Re: Ad-hoc survey of existing Debian git integration tools

[Ian Jackson <ijackson@chiark.greenend.org.uk>]

Vincent Bernat writes ("Re: Ad-hoc survey of existing Debian git integration tools"):
> On the other hand, since the previous workflow was quilt, still using
> quilt is useful for someone who doesn't want to learn about specific git
> workflows.

That depends on who you are.

For a maintainer who is used to using quilt, and who does not touch
other packages, then maybe yes, it is easier.  Just as a maintainer
who doesn't want to learn git - they should probably ignore dgit.

But for anyone with a wider focus, who finds git useful, dgit is the
best way to (a) get a git view of /any/ package, no matter the
maintainer's preferences (b) share that git view with others (whether
via Debian proper, or via a side channel, or elsewhere).

> Personally, I am waiting for the number of git-based workflows to
> decrease before settling for something else.

I don't think the number of git-based workflows is going to decrease.
I'm hoping (betting!) that the proportion of packages whose git
maintainers' git workflows involve strange[1] git trees will go down.

[1] Strange means any tree which you can't work on, immediately, with
git, without using some special tool on it first.

Ian.