Rev 2367: (John Arbash Meinel) allow win32 to grab a write lock after 'bzr status', so that we can record the updated stat values. in /home/pqm/archives/thelove/bzr/+trunk/

Canonical.com Patch Queue Manager pqm at pqm.ubuntu.com
Tue Mar 20 22:39:40 GMT 2007


------------------------------------------------------------
revno: 2367
revision-id: pqm at pqm.ubuntu.com-20070320223938-97fdc295a1111e36
parent: pqm at pqm.ubuntu.com-20070320222307-30c846e90ac18c65
parent: john at arbash-meinel.com-20070320215959-1vwhp1nl861bf7ff
committer: Canonical.com Patch Queue Manager<pqm at pqm.ubuntu.com>
branch nick: +trunk
timestamp: Tue 2007-03-20 22:39:38 +0000
message:
  (John Arbash Meinel) allow win32 to grab a write lock after 'bzr status', so that we can record the updated stat values.
added:
  bzrlib/tests/per_lock/test_temporary_write_lock.py test_temporary_write-20070314233412-xp3ocbyvw3woa03w-1
modified:
  NEWS                           NEWS-20050323055033-4e00b5db738777ff
  bzrlib/dirstate.py             dirstate.py-20060728012006-d6mvoihjb3je9peu-1
  bzrlib/errors.py               errors.py-20050309040759-20512168c4e14fbd
  bzrlib/lock.py                 lock.py-20050527050856-ec090bb51bc03349
  bzrlib/tests/per_lock/__init__.py __init__.py-20070314201444-u92yjsqrkh2m3qcb-1
  bzrlib/tests/per_lock/test_lock.py test_lock.py-20070313190612-mfpoa7t8kvrgrhj2-1
  bzrlib/tests/test_commit_merge.py test_commit_merge.py-20050920084723-819eeeff77907bc5
  bzrlib/tests/test_dirstate.py  test_dirstate.py-20060728012006-d6mvoihjb3je9peu-2
    ------------------------------------------------------------
    merged: john at arbash-meinel.com-20070320215959-1vwhp1nl861bf7ff
    parent: john at arbash-meinel.com-20070319221128-iphfdqrrfqypfjva
    committer: John Arbash Meinel <john at arbash-meinel.com>
    branch nick: simple_locking
    timestamp: Tue 2007-03-20 16:59:59 -0500
    message:
      NEWS
    ------------------------------------------------------------
    merged: john at arbash-meinel.com-20070319221128-iphfdqrrfqypfjva
    parent: john at arbash-meinel.com-20070319214503-1t048spp6n2tbro7
    committer: John Arbash Meinel <john at arbash-meinel.com>
    branch nick: simple_locking
    timestamp: Mon 2007-03-19 17:11:28 -0500
    message:
      make Write locks not block on Read locks, so that revert tests don't fail
      Specifically: 
        test_revert.TestRevert.test_preserve_execute
        test_revert.TestRevert.test_revert_executable
      
