| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
The reason for this is documented in the middleware itself. Without
this, pgbouncer is of little use to us since it has to throw away every
connection we try to route through it because of unclean disconnects.
In theory, with the switch to using pgbouncer for all WSGI originating
connections and adding this middleware, we should see a notable decrease
in connection time to the database.
Signed-off-by: Dan McGee <dan@archlinux.org>
|
|
It's time to stop serving up stale pages. Remove this middleware caching
and start pushing it down to spots where we can actually control it more
appropriately (and only cache things that are expensive anyway).
Signed-off-by: Dan McGee <dan@archlinux.org>
|
|
Signed-off-by: Dan McGee <dan@archlinux.org>
|
|
This was having some serious effects on caching as we would always have to
access the user in the session, marking every page with a "Vary: Cookie"
header. This is the start of stamping that out. The way we get the user for
news item creation is now more similar to that from the todo lists, but not
quite. That should be adjusted to be more like the news item creation.
Signed-off-by: Dan McGee <dan@archlinux.org>
|
|
This is to address a rather large issue with caching of feed objects in
Django. Because they are built up using an XML library that does multiple
writes on a file-like object, a single feed object, even when pulled from
memcached, generates 1582 writes to the open socket rather than the optimal
one it could do.
Some version of this fix will be making it upstream, but I need to figure
out how to approach that before I do so and for now this will address one of
our larger performance issues on the live site since the packages feed is
hit as often as it is.
Signed-off-by: Dan McGee <dan@archlinux.org>
|
|
First off, the copyright notice is from the other middleware previously
removed, so all of that can go, along with some stragglers in the import
list.
Next, as stated on
http://code.djangoproject.com/wiki/CookBookThreadlocalsAndUser, you need to
be sure to have the import in the model class be the same as the import used
in the middleware declaration, which was not true. Whoops.
Signed-off-by: Dan McGee <dan@archlinux.org>
|
|
|
|
|
|
|
|
|
