Hello,
On Mon, Aug 01, 2016 at 10:42:53AM +0100, Ian Jackson wrote:
> Sean Whitton writes ("Bug#833025: dgit: quiltify smash strategy shouldn't
> fail when current patch queue fails to apply"):
> > I am the maintainer, but I think that there are use cases for a dgit
> > quilt strategy that is essentially a one-shot single-debian-patch.
> >
> > Suppose that the package was team-maintained, such that introducing
> > single-debian-patch would be disruptive -- perhaps because other team
> > members expect to introduce new, separate patches in the future.
> > However, for now there is only one patch, so a one-shot
> > single-debian-patch is the easiest way to fix the quilt series after
> > merging a new upstream version.
>
> The real problem here is surely that it is "too hard" to work on the
> patch series in git.The elephant in the room! > Since dgit needs to gateway between git and dpkg-source, it inevitably > ends up containing some pieces of a general version control system. > But I'm trying to avoid turning it into a general tool for patch stack > manipulation... I agree with the general thought. Are you saying you wouldn't accept a patch adding the smash-all strategy? > I do intend for dgit to, when generating synthetic history from a > non-dgit .dsc, produce a sort of patch stack representation. That sounds cool. -- Sean Whitton
signature.asc
Description: PGP signature
