On 11/14/13 2:01 AM, Philip Martin wrote: > Looking at that again, the code doesn't rely on repos_basename being > valid when there is an error but it is checking for non-NULL. So it's > still a bit odd that we are checking repos_basename when there is an > error as it means this code relies on dav_svn_split_uri not setting > repos_basename on error. > > I still think the test of repos_basename should be removed. If we need > to test something then look for HTTP_FORBIDDEN in err.
Thanks a lot for the review. After talking with stefan2 on IRC I came up with a much better way to deal with this. I myself wasn't thrilled with what I had done.
