push: move local lock logic in pushoperation

During push, we try to lock the local repo to move local phase according to what
we saw/pushed on the remote repo. Locking the repo may fail, in that case we let
the push proceed without local phase movement (printing warning).

This mean we have code in phase synchronisation that will check if the local repo is
locked or not. we need to move this information in the push object to be able to
extract the phase synchronisation in its own function. This is done as a boolean
because putting reference to the actual lock outside of the main function
sounded a bad idea.

mercurial/exchange.py

@@ -33,6 +33,8 b' class pushoperation(object):'
         self.revs = revs
         # allow push of new branch
         self.newbranch = newbranch
+        # did a local lock get acquired?
+        self.locallocked = None
 
 def push(repo, remote, force=False, revs=None, newbranch=False):
     '''Push outgoing changesets (limited by revs) from a local
@@ -66,7 +68,7 b' def push(repo, remote, force=False, revs'
     unfi = pushop.repo.unfiltered()
     def localphasemove(nodes, phase=phases.public):
         """move <nodes> to <phase> in the local source repo"""
-        if locallock is not None:
+        if pushop.locallocked:
             phases.advanceboundary(pushop.repo, phase, nodes)
         else:
             # repo is not locked, do not change any phases!
@@ -81,7 +83,9 b' def push(repo, remote, force=False, revs'
     locallock = None
     try:
         locallock = pushop.repo.lock()
+        pushop.locallocked = True
     except IOError, err:
+        pushop.locallocked = False
         if err.errno != errno.EACCES:
             raise
         # source repo cannot be locked.