pushrebase: update fastmanifest test

This test looked like it's been funky since it was introduced. It harded code
an exception from a bug.  The bug was that we're storing bundle manifests in a
cache during pushrebase, and because those manifests are hybrid manifests they
haven't actually loaded from disk. Later we close the bundle, then after that we
try to access these cached manifests, which throws because they can't read from
the file.

The recent change to read the copied files from the manifest before the lock is
taken, forces these manifests to be in loaded, thus fixing the problem. So let's
update the test.

tests/test-pushrebase-fastmanifest.t

@@ -53,27 +53,25 @@ Create the dummy commit on client 1
   $ hg add newfile
   $ hg commit -m 'dummy commit'
 
-Test that pushing to a remotename gets rebased (client1 -> client2), but fails.
+Test that pushing to a remotename gets rebased (client1 -> client2) works
 
-  $ hg push --to mybook ssh://user@dummy/client2 2>&1 | tail -4
-  remote:   File "*changegroup.py", line *, in seek (glob)
-  remote:     return self._stream.seek(pos)
-  remote: ValueError: I/O operation on closed file
-  abort: stream ended unexpectedly (got 0 bytes, expected 4)
-
-Disable fastmanifest from client2
-
-  $ cd ../client2
-  $ cat >> .hg/hgrc << EOF
-  > [extensions]
-  > fastmanifest = !
-  > EOF
-
-Test that pushing to a remotename gets rebased (client1 -> client2)
-
-  $ cd ../client1
   $ hg push --to mybook ssh://user@dummy/client2
   pushing to ssh://user@dummy/client2
   searching for changes
   remote: pushing 1 changeset:
   remote:     eb7a4df38d10  dummy commit
+
+  $ cd ../client2
+  $ hg log -G
+  o  changeset:   1:eb7a4df38d10
+  |  tag:         tip
+  |  user:        test
+  |  date:        Thu Jan 01 00:00:00 1970 +0000
+  |  summary:     dummy commit
+  |
+  @  changeset:   0:2bb9d20e471c
+     bookmark:    master
+     user:        test
+     date:        Thu Jan 01 00:00:00 1970 +0000
+     summary:     initial
+