cleanupnodes: preserve phase of parents of new nodes

As Yuya noted in the review of D3818, passing in
targetphase=phases.draft would result in advancing the phase boundary
of a secret-phase parent. We never pass targetphase=phases.draft so
far, but it's a bug waiting to happen.

I tried to refactor it so max(parentphase, X) happened in one place,
but I couldn't come up with good variables names and I ended up with a
"newphase = max(newphase, parentphase)" line, which made the whole
block not look any better to me.

Differential Revision: https://phab.mercurial-scm.org/D3824

mercurial/scmutil.py

@@ -842,13 +842,13 b' def cleanupnodes(repo, replacements, ope'
             return newphases.get(ctx.node(), ctx.phase())
         for newnode in allnewnodes:
             ctx = unfi[newnode]
+            parentphase = max(phase(p) for p in ctx.parents())
             if targetphase is None:
                 oldphase = max(unfi[oldnode].phase()
                                for oldnode in precursors[newnode])
-                parentphase = max(phase(p) for p in ctx.parents())
                 newphase = max(oldphase, parentphase)
             else:
-                newphase = targetphase
+                newphase = max(targetphase, parentphase)
             newphases[newnode] = newphase
             if newphase > ctx.phase():
                 toretract.setdefault(newphase, []).append(newnode)