| Age | Commit message (Collapse) | Author |
|
|
|
|
|
|
|
|
|
|
|
|
|
database routers.
Partially reverted refs #27054 except for one of the tests as this
solution supersedes that one.
Thanks Shai Berger for the review.
|
|
Thanks to Simon Litchfield for the report, Uninen for the initial
patch, akaihola, jamesp, b.schube, and Florian Appoloner for
subsequent patches, tests, and documentation.
|
|
|
|
|
|
Thanks Tim Graham for the review.
|
|
|
|
|
|
|
|
Thanks Konrad Świat for the original patch.
|
|
|
|
|
|
|
|
Previously, the MemcachedCache backend ignored `OPTIONS` and
PyLibMCCache used them to set pylibmc behaviors. Both backends now
pass `OPTIONS` as keyword arguments to the client constructors.
|
|
|
|
|
|
Thanks Tim for the review.
|
|
|
|
Passing an Unicode string on Python 2 was crashing the connection.
Thanks slavugan@gmail.com for the report, and Tim Graham for the review.
|
|
|
|
|
|
|
|
|
|
|
|
import errors.
Thanks Tim for the review.
|
|
|
|
get_constraints().
|
|
|
|
|
|
same test run.
|
|
The first half of the test fails when using pylibmc (so will need
to be skipped).
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Since the `cull` and `zero_cull` test cache configs set `MAX_ENTRIES`
and `CULL_FREQUENCY` in `OPTIONS`, which are only intended for use with
the locmem, filesystem, and database backends. This prevents test
failures once refs #20892 is fixed.
|
|
|
|
|
|
|
|
|
|
|
|
Thanks Berker Peksag and Tim Graham for the reviews.
|