Conflict storage

["Hyrum K. Wright" <hy...@mail.utexas.edu>]

As I understand it, the conflict storage handling updates are one of
the blocking items for 1.7[1].  In an effort to better understand the
problem, and the proposed solution, I'm writing this mail with the
intent to start some discussion.

The content of notes/wc-ng/conflict-storage seem to describe the end
result, but doesn't do very much to describe how to move between the
current implementation to the later one.  Additionally, that document
still has many pending questions.

So, for somebody eager to start hacking that part of the code, how
would $EXPERT suggest I go about doing so?  What bits are the
low-hanging fruit?  And where are the gotchas, where things are a bit
murky?

-Hyrum

[1] I don't fully understand *why* this is true, just that it is.  My
guess is that people have thought "hey, while we're changing
everything in the working copy, let's change this, too."  If that is
the sole reason for making 1.7 block on the conflict storage work,
then I would submit that in that way lies madness, and we need to
seriously rethink this dependency.

Re: Conflict storage

[Greg Stein <gs...@gmail.com>]

On Mon, Oct 25, 2010 at 16:57, Stefan Sperling <st...@elego.de> wrote:
> On Mon, Oct 25, 2010 at 03:42:16PM -0500, Hyrum K. Wright wrote:
>> [1] I don't fully understand *why* this is true, just that it is.  My
>> guess is that people have thought "hey, while we're changing
>> everything in the working copy, let's change this, too."  If that is
>> the sole reason for making 1.7 block on the conflict storage work,
>> then I would submit that in that way lies madness, and we need to
>> seriously rethink this dependency.
>
> See http://svn.haxx.se/dev/archive-2010-08/0555.shtml
> (Note that Bert is replying to himself, i.e. the first-level quote and
> the normal text was written by him.)

Well... Bert is talking about recording the filenames of the conflict
files. That is different from the more general "recording conflicts"
problem.

IOW, we could record the svnpatch reject file in the "conflicts" skel,
but not worry about migrating the directory-level tree conflict
information to the new format.

I'd prefer to see a full migration, but figured the point needs to be
brought up.

Cheers,
-g

Re: Conflict storage

[Stefan Sperling <st...@elego.de>]

On Mon, Oct 25, 2010 at 03:42:16PM -0500, Hyrum K. Wright wrote:
> [1] I don't fully understand *why* this is true, just that it is.  My
> guess is that people have thought "hey, while we're changing
> everything in the working copy, let's change this, too."  If that is
> the sole reason for making 1.7 block on the conflict storage work,
> then I would submit that in that way lies madness, and we need to
> seriously rethink this dependency.

See http://svn.haxx.se/dev/archive-2010-08/0555.shtml
(Note that Bert is replying to himself, i.e. the first-level quote and
the normal text was written by him.)