Mike Kupfer wrote: >>>>>> "Rich" == Richard Lowe <richlowe at richlowe.net> writes: > > Mike> I just filed issue 576 against Mercurial for this. AFAICT, the only > Mike> issue is the scary warning message when the child is updated. Can > Mike> you clarify what you meant by "a scarier (but apparently ignored) > Mike> dirstate"? > > Rich> In my experience (and that of another opensolaris person) the > Rich> dirstate represents the not-really-versioned files from that > Rich> directory as 'modified', when clearly they should be unversioned. > Rich> It's not clear to me whether if committed, they would become > Rich> versioned. > > Rich> That aspect is why I consider it a stopper (or close to it), > Rich> webrev will show .make.state's for instance, and, as I said, it's > Rich> possible commit would cause them to be versioned. > > At least with my trivial test case, that doesn't seem to be a problem. > "hg stat" shows no changed files, "hg manifest" only shows the file that > was already versioned, and "hg commit" says "nothing changed".
What version of Mercurial, out of interest? I've seen the behaviour described above after bringing over any directory move (to include removal, given deleted_files), after a workspace is built. -- Rich
