\cache::make() may return a cache_disabled subclass but callers, specifically
user/lib.php:user_create_user(), invoke \cache::purge_current_user() which was
only implemented by the cache_session subclass. Added empty
\cache::purge_current_user() which subclasses can implement, i.e.
cache_session, or not, i.e. cache_disabled.
The current logic in the cache_session::check_tracked_user() is not
right. We must always set the current session id. A typical use case is
when the cache instance is instantiated for a not logged in user. We
can't let the sessionid property null in that case as it forms an
important part of the parsed key.
Similarly, even if we have the same user currently loaded, we must still
set the sessionid to make sure the data will be associated with the
current PHP session. Same user (including visitors or guest users) can
access the site from different browsers and each must end up with its
own key prefix.
It is now safe to cache a reference to a cache and expect consistent results.
Changing identifiers altered cache results where a reference was
held to the cache. Identifiers have been set to be cached with
identifiers included so the caches are separate.
As a consequence of this it was identified that invalidation events
and identifiers don't easily work together as an event can't determine
which identifiers should be used for cache invalidation. So invalidation
events have been made incompatible with identifiers being set. No core
code used this combination as it's not possible to understand any expected
behaviour.
Event invalidation for application and session caches was centralised to the same
location. The only difference was the name of the lastinvalidation variable. This
improves support and consistency of invalidation code.
Previously a purge_store was only purging caches but it was leaving a
reference to the store instance for the cache definition.
As a result, the cache was never cleared away correctly and, in cases
where the cache store makes a TCP connection to a backend system, the
number of open sockets was increasing for each store instance.
After a complete purge, as found in the unit test reset, the cache
definitions should be both purged and removed.
This change should also lead to a drop in memory consumption for unit
tests as there will no longer be references to unused definition stores.
Ensure that the cache store requirements are met prior to attempting
to instantiate a class. Many of the constructors create connections
to external services requiring the installation and availability of
classes that are specified in are_requirements_met(). If they are
not checked properly you end up with PHP Fatal errors rather than
falling back to a different store that is capabable without the
the extra classes loaded.
purge_all() and purge_by_definition() look in the configuration
for which caches are available and then creates them to purge them.
The configuration stores the values used by initialise(), not
initialise_unit_test_instance() and would therefore fail to purge
all caches if they were not purged by another means.
In the case of filestore, it's purged by unit tests, in the case
of memcache(d), it purges the whole store when a single definition
is requested.
Therefore all configuration was moved into the configuration file
during unit tests and does not have any special override codes in
the unit test infrastructure.
Static cache has a very flexible storage capability
and does not need all the checks performed on it.
Enabling dereferencing handling and multiple identifiers
removes a lot of unneeded overhead for request caches that
have many get calls.
The changes were modelled on the static acceleration changes
done in MDL-53208.
When a cachable object is store in the static cache from
the backing store, it was incorrect serialised rather than
using the wake function. This has been resolved and tests added.
During the investigation into cacheable_object, it was discovered
that set_identifiers never removes identifiers when you call it,
so set_identifiers(array('a')) and set_identifiers(array('b')) really
resulted in array('a','b') as the identifiers rather than 'b'.
The fix for this issue depends on the set_identifiers fix and
they have been coupled together as a result.
- We only now serialize when required, eg when the store doesn't do it.
- static acceleration now serializes on set to reduce overhead on load.
- null now works in static acceleration as we use an array for storage metadata
- static acceleration delete uses isset rather than an array_search.
- We only now serialize when required, eg when the store doesn't do it.
- static acceleration now serializes on set to reduce overhead on load.
- null now works in static acceleration as we use an array for storage metadata
- static acceleration delete uses isset rather than an array_search.
If the cache does all the dereferencing when it stores and loads
the objects, then the cache loaders don't need to do that work.
This is true of all caches that use something other than PHP's memory
to store their results.
The cache store mode is now included in the performance stats
printed at the bottom of the page.
It is represented as either [a] [s] or [r] and a title is used
to actually state the mode.
behat/lib.php script should not be usually included, neither any of its functions
used within mooodle code at all. It's for exclusive use of behat and
moodle setup.php. For places requiring a different/special behavior
needing to check if are being run as part of behat tests, use:
if (defined('BEHAT_SITE_RUNNING')) { ...
Cache stores are now responsible for deciding if they are suitable
for use as the primary cache store during unit and acceptance tests
Changes are as follows:
* New method: cache_store::ready_to_be_used_for_testing
* Core cache store have been updated to override this method as
required.
* MongoDB usesafe default changed to true to match actual default
behaviour.
Cache data source aggregate functionality was found to be broken
and unused, because of this the decision was made to remove it
rather than fix it.
As it was broken we did not follow typical deprecation methods and
instead the code was removed outright with only structure
remaining and left deprecated.
