Distribution:+ppas and Distribution:+questions issue unlimited queries when memo=0&direction=backwards

Looks like a storm bug SQLObjectResultSet[0:0]:

{{{
>>> from lp.registry.interfaces.distribution import IDistributionSet
>>> from lp.registry.interfaces.distroseries import IDistroSeriesSet
>>> ubuntu = getUtility(IDistributionSet).getByName("ubuntu")
>>> distroseriesset = getUtility(IDistroSeriesSet)

>>> list(distroseriesset.search(distribution=ubuntu))
[<DistroSeries u'warty'>, <DistroSeries u'hoary'>, <DistroSeries
u'grumpy'>, <DistroSeries u'breezy-autotest'>]
# Good
# SELECT DistroSeries.backports_not_automatic,[...] FROM DistroSeries
WHERE distribution = 1 ORDER BY DistroSeries.distribution,
DistroSeries.version

>>> list(distroseriesset.search(distribution=ubuntu)[1:1])
[]
# Good
# SELECT DistroSeries.backports_not_automatic,[...] FROM DistroSeries
WHERE distribution = 1 ORDER BY DistroSeries.distribution,
DistroSeries.version LIMIT 0 OFFSET 1

>>> list(distroseriesset.search(distribution=ubuntu)[0:0])
[<DistroSeries u'warty'>, <DistroSeries u'hoary'>, <DistroSeries
u'grumpy'>, <DistroSeries u'breezy-autotest'>]
# No limit nor offset in the generated query ... ?!
# SELECT DistroSeries.backports_not_automatic,[...] FROM DistroSeries
WHERE distribution = 1 ORDER BY DistroSeries.distribution,
DistroSeries.version
}}}

The plan here is to:
- fix the bug in Storm
- fix the batchnavigator to not even issue a query if the slice is going to be of the form [N:N]

This is nearly right :) the if block you remove needs to stay and be tweaked though:
>>> range(10)[None:None]
[0, 1, 2, 3, 4, 5, 6, 7, 8, 9]

Profiling the queries issued by the page on qastaging confirms the limit is now there in the sql query ("ORDER BY Question.datecreated DESC LIMIT 0 OFFSET 0").

Can we do a release of lazr.batchnavigator then, clear the inventory out?

I've released lazr.batchnavigator 1.2.11, and will upgrade Launchpad to that.

And I've confirmed that this does indeed get rid of the extra LIMIT 0 query.