On Sat, Feb 25, 2017 at 08:56:33AM +0100, b...@qqmail.nl wrote: > That there is a collision now doesn’t change that we always assumed > there would be collisions, and designed the current behavior with that > in mind.
Yes, that is right. My line of thinking there was not meant to be the one we use for an immediate fix. It must be pursued only if we decide that SVN should be able to store files with SHA1 collisions as distinct content. Regardless of whether we change hashes or not, our immediate attention is best spent on fixing the behaviour of the existing system when faced with SHA1 collisions. Obviously there are several problems throughout the code base that need to be addressed even before we even change formats etc.