On Thu, Apr 30, 2015 at 8:02 PM, Daniel Shahaf <d.s@daniel.shahaf.name> wrote:
Dan Ellis wrote on Thu, Apr 30, 2015 at 14:52:23 -0700:
> Here's a snippet of the running scenario that can NOT reproduce the issue
> in a new repo:

OK, we are pretty confident we have narrowed the behavior down.  Unfortunately, the issue is somewhat (very) embarrassing to report. 

Our automation tools had been updated to monitor file system changes and one of the semi-recent changes had been to clear properties when it detects an SVN move (which catches SVN renames).  As you can guess, the randomness and non-reproducibility was induced by the timing of the file system watcher callbacks and our understanding of what occurred during these watches.

I appreciate everyone's time and apologize for the wild goose chase.

How I wish I could delete this thread.