[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: [PATCH][merge-tracking] Fix for failing testcase merge-tests.py:35

From: Daniel Rall <dlr_at_collab.net>
Date: 2006-09-18 04:04:02 CEST

On Tue, 12 Sep 2006, Kamesh Jayachandran wrote:
...
> 1. Should we record the mergeinfo in case of conflicting merges?

In merge-tracking/TODO, we said the following:

    * If a conflict is encountered, invoke a conflict resolution
      callback to give a Subversion client a chance to intervene. If
      resolution is successful, convert the notification from a 'C' to
      something else (e.g. 'M'). (Phase 2?)

    * Otherwise, stop applying merge ranges as soon as a second
      conflict is encountered in a WC item (as it might generate
      overlapping conflict makers, or apply a merge inside a conflict
      marker!), being sure to record partial application of merge
      ranges.

So for merge test #35, we should record merge info the conflict, as
only one occurs in a WC item (as per your patch).

> 2. Shouldn't we check for historical relationships between source
> and destination before recording merge info?

No. Subversion has never required a historical relationship to exist
between merge source and target -- merging two pieces of unrelated
text together is perfectly fine. I don't see any reason to adopt this
requirement for Merge Tracking.

I've tweaked the formatting of your patch and committed it as r21525.
Thanks Kamesh!

- Dan

  • application/pgp-signature attachment: stored
Received on Mon Sep 18 04:05:19 2006

This is an archived mail posted to the Subversion Dev mailing list.