      fail if we fail to take a Write lock if we have a Read lock.
    ------------------------------------------------------------
    merged: john at arbash-meinel.com-20070319214503-1t048spp6n2tbro7
    parent: john at arbash-meinel.com-20070319205618-21plqrdf0feihonk
    committer: John Arbash Meinel <john at arbash-meinel.com>
    branch nick: simple_locking
    timestamp: Mon 2007-03-19 16:45:03 -0500
    message:
      Disable read locks blocking on write locks.
      This requires some small test updates to disable that sort of check.
      But it should allow the rest of the test suite to pass on non-win32
    ------------------------------------------------------------
    merged: john at arbash-meinel.com-20070319205618-21plqrdf0feihonk
    parent: pqm at pqm.ubuntu.com-20070317015305-7b7562331da9f786
    parent: john at arbash-meinel.com-20070319204546-xxywya2iuemwqdt9
    committer: John Arbash Meinel <john at arbash-meinel.com>
    branch nick: simple_locking
    timestamp: Mon 2007-03-19 15:56:18 -0500
    message:
      [merge] locking changes into bzr.dev tip
    ------------------------------------------------------------
    merged: john at arbash-meinel.com-20070319204546-xxywya2iuemwqdt9
    parent: john at arbash-meinel.com-20070315233905-1keqs8pqeas1smq2
    committer: John Arbash Meinel <john at arbash-meinel.com>
    branch nick: locking
    timestamp: Mon 2007-03-19 15:45:46 -0500
    message:
      Remove the unused _ignore_write_lock parameter.
    ------------------------------------------------------------
    merged: john at arbash-meinel.com-20070315233905-1keqs8pqeas1smq2
    parent: john at arbash-meinel.com-20070315232459-tf23624lgm7zz1yj
    parent: bialix at ukr.net-20070315232743-ol8q5vyqzpvdkrjy
    committer: John Arbash Meinel <john at arbash-meinel.com>
    branch nick: locking
    timestamp: Thu 2007-03-15 18:39:05 -0500
    message:
      [merge] Alexander Belchenko: forgotten import of TestSkipped
        ------------------------------------------------------------
        merged: bialix at ukr.net-20070315232743-ol8q5vyqzpvdkrjy
        parent: john at arbash-meinel.com-20070315191520-3ogywmtl6v5gop7q
        committer: Alexander Belchenko <bialix at ukr.net>
        branch nick: jam-locking-bialix
        timestamp: Fri 2007-03-16 01:27:43 +0200
        message:
          forgotten import of TestSkipped
    ------------------------------------------------------------
    merged: john at arbash-meinel.com-20070315232459-tf23624lgm7zz1yj
    parent: john at arbash-meinel.com-20070315191520-3ogywmtl6v5gop7q
    parent: pqm at pqm.ubuntu.com-20070315231915-000d2bef502ae12b
    committer: John Arbash Meinel <john at arbash-meinel.com>
    branch nick: locking
    timestamp: Thu 2007-03-15 18:24:59 -0500
    message:
      [merge] bzr.dev 2359
    ------------------------------------------------------------
    merged: john at arbash-meinel.com-20070315191520-3ogywmtl6v5gop7q
    parent: john at arbash-meinel.com-20070315185559-xtb5q6p8ldl2mvs1
    committer: John Arbash Meinel <john at arbash-meinel.com>
    branch nick: locking
    timestamp: Thu 2007-03-15 13:15:20 -0600
    message:
      Use the right lock object.
    ------------------------------------------------------------
    merged: john at arbash-meinel.com-20070315185559-xtb5q6p8ldl2mvs1
    parent: john at arbash-meinel.com-20070315181445-8ig26uqlkq8bso3b
    committer: John Arbash Meinel <john at arbash-meinel.com>
    branch nick: locking
    timestamp: Thu 2007-03-15 13:55:59 -0500
    message:
      Change the temporary_write_lock api, so that it always returns a lock object,
      and a flag indicating whether it was able to grab a write lock.
      This is because callers don't know whether it had to unlock the original lock, so it
      should always return something that lets callers maintain proper state.
    ------------------------------------------------------------
    merged: john at arbash-meinel.com-20070315181445-8ig26uqlkq8bso3b
    parent: john at arbash-meinel.com-20070315180216-jddyle9fxhr8tlde
    committer: John Arbash Meinel <john at arbash-meinel.com>
    branch nick: locking
    timestamp: Thu 2007-03-15 13:14:45 -0500
    message:
      ctypes locks should return ctypes locks.
    ------------------------------------------------------------
    merged: john at arbash-meinel.com-20070315180216-jddyle9fxhr8tlde
    parent: john at arbash-meinel.com-20070315171948-5m28gvw153egui90
    committer: John Arbash Meinel <john at arbash-meinel.com>
    branch nick: locking
    timestamp: Thu 2007-03-15 13:02:16 -0500
    message:
      Update DirState to use the new 'temporary_write_lock', and add tests that it works.
    ------------------------------------------------------------
    merged: john at arbash-meinel.com-20070315171948-5m28gvw153egui90
    parent: john at arbash-meinel.com-20070314234034-udg2sm249k9dqj8r
    committer: John Arbash Meinel <john at arbash-meinel.com>
    branch nick: locking
    timestamp: Thu 2007-03-15 12:19:48 -0500
    message:
      Implement temporary_write_lock, and restore_read_lock for win32 locks.
    ------------------------------------------------------------
    merged: john at arbash-meinel.com-20070314234034-udg2sm249k9dqj8r
    parent: john at arbash-meinel.com-20070314223138-wfmodqhruo0g5l0e
    committer: John Arbash Meinel <john at arbash-meinel.com>
    branch nick: locking
    timestamp: Wed 2007-03-14 18:40:34 -0500
    message:
      Implement a 'ReadLock.temporary_write_lock()' to upgrade to a write-lock in-process.
      Implement it for _fcntl.
      This is the api that can be used when we want to update a dirty dirstate
      even though we only have a read lock.
      On win32, this will actually unlock and re-lock the file. Which is part of
      the api description. But we don't have to limit ourselves to the lowest
      common denominator on all platforms.
    ------------------------------------------------------------
    merged: john at arbash-meinel.com-20070314223138-wfmodqhruo0g5l0e
    parent: john at arbash-meinel.com-20070313233914-bigtq3yxnmf90d9s
    parent: john at arbash-meinel.com-20070314205703-taih3n87g9n4afdv
    committer: John Arbash Meinel <john at arbash-meinel.com>
    branch nick: locking
    timestamp: Wed 2007-03-14 17:31:38 -0500
    message:
      [merge] LockCleanup changes.
    ------------------------------------------------------------
    merged: john at arbash-meinel.com-20070313233914-bigtq3yxnmf90d9s
    parent: john at arbash-meinel.com-20070313231812-h74g1zz32v12lv6s
    committer: John Arbash Meinel <john at arbash-meinel.com>
    branch nick: locking
    timestamp: Tue 2007-03-13 17:39:14 -0600
    message:
      (broken) Change fcntl locks to be properly exclusive within the same process.
      This exposes the location where we are opening the same WorkingTree in both read and write
      mode, which breaks several tests.



More information about the bazaar-commits mailing list