[PATCH 4 of 8 STABLE] pager: rename 'pager.enable' to 'ui.pager'

Augie Fackler raf at durin42.com
Mon May 1 21:18:42 EDT 2017


On Tue, May 02, 2017 at 12:11:28AM +0200, Pierre-Yves David wrote:
> On 05/01/2017 10:23 PM, Kevin Bullock wrote:
> > > On May 1, 2017, at 11:12, Pierre-Yves David <pierre-yves.david at ens-lyon.org> wrote:
> > >
> > This turns out to be a bit of a can of worms that Augie, Pierre-Yves, Jun, Sean, I, and others have been discussing via VC and IRC today. To summarize:
> >
> > * pager.enable is inconsistent with ui.color=yes/no/auto/always
> >
> >   * ...but ui.pager is confusing in comparison to ui.editor and pager.pager
> >
> > * We've already released pager.enable in 4.2-rc, and while that isn't subject to the normal BC guarantees, it would suck to change the name of the config setting again for our most active testers (i.e. those who bothered to install the RC)
>
> To summarize my main concerns:
>
> * 4.2 introduces a large behavior change by adding a pager by default. So
> their will be a reasonable share of users who will seek to disable (back)
> that pager in the coming months.
>
>   So we should ensure this 4.2 release is smooth on that regard and getting
> the config option in its definitive spot is part of this smoothing process.
>
> * Even after the migration, this will be an important config knob that
> people will reach out to.
>
>   This is my main motivation to have it in the main '[ui]' config next to
> the similar 'ui.color' knob. (the exact form is not too important).

I agree that on the one hand it'd be nice to have ui.color and
ui.pager next to each other, but I feel it's too confusing to have
ui.editor and ui.pager be wildly different (one is a program to
invoke, the other a boolean).

I'd have suggested moving ui.color to color.enabled but the free-form
nature of the existing [color] namespace means that's probably
unwise. It's my opinion, after reflecting on this for most of today,
that the existing names are probably the path of least confusion overall.


More information about the Mercurial-devel mailing list