revset: make use of natively-computed set for 'draft()' and 'secret()'

If the computation of a set for each phase (done in C) is available,
we use it directly instead of applying a simple filter. This give a
massive speed-up in the vast majority of cases.

On my mercurial repo with about 15000 out of 40000 draft changesets:

revset: draft()
   plain         min           first         last
0) 0.011201      0.019950      0.009844      0.000074
1) 0.000284   2% 0.000312   1% 0.000314   3% 0.000315 x4.3

Bad performance for "last" come from the handling of the 15000 elements set
(memory allocation, filtering hidden changesets (99% of it) etc. compared to
applying the filter only on a handfuld of revisions (the first draft changesets
being close of tip).

This is not seen as an issue since:

* Timing is still pretty good and in line with all the other one,
* Current user of Vanilla Mercurial will not have 1/3 of their repo draft,

This bad effect disappears when phase's set is smaller. (about 200 secrets):

revset: secret()
   plain         min           first         last
0) 0.011181      0.022228      0.010851      0.000452
1) 0.000058   0% 0.000084   0% 0.000087   0% 0.000087  19%

contrib/all-revsets.txt

@@ -124,3 +124,7 b' head()'
 head() - public()
 draft() and head()
 head() and author("mpm")
+
+# testing the mutable phases set
+draft()
+secret()

mercurial/revset.py

@@ -1464,6 +1464,13 b' def parents(repo, subset, x):'
 
 def _phase(repo, subset, target):
     """helper to select all rev in phase <target>"""
+    repo._phasecache.loadphaserevs(repo) # ensure phase's sets are loaded
+    if repo._phasecache._phasesets:
+        s = repo._phasecache._phasesets[target] - repo.changelog.filteredrevs
+        s = baseset(s)
+        s.sort() # set are non ordered, so we enforce ascending
+        return subset & s
+    else:
     phase = repo._phasecache.phase
     condition = lambda r: phase(repo, r) == target
     return subset.filter(condition, cache=False)