Re: [RFC] Move tracking - design summary and mock-up proposal
From: Julian Foad <julianfoad_at_btopenworld.com>
Date: Wed, 21 May 2014 13:14:18 +0100 (BST)
Julian Foad wrote:
I have written lots more about various parts of move tracking at
http://www.foad.me.uk/tmp/svn/moves/
with an emphasis on trying to explain why it needs to work this way in terms of the theoretical basis. My writing there is a bit scattered and out of date, but you might nevertheless find some of it interesting and informative.
The docs listed first are the most readable, coherent and up to date, and are supposed to make sense top-down in the order given. Don't put too much weight on their titles.
rationale
One out-of-date concept in the docs there is that I was talking about the definition of a branches and branch families being mutable: that any directory (or file) could become a branch root element on demand, which led to problems of how or whether a branch could later be converted back to a non-branch, and how to decide whether it currently is or isn't, and so on. From discussions with the others I've changed my mind: a directory (or file) either is or isn't a branch root element, immutably for its whole lifetime.
- Julian
|
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.