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

Re: Code reformatting

On Wed, Mar 12, 2008 at 10:38:16PM -0700, Steve Langasek wrote:
> On Thu, Mar 13, 2008 at 02:19:14PM +1000, Anthony Towns wrote:
> > On Tue, Mar 11, 2008 at 04:14:45PM -0700, Steve Langasek wrote:
> > > Because this is being done *in lieu of* merging the triggers branch, with
> > > the result that the triggers branch becomes harder to merge afterwards.
> > The triggers branch is already difficult to merge because it has numerous
> > unrelated changes.
> That's a fair criticism, but I don't think it changes the fact that making
> style changes while there's a major branch merge outstanding is precisely

I don't believe there's any intention to automatically merge the branch,
but instead to reimplement/port some of its features (including triggers)
into the main branch by hand.

> > From at least August last year, the triggers branch has had significant
> > stylistic changes compared to the dpkg codebase.
> If this means style changes to existing code, I agree that this is bad.  But
> two wrongs don't make a right.

I believe it means introducing new code that doesn't conform with the
style guidelines, which in turn doesn't conform with the dpkg team's
patch acceptance guidelines. My impression from other mails is that's
not a complete summary of the style changes in Ian's tree.

> But a "preliminary injunction" would still be an
> appropriate response to actions that are actively harming the merge of a
> feature that's important to Debian, would it not?

I don't think so, no. Trying to control how people use VCS repos to
prepare their patches seems entirely out of our remit; and if the dpkg
team decided to point dpkg.org at their own machines instead of using
alioth, entirely out of our control.

Of course, precedent would be for us to simply say "go away, code style
isn't a technical issue".

> > http://lists.debian.org/debian-dpkg/2007/08/msg00014.html
> It's not clear to me that these are stylistic *changes*, which to me implies
> that the style of existing code has been altered,

The package changelog Ian used closed Bug#375711, which includes
formatting changes. Ian's patch for that bug was prior to triggers,
and my impression from the discussions was that it was included in the
triggers tree from the word go. If there was a branch that applied to
the dpkg git tree as of 17th August 2007 or 10th March 2007, without
spurious changes, that'd be one thing, but without having gone through
the entire git history over that period of both trees, everything I've
seen indicates there's not.


Attachment: signature.asc
Description: Digital signature

Reply to: