On Friday 07 May 2010 07:34:23 pm Martin Geisler wrote:

I am not sure this is the direction I would like to see happen. I understand that "closed history" can cause some cluttering, but I think not sending some changesets on push/pull is a disruptive behavior. Do we already do that in some situation, except explicitly requested through a "-r" option. And after all, closed history is history, so why not push it anyway ?

I am more concerned the close-branch behavior itself. May be I am stupid, but I have hard time to understand the logic behind the current implementation. Actually, I would really like to use that feature, but never did it because I am confused how all the commands reacts. And I must not be alone as I remember several questions on that topic in the mailing list.

Setting aside compatibility issues with the CLI, repo format, perf issues, my ideal requirements would be: 1/ It shall be possible to close a head, no distinction between branch head or topological heads (too confusing for newbie IMO) 2/ Trying to close a changeset that is not a head (either branch head or topological) shall raise an error. Why would we allow that ? 3/ the "heads" command shall not show a changeset that is closed, whether it's a branch or a topo head, except if the --closed option is set. 4/ The "branches" commands shall not show a branch if all its branch heads are closed, except if the --closed option is set. 5/ Closed heads shall not be considered for default/implicit head selection of the update or merge commands 6/ The user shall be warned when committing from a closed parent head. Either: 6a/ This should raise a error, except if a --force option to commit is set. May be too extreme solution I guess. 6b/ When committing over a closed head parent, the commit command shall display a message: "Resurrecting closed head" 7/ Closed changesets shall be marked more explicitly in showed changesets, i.e. for commands like log, incoming, ..., and may be also in identify and summary.

With that, I would be satisfied if push/pull stays the same. Are those requirements are: - already the case - reasonable - achievable

Regards. Gilles.