[PATCH 1 of 7] obsolete: cheap detection of nullid as successors

[pierre-yves.david at logilab.fr]

# HG changeset patch
# User Pierre-Yves David <pierre-yves.david at logilab.fr>
# Date 1349257585 -7200
# Node ID af731e6a342308e9111c11287bd73a1a0fcbbd47
# Parent  57fe5aca86af8d82b9c932bf68ae536a99365c86
obsolete: cheap detection of nullid as successors

Nullid as successors create multiple issues:

- Nullid revnum is -1, confusing algorithm that use revnum unless you add
  special handling in all of them.
- Nullid confuse "divergent" changeset detection and resolution. As you can't
  add any successors to Nullid without being in even more troubles

Fortunately, there is not good reason to use nullid as a successor. The only
sensible meaning of "succeed by nullid" is "drooped" and this meaning is already
covered by obsolescence marker with empty successors set.

However, letting some nullid successors to slip in may cause terrible damage in
such algorithm difficult to debug. So I prefer to performs and clear detection
of of such pathological changeset. We could be much smarter by cleaning up
nullid successors on the fly but it would be much for expensive. As core
Mercurial does not create any such changeset, I think its fine to just abort
when suspicious situation is detected.

Earlier experimental version created such changeset, so there is some out there.
The evolve extension added the necessary logic to clean up its mess.

diff --git a/mercurial/obsolete.py b/mercurial/obsolete.py
--- a/mercurial/obsolete.py
+++ b/mercurial/obsolete.py
@@ -50,11 +50,11 @@ The header is followed by the markers. E
   string contains a key and a value, separated by a color ':', without
   additional encoding. Keys cannot contain '\0' or ':' and values
   cannot contain '\0'.
 """
 import struct
-import util, base85
+import util, base85, node
 from i18n import _
 
 _pack = struct.pack
 _unpack = struct.unpack
 
@@ -235,10 +235,13 @@ class obsstore(object):
             self._all.append(mark)
             pre, sucs = mark[:2]
             self.precursors.setdefault(pre, set()).add(mark)
             for suc in sucs:
                 self.successors.setdefault(suc, set()).add(mark)
+        if node.nullid in self.successors:
+            raise util.Abort('Bad obsolescence marker detected: '
+                             'invalid successors nullid')
 
 def _encodemarkers(markers, addheader=False):
     # Kept separate from flushmarkers(), it will be reused for
     # markers exchange.
     if addheader:
diff --git a/tests/test-obsolete.t b/tests/test-obsolete.t
--- a/tests/test-obsolete.t
+++ b/tests/test-obsolete.t
@@ -98,10 +98,17 @@ Register two markers with a missing node
   245bde4270cd1072a27757984f9cda8ba26f08ca cdbce2fbb16313928851e97e0d85413f3f7eb77f 0 {'date': '56 12', 'user': 'test'}
   cdbce2fbb16313928851e97e0d85413f3f7eb77f ca819180edb99ed25ceafb3e9584ac287e240b00 0 {'date': '1337 0', 'user': 'test'}
   ca819180edb99ed25ceafb3e9584ac287e240b00 1337133713371337133713371337133713371337 0 {'date': '1338 0', 'user': 'test'}
   1337133713371337133713371337133713371337 5601fb93a350734d935195fee37f4054c529ff39 0 {'date': '1339 0', 'user': 'test'}
 
+Refuse pathological nullid successors
+  $ hg debugobsolete -d '9001 0' 1337133713371337133713371337133713371337 0000000000000000000000000000000000000000
+  transaction abort!
+  rollback completed
+  abort: Bad obsolescence marker detected: invalid successors nullid
+  [255]
+
 Check that graphlog detect that a changeset is obsolete:
 
   $ hg glog
   @  changeset:   5:5601fb93a350
   |  tag:         tip