Evolve naming troubles

Matt Harbison mharbison72 at gmail.com
Mon Jan 23 00:57:36 EST 2017


On Fri, 20 Jan 2017 13:11:40 -0500, Sean Farley <sean at farley.io> wrote:

> While we are in a freeze and have some time to think about things
> outside of features, I'd like to bring up naming evolution terms before
> they are permanently frozen.
>
> I'll keep this mostly short with a list of current names and suggested
> replacements but I hope we can finalize this at the next sprint. The
> following suggestions are from an IRC a little while ago.
>
> Please chime in if you have a better naming scheme or suggestion.
>
> == Troubles ==
>
> The worst offender is 'troubles'. It is kind of a like a merge conflict,
> in my opinion.
>
> Suggestions: inapplicable, conflicted, invalid, unevolved, dirty

I like 'unevolved', but given that there's an evolve command, I think the  
naive user will naturally reach for that to fix all csets in this state.   
And I'm not sure that's possible with divergent and bumped.

I guess 'conflicts' or 'conflicted' would be my second choice.  (Would it  
be confusing to overload  a term already associated with merges?  I never  
really thought of these as being 'like a merge conflict', but I guess I  
can see similarities.)

> == Bumped ==
>
> Does not in the slightest convey what is going on.
>
> Suggestions: invalidated, tardy, stale, behind, superseded, lagging,  
> obviated

The metaphor here is "passenger on a plane gets bumped from his seat".   
Knowing that, I can kinda see what the help is saying.  But I probably  
wouldn't understand the significance of what the help is saying if I had  
missed that discussion on the mailing list.  Not sure how to improve the  
help (I was looking at revsets.bumped()).  For me, I wouldn't associate  
the scenario in 'bumped' with any of the suggestions, but don't have any  
to add either.

> == Unstable ==
>
> This one seems fine to me as-is. But a few people suggested these:

Me too.

> Suggestions: unsettled,uprooted
>
> == Divergent ==
>
> Not too bad as-is but duplicated seems better.
>
> Suggestions: duplicated

If I see 'duplicated', I think carbon copy.  But most of the time I've hit  
this, there was something different between the related csets.  Divergent  
seems good to me.

> == Successor ==
>
> Suggestions: replacement, next
>
> == Precursor ==
>
> If we keep 'successor' then we have to change this one to 'predecessor'.
>
> Suggestions: original, predecessor, progenitor, previous, prior,  
> antecedent

I don't feel strongly about these.  Shorter would be preferable, and I can  
see misspelling 'predecessor' or 'progenitor' when checking to make sure I  
didn't lose something in a rebase conflict resolution with:

     $ hg extdiff --patch --hidden -r 'precursors(.)' -r .

The other cautionary note is that there is already an 'origin' revset,  
which finds the original graft/rebase/transplant.  Not sure how closely we  
want to bring unrelated concepts (i.e. use 'original' here).

> _______________________________________________
> Mercurial-devel mailing list
> Mercurial-devel at mercurial-scm.org
> https://www.mercurial-scm.org/mailman/listinfo/mercurial-devel


More information about the Mercurial-devel mailing list