Q objects don't work properly when using spanned relationships

Note: Of course, lines 5 and 6 should read as follows:

>>> websites.filter(domain=d)
[<Website: 182>, <Website: 183>]

This is #2080, which has now been fixed on the queryset-refactor branch and will be merged into trunk when that work is finished.

Sorry for reopening it, but it still doesn't work properly in the queryset-refactor branch:

>>> Website.objects.filter(pk__in=[182,183]).filter(Q(domain=d) | Q(websitealias__domain=d)).distinct().count()
41L
>>> Website.objects.filter(Q(domain=d) | Q(websitealias__domain=d)).filter(pk__in=[182,183]).distinct().count()                        
2L

In that case, please construct a small, complete example that demonstrates the problem. At the moment, all I have to go on is the "shape" of the query, which is why it looks like #2080, which did have problems.

Since it's apparently not that issue, what is a small model and data set that can be used to repeat the bug for testing purposes (I have no idea what your Website contains and which of the relations in your setup are significant here).

from django.db import models

class A(models.Model):
    def __unicode__(self):
        return '%d' % self.id

class B(models.Model):
    a = models.ForeignKey(A)
    def __unicode__(self):
        return '%d' % self.id

class C(models.Model):
    a = models.ForeignKey(A)
    b = models.ForeignKey(B)
    def __unicode__(self):
        return '%d' % self.id

>>> from django.db.models import Q
>>> from qs.models import A,B,C
>>> a1 = A.objects.create()
>>> a2 = A.objects.create()
>>> b1 = B.objects.create(a=a1)
>>> b2 = B.objects.create(a=a1)
>>> b3 = B.objects.create(a=a2)
>>> c1 = C.objects.create(a=a1, b=b1)
>>> a1, a2, b1, b2, b3, c1
(<A: 1>, <A: 2>, <B: 1>, <B: 2>, <B: 3>, <C: 1>)
>>> B.objects.filter(pk__in=[b1.id,b2.id]).filter(Q(a=a1.id) | Q(c__a=a1.id))           
[<B: 1>, <B: 2>]

Until now, everything is okay. Now let's create another C object:

>>> c2 = C.objects.create(b=b3, a=a1)
>>> B.objects.filter(pk__in=[b1.id,b2.id]).filter(Q(a=a1.id) | Q(c__a=a1.id))
[<B: 1>, <B: 2>, <B: 3>]
>>> B.objects.filter(Q(a=a1.id) | Q(c__a=a1.id)).filter(pk__in=[b1.id,b2.id])
[<B: 1>, <B: 2>]

b3 shouldn't appear here, because it is filtered. It works correctly when filtering the PK at the end.

(In [6957]) queryset-refactor: Fixed a problem when adding certain additional filters to a queryset that has precisely one filter attached already.

Refs #6154.

(In [7477]) Merged the queryset-refactor branch into trunk.

This is a big internal change, but mostly backwards compatible with existing code. Also adds a couple of new features.

Fixed #245, #1050, #1656, #1801, #2076, #2091, #2150, #2253, #2306, #2400, #2430, #2482, #2496, #2676, #2737, #2874, #2902, #2939, #3037, #3141, #3288, #3440, #3592, #3739, #4088, #4260, #4289, #4306, #4358, #4464, #4510, #4858, #5012, #5020, #5261, #5295, #5321, #5324, #5325, #5555, #5707, #5796, #5817, #5987, #6018, #6074, #6088, #6154, #6177, #6180, #6203, #6658