implemented in my patch) the default. Then add
>
>   hg fold --from REV
>
> to restore the current (REV::.) behaviour.
>
> Hmmm. Time for a poll? Here are the options:
>
>   1) status quo: "hg fold REV" and "hg fold -r REV" are different
>
>   2) my proposal: "hg fold [-r] REV...", with --from option
>      to satisfy people who like the current "hg fold REV" semantics
>
>   3) Pierre-Yves' proposal: "hg fold [-r] REV" same as current
>      "hg fold REV", with "hg fold --only REV..." option to
>      satisfy people who like the current "hg fold -r REV..." semantics
>
> I'm pretty sure no one will vote for #1. Go ahead, surprise me!
>
> Sounds like Sean, Augie, and I are for #2.

There's a big of hack we could do here:

a) $ hg fold
b) $ hg fold SINGLEREV
c) $ hg fold REVSET 

Currently, (a) doesn't make sense so we could have it default to folding
with the parent '.'. Also, (b), I think, only makes sense if you want to
fold from '.' to SINGLEREV. That leaves (c) to be an arbitrary fold that
could work with an in-memory implementation.

> Clearly Pierre-Yves likes #3. Who else?

No one. Pierre-Yves is dangerously close to:

https://yourlogicalfallacyis.com/anecdotal
https://yourlogicalfallacyis.com/appeal-to-nature


More information about the Mercurial-devel mailing list