John Beck <jbeck at eng.sun.com> writes: > Gary> 1) There seems to be no equivalent to wx/active where I can define only > Gary> those files I wish to be active and putback comments for each file. > > Gary> I almost always have multiple different things with multiple different > Gary> files going on in my TeamWare workspaces. I manage them multiple > Gary> wx/active files for webrev and putback from a single TeamWare workspace. > > Gary> How do I get the same functionality from Hg+Cd? > > You don't, at least for now. Use multiple workspaces and keep your work > separate.
That's not an "at least for now" thing, that's the way things are. Commits are repository wide, so once your changes are committed they are there, there is no way to push only a subset of the files in a given commit (as you could putback only a subset of altered files in TeamWare, given it only cares about the SCCS delta). Given that limitation, there's little utility in attempting to allow people to work on subsets of their total change. > > Gary> 4) If I reference the parent via ssh I'm prompted for my id_dsa > Gary> passphrase. If I ssh to a SWAN machine I'm prompted for my id_rsa > Gary> passphrase. > > Gary> Why the difference? > > Don't know; someone else (Dave? Mark?) might. > Gary> How do I stop having to give my id_dsa passphrase for onnv.eng? > > My only suggestion is to generate a key with no pass-phrase, but others > might have better ideas. ssh-agent is certainly a better approach than a password-less key, for interactive purposes. > Gary> 5) hg pbchk -p ssh://onhg at onnv.eng//export/onnv-gate > Gary> succeeds happily, but the immediately following hg push -p > Gary> ssh://onhg at onnv.eng//export/onnv-gate fails. > > Gary> Is there some deficiency in pbchk? > > No, that's not part of what it's checking. Push doesn't take -p, so, at the very least, it fails because that's a syntax error. If the gate sanity hooks find errors that pbchk does not, I would probably consider that a bug, however. Further than that, I'd need to know more about what "fails" suggests (how does it fail?) > Gary> 6) hg push -p ssh://onhg at onnv.eng//export/onnv-gate fails complaining > Gary> about files I'm not modifying. > > Gary> Why is push deficient in its updates? > > It isn't; that's by design AFAIK. I don't entirely understand that question. If you're meaning that it fails saying you would add multiple heads, John's answer is correct. If you are saying that the hooks on the gate that check for cstyle, and similar, are checking files you don't feel you have modified, that's a separate problem and I would need more information (as, no doubt, would gatekeeper@). > Gary> 8) So far, I've not had a conflict (only been doing Hg for a week), but > Gary> merges seem to be required to happen automagically. I understand this > Gary> is required for the convenience of the conflicting push issue I noted in > Gary> 6). I'm happy to have all non-conflicting files automatically merged > Gary> (just like when I did a bringover). I never what any of my "active" > Gary> files to ever be merged by any tool other than my fingers using $EDITOR. > > Gary> The way I did this with TeamWare by making a copy of any files in > Gary> conflict, running resolve and always accepting parent, then reediting > Gary> and merging again. Finally redelget. > > Gary> How do I affect the same work flow in Hg+Cd? > > I don't know if you can. There is an outstanding request from someone else to adjust hgsetup(1) such that the hgrc it creates prevents Hg attempting to automerge any file it views as in conflict. I'm quite happy to do this, but I had thought that there were those as vocally in favor of auto-merge as some are against. -- Rich
