mercurial-scm/hg: mercurial/hbisect.py comparison

comparison mercurial/hbisect.py @ 15146:b39d85be78a8

hbisect.get: use simpler code with repo.set(), fix 'pruned' set Use repo.set() wherever possible, instead of locally trying to reproduce complex graph computations. 'pruned' now means 'all csets that will no longer be visited by the bisection'. The change is done is this very patch instead of its own dedicated one becasue the code changes all over the place, and the previous 'pruned' code was totally rewritten by the cleanup, so it was easier to just change the behavior at the same time. The previous series went in too fast for this cleanup pass to be included, so here it is. ;-) Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>

author	"Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>
date	Tue, 20 Sep 2011 20:19:48 +0200
parents	883d28233a4d
children	395ca8cd2669

comparison

equal deleted inserted replaced

-:ff26712a0c50
+:b39d85be78a8
 """
 Return a list of revision(s) that match the given status:
 - ``good``, ``bad``, ``skip``: as the names imply
 - ``range``              : all csets taking part in the bisection
-- ``pruned``             : good|bad|skip, excluding out-of-range csets
+- ``pruned``             : csets that are good, bad or skipped
 - ``untested``           : csets whose fate is yet unknown
 """
 state = load_state(repo)
 if status in ('good', 'bad', 'skip'):
 return [repo.changelog.rev(n) for n in state[status]]
 else:
-# Build sets of good, bad, and skipped csets
+# In the floowing sets, we do *not* call 'bisect()' with more
-goods = set(repo.changelog.rev(n) for n in state['good'])
+# than one level of recusrsion, because that can be very, very
-bads  = set(repo.changelog.rev(n) for n in state['bad'])
+# time consuming. Instead, we always develop the expression as
-skips = set(repo.changelog.rev(n) for n in state['skip'])
+# much as possible.
-# Build kinship of good csets
+# 'range' is all csets that make the bisection:
-ga = goods.copy()   # Goods' Ancestors
+#   - have a good ancestor and a bad descendant, or conversely
-gd = goods.copy()   # Goods' Descendants
+# that's because the bisection can go either way
-for g in goods:
+range = '( bisect(bad)::bisect(good) | bisect(good)::bisect(bad) )'
-ga |= set(repo.changelog.ancestors(g))
-gd |= set(repo.changelog.descendants(g))
+# 'pruned' is all csets whose fate is already known:
+#   - a good ancestor and a good ascendant, or
-# Build kinship of bad csets
+#   - a bad ancestor and a bad descendant, or
-ba = bads.copy()    # Bads' Ancestors
+#   - skipped
-bd = bads.copy()    # Bads' Descendants
+# But in case of irrelevant goods/bads, we also need to
-for b in bads:
+# include them.
-ba |= set(repo.changelog.ancestors(b))
+pg = 'bisect(good)::bisect(good)'   # Pruned goods
-bd |= set(repo.changelog.descendants(b))
+pb = 'bisect(bad)::bisect(bad)'     # Pruned bads
+ps = 'bisect(skip)'                 # Pruned skipped
-# Build the range of the bisection
+pruned = '( (%s) | (%s) | (%s) )' % (pg, pb, ps)
-range  = set(c for c in ba if c in gd)
-range |= set(c for c in ga if c in bd)
+# 'untested' is all cset that are- in 'range', but not in 'pruned'
+untested = '( (%s) - (%s) )' % (range, pruned)
 if status == 'range':
-return [c for c in range]
+return [c.rev() for c in repo.set(range)]
 elif status == 'pruned':
-# We do not want skipped csets that are out-of-range
+return [c.rev() for c in repo.set(pruned)]
-return [c for c in range if c in (goods | bads | skips)]
 elif status == 'untested':
-# Return the csets in range that are not pruned
+return [c.rev() for c in repo.set(untested)]
-return [c for c in range if not c in (goods | bads | skips)]
 else:
 raise error.ParseError(_('invalid bisect state'))

Mercurial > public > mercurial-scm > hg

comparison mercurial/hbisect.py @ 15146:b39d85be78a8