[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Triggers status?

On Wed, Oct 24, 2007 at 01:27:22PM -0400, Joey Hess wrote:
> Raphael Hertzog wrote:
> > I just gave a quick look to your branch:

> > - please rebase it on the current master branch (that way you're sure that
> >   there are no conflicts)

> I have yet to see a use of git rebase that I am comfortable with.
> History is very important to me, even the ugly bits. Being able to see
> every misttep that was committed actually helps understanding the end
> result.

Loosing the missteps like that is one of the explicit goals of doing
this that Linus pushes for rebasing.  The other thing with it
(especailly with -i) is that it supports a workflow where patches are
submitted using git-format-patch and git-send-email and reviews are
iterated - this gives a patch per commit, allowing piecemeal review, but
means that you need to edit history in order to do the iterations.

> Merging it with the current upstream master would have the same effect
> with a more "ugly" history, surely?

Yes.  As I understand it what's going on with the kernel is that it's
aiming for a revision control history that looks a lot like things being
submitted and reviewed as patches even where that's not technically what
happens in all the stages the patch flows through.

> > - you can also restructure the serie of changes with "git rebase -i origin/master"
> >   (provided that origin/master is an up-to-date copy of the master branch
> >   of the main dpkg repository)
> >   (merge several commits in a single, in particular bug fixes with initial
> >   implementation so that while reviewing diff you don't find bugs which
> >   are in fact fixed by a subsequent commit)

> This loses history, is highly complex, requires a significant knowledge
> of git (more than I am comfortable with after using it for a month), and
> AFAICS is an anourmous time sink for no benefit. I'd rather be coding.

It can be useful if used with a workflow that it supports (I've used it
with ones where individual commits get a fairly large amount of review)
but I'm personally less sure about it when it's just the head of a
branch that's getting reviewed.

"You grabbed my hand and we fell into it, like a daydream - or a fever."

Attachment: signature.asc
Description: Digital signature

Reply to: