<hr /><i>anonymous claiming to be Joerg Sonnenberger added on 2010-10-15 14:03:15:</i><br />
There are two potential questions related to this:

(1) Should fossil always scan for the markers on files with modified mtime. This works for conflicts from manual patches too. It has the strong disadvantage of requiring reading potentially large files to determine the conflict. Such a check should at least be done pre-commit, I think.

(2) Should conflict resolution be implicit or explicit? I'm not a fan of the Mercurial approach of explicit "hg resolve", but it has its merits too. At the very least, it would make the status operation very cheap and force people to more explicit think about it.