Re: Inherited properties document

Molle Bestefich wrote:
> Greg Hudson wrote:
>
>>I don't really object to inherited properties; I just have some doubts
>>that we can get to the point of having a design which (a) we can
>>implement, and (b) doesn't commonly bite users with edge cases.
>
>
> Just out of curiosity.
> What exactly are the edge cases wrt. (b) that you are thinking about?

The file that I posted had a couple of the more obvious ones (read 5a
and 5b).

Specifically, if the inheritance is by path in the repository and you
copy a subtree to another place in the repository (typically as either a
branch or a tag) what happens to any properties which were inherited
(from above the copied portion) at the old location that no longer exist
in the new location. In this case, you can either freeze the existing
property values (which works well for tags but not branches) or you can
ignore existing inheritance (which would in general work for branches
but not tags). Either way, someone is going to be disappointed; there
are other reasons why tag vs. copy may need to eventually be two
different [related] operations (if only with syntactic sugar).

The other one that I've already thought of is when you get to the point
of, for example, using an inherited property to expand as a custom
keyword, when you copy the file to a location that no longer has that
property assigned, the file is now "broken" in the sense that the
keyword will no longer expand.

John

-- 
John Peacock
Director of Information Research and Technology
Rowman & Littlefield Publishing Group
4501 Forbes Boulevard
Suite H
Lanham, MD  20706
301-459-3366 x.5010
fax 301-429-5748
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org