[PATCH 3 of 5 flagprocessor v6] revlog: pass revlog flags to addrevision

Rémi Chaintron remi.chaintron at gmail.com
Sat Dec 31 06:47:31 EST 2016


I'm convinced already :)
On Sat, 31 Dec 2016 at 06:46, Pierre-Yves David <
pierre-yves.david at ens-lyon.org> wrote:

>
>
> On 12/30/2016 05:38 PM, Rémi Chaintron wrote:
> > You're right, this is a mistake on my part. The reasoning behind
> > dropping this code is that in my current implementation, processflags()
> > handles checking whether the flags are known (both for revision() and
> > _addrevision()) so I moved this snippet to processflags().
> > Still makes sense to not spread the refactor across two patches, will
> > update.
>
> The issue with not grouping this is that this check is effectively gone
> between the commit that remove it and the commit that adds it back. This
> means we no longer have a "Code is correct at all commit" property. This
> can prevent us to take the first part of the series without the second
> part (which might need rework) and this make potential backout (if we
> find a regression for example) much more harder for use since we have do
> more extensive checks of potential other commit entangled with the one
> we would backout.
>
> Cheers,
>
> --
> Pierre-Yves David
>
-- 
Rémi
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.mercurial-scm.org/pipermail/mercurial-devel/attachments/20161231/ea831443/attachment.html>


More information about the Mercurial-devel mailing list