D4813: narrow: the first version of narrow_widen wireprotocol command

[martinvonz (Martin von Zweigbergk)]

martinvonz added a comment.


  > I need to specify the arguments instead of just specifying "*".
  
  Don't forget this one.

INLINE COMMENTS

> pulkit wrote in narrowcommands.py:293-295
> We can have a 'exp-ellipses-2' which will tell whether the server supports ellipses widening using narrow_widen() wireprotocol command or not. I think that should help in the meantime. Also will a week, or 10-15 days be enough for you? I think it will be better if can prevent releasing this compatibility because exp-ellipses-1 was introduced in this cycle only.

We don't really care whether a version is released or not. It would be nice to have a version of the client that would use the new wire protocol with ellipses if the server said it supported that but would otherwise fall back to the old getbundle-based call.

> pulkit wrote in narrowwirepeer.py:49
> I implemented the peer initially in core only, but while implementing server side, I realized it rely on logic in narrowbundle2.py which also needs to be moved to core. Then I decided to implement it cleanly in the extension and then move it to core.

It seems to depend only on `widen_bundle`, which doesn't seem to depend on anything else, so it would probably be easy to move it to core, but I won't insist.

> pulkit wrote in narrowwirepeer.py:79-82
> I am not sure, I just copied from getbundle() handling: https://www.mercurial-scm.org/repo/hg/file/1a4c1a3cc3f5/mercurial/wireprotov1server.py#l404

That code was added in https://www.mercurial-scm.org/repo/hg/rev/3e7f675628ad. Maybe @indygreg can tell us if he thinks we should support the same values for a new wireprotocol command.

REPOSITORY
  rHG Mercurial

REVISION DETAIL
  https://phab.mercurial-scm.org/D4813

To: pulkit, durin42, #hg-reviewers, martinvonz
Cc: indygreg, mercurial-devel