On Wed, Jan 6, 2010 at 10:48 AM, Julian Foad <julianf...@btopenworld.com> wrote:
> My concern is whether the new rules are just papering over a crack or > whether they correctly take into account all the ways in which a client > might depend on accessing the newly created revision. My understanding is that these commits fail, so I do not think it is about accessing the new revision at all. I think it has something to do with establishing the baseline revision for the commit and picking an out of date revision from the proxy and then having the master reject it. It goes back to this I think: > * The following operations are handled by the proxy if referring to a > cached revision, else passed through to the master: > PROPFIND GET ... Apparently, the PROPFIND is using the proxy at a time when it needs to be using the master. I think your general question stands, although this is only ever going to be an issue when we are talking about some kind of write operation, so it ought to be fairly easy to know where we have a problem. -- Thanks Mark Phippard http://markphip.blogspot.com/
