The change in f105fbe52b21da206bfbaedf0e92326667d7b2d4 made through=None
m2m fields fail in cases where they worked before. It isn't possible to
create such fields using public APIs. The fix is trivial, so it seems
worth fixing this for custom m2m field users.
This is not a backport from master. Master has gotten enough other
changes to related fields internal API that this fix alone isn't enough
to do any good.
SQLite used INSERT INTO tbl SELECT %s UNION SELECT %s, the problem
was that there should have been UNION ALL instead of UNION.
Refs #19351
Backpatch of a27582484cf814554907d2d1ad077852de36963f
The problem is the same as in #10888 which was reintroduced when
bulk_insert was added. Thanks to Jani Tiainen for report, patch and
also testing the final patch on Oracle GIS.
Backpatch of 92d7f541da8b59520c833b19fbba52d3ecef2428
There was a potential data-loss issue involved -- when clearing
instance's m2m assignments it was possible some other instance's
m2m data was deleted instead.
This commit also improved None handling for to_field cases.
Backpatch of 611c4d6f1c24763e5e6e331a5dcf9b610288aaa8
The problem description in #18239 asserted that
http://bugs.python.org/issue670664 was fixed in Python 2.6.8, but based on
http://bugs.python.org/issue670664#msg146770 it appears that's not correct; the
fix was only applied in 2.7, 3.2, and Python trunk. Therefore we must use our
patched HTMLParser subclass in all Python 2.6 versions.
Backport of fcec904e4f from master. Fixes#19148.
The qs.bulk_create() method did not work with large batches together
with SQLite3. This commit adds a way to split the bulk into smaller
batches. The default batch size is unlimited except for SQLite3 where
the batch size is limited to 999 SQL parameters per batch.
Thanks to everybody who participated in the discussions at Trac.
Backpatch of 29132ebdef0e0b9c09e456b05f0e6a22f1106a4f from master (with
documentation changes removed).
Unlike other model fields, the newly introduced (1.4)
GenericIPAddressField did not accept verbose_name and name as the
first positional arguments. This commit fixes it.
Thanks Dan McGee for the report and the patch.
Backport of 306d34873cff2 from master.
The piece of state is DB connections' allow_thread_sharing attribute
which gets munged when test are run when in-memory SQLite databases.
Thanks Anssi for suggesting the possible root cause and Julien for
implementing the fix.
Backport of ea667ee3aeed33bce1dd681d9c0ea42f9926db5a from master.
On MySQL when checking the server version, a new connection could be
created but never closed. This could result in open connections on
server startup.
Backport of 4423757c0c50afbe2470434778c8d5e5b4a70925.
This should prevent a race condition if running collectstatic is
canceled or its cache is accessed from other processes, leaving the
cache in a corrupt state.
Backport of 9350d1d59c1a4e6a9ac246a808f55da35de0df69 from master.
This was a regression in Django 1.4.
Thanks bronger for the report and claudep for the patch.
The previous commit didn't work with PyLibMC.
This solution appears to be the best compromise
at this point in the 1.4 release cycle.
git-svn-id: http://code.djangoproject.com/svn/django/trunk@17797 bcc190cf-cafb-0310-a4f2-bffc1f526a37
Catches memcached session key errors related to overly long session keys.
This is a long-standing bug, but severity was exacerbated by the addition
of cookie-backed session storage, which generates long session values. If
an installation switched from cookie-backed session store to memcached,
users would not be able to log in because of the server error from overly
long memcached keys.
git-svn-id: http://code.djangoproject.com/svn/django/trunk@17795 bcc190cf-cafb-0310-a4f2-bffc1f526a37
