A new optional parameter $context has been added to the
core_group::groups_get_members_join() function.
Besides, some core_group methods now accept -1 (USERSWITHOUTHGROUP) for
the groupid field.
Previously, reset_course_userdata() was deleting the role_capabilities
directly, but has_capability() checks the local capabilities cache
which was still returning the previous value.
Whenever the role_capabilities table is changed, we need to remember
to clear the cache for the role(s) being modified. This is far simpler
when all of those changes happen in a single file, i.e. accesslib.php,
allowing other parts of the application to use the public functions
without requiring knowledge of the cache internals.
Thanks to MDL-49398, we can separate the combined user session and role
definition cache clearing function into two separate functions. At the
same time, we want to identify and remove mark_dirty() calls that were
added for role definition changes but were incorrectly left behind.
Change highlights:
- Remove unnecessary mark_dirty() calls performed after
assign_capability(), unassign_capability(), delete_role(),
deleted contexts, brand new contexts
- Move role definition cache clear from the user-centric
accesslib_clear_all_caches() to the newly created,
role-dedicated accesslib_reset_role_cache()
This is another preliminary commit, rationalising the ->work... fields
and also moving the code that sets them up into a separate function.
This is also to prepare for the next commit.
This eliminates a lot of necessary string manipulation for something
that preg can just do with the \b assertion.
The change also extracts all the work to prepare ->work_phrase (renamed
->regexp) into a separate function. This is to pave the way for future
efficiency gains, but for now I have not done them, so it is easier to
verify by inspection that this commit does not change behaviour.
Also, some tidy-ups to the filterobject structure, to eliminate some
redundant fields.
Instead of using set_coursemodule_name() to proceed with the name
change we perform the change earlier in the process, both to avoid
duplicate events caused by the renaming function and also to get the new
name used by all proper generated events.
