Markus Schaber <m.scha...@3s-software.com> writes: > To be honest, after thinking twice, I'm not sure whether my change > actually changes user-visible behaviour for non-binary files. My > assumption was that merge_text_file() call following the > merge_file_trivial() eventually produces the same result in that case, > just using a different path to arrive there.
That depends on the text merge algorithm. At present svn up --diff3-cmd /bin/false would cause a conflict but with your patch the merge code is not invoked so the "merge" succeeds. Does that matter? I don't know. One thing that is still not quite right is that the final state of the file is unmodified but NODES.last_mod_time is not set. It's not caused by your patch but ideally we would update last_mod_time. -- uberSVN: Apache Subversion Made Easy http://www.uberSVN.com
