Re: r27729 breaks API of svn_ra_get_log2

David Glasser wrote:
>>> I haven't seen this in real life yet, but I believe r27737 introduced an
>>> inconsistency with the API spec.
>>
> Actually, looks like it was r27729

The subject was right all along, but while writing the body I must have
copied from the wrong tab. Sorry for any confusion I caused.

> I think putting the whole loop in an if would be more clear.

Slightly better for performance as well, but I guess that doesn't count
much here.

By now I encountered this in real live as well, and the first suggested
fix works as expected. https://bugs.launchpad.net/bzr-svn/+bug/234010

Although I'm an outsider, I'd say the risk of breaking anything due to
this fix is considerably lower than breakage due to the bug, so I'd vote
to place it into 1.5.0.

Looking at the email that just arrived, it seems to be decided that way.

Greetings,
  Martin von Gagern