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

RE: regarding tree-conflicts work

From: Kamesh Jayachandran <kamesh_at_collab.net>
Date: Fri, 16 May 2008 23:10:55 +0530

>Kamesh wrote:
>> I was going through the tree-conflicts work to see if I can make any
>> contribution there.

>Hi Kamesh. It would be great to have you helping.

>Is there any reason not to move this conversation to the public dev list? It might help to awaken some interest from other people. If you agree, >I'll copy this reply to dev_at_s.t.o.

Yes I CCed dev_at_s.t.o now.

>> Do you maintain kind of TODO file somewhere?

>As Stefan mentioned, there are issues in the tracker. I converted the tagging from using "tree-conflicts" in the URL field to using >"tree-conflicts" in the Keywords field.

>At the top of my personal to-do list are:

> * Fix the failing non-tree-conflict tests on the tree-conflicts branch.

> * [Main task] Directory conflicts. In parallel, I want to both "fix" the most important individual cases that aren't yet handled, and also design a cleaner way to handle all the cases in some consistent manner, perhaps using a "conflict resolver callback" idea.

> * Resolve the problem that prevents the "double-delete" branch being completed - Issue #3156.

>And then some other tasks:

> * Investigate Alexander Kitaev's "1.5.x bug: property that have to be deleted by update remains in the working copy after update"

> * Issue #1962 "merge of non-empty subdir committed incorrectly". Needs design. Eric H had some thoughts, a test, and a reverted fix attempt.

> * "svn resolve" should force you to specify which conflict in a directory you are resolving, and not just resolve all of them.

> * Merge into conflicted directories - how to allow it for a bit of manual merging that may be needed in order to resolve conflicts within a >larger merge.

> * Re-vamp the use cases: state the minimum acceptable behaviour; state all the cases that we plan to make work (esp. directory clashes). We >should get Phillips to correct/clarify/update their own use cases, and we should write our own ones too.

> * Turn some of the use cases into acceptance tests.

>If you want to tackle any of these, that would be great.

>> I could see some tests failing, I am yet to go closer to fix
>> them(checkout_tests-7 for instance).

>Thanks. This would be very useful. I started to look at them but the first two or three I looked at were more difficult than I expected. Some of >them probably are broken due to bugs in the tree conflicts code (perhaps including accidentally interfering with non-conflict cases), and others >will need updating to expect a "conflict" result where they didn't expect it before.

I will focus on making the testsuite pass and pickup few items from the above list.

With regards
Kamesh Jayachandran
Received on 2008-05-16 19:44:38 CEST

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

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.