Re: svn_fs_merge not used?

Branko Čibej <brane@xbc.nu> writes:

> Don't we already have that? I distinctly remember we had a real rename
> operation in the FS at one point. Um. It's possible that we threw it
> out, though.

No, we had a partially finished "real" rename operation that
SEGFAULTed in simple test cases and allowed us to violate our primary
key uniqueness given the right opportunity. That code exists in the
fs-enhancements branch as a known-imperfect, proposed implementation
of svn_fs_rename(), but until a better algorithm comes along, that
code will die on that branch.

Lately I've been tossing a few cycles here and there at the Right
implementation, but quite frankly, this problem is so post-1.0 (and
I'm not yet convinced that *ever* needs to be solved) that even those
few cycles are being misused.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sat Oct 18 20:25:29 2003