Also deprecated the following functions -
1. message_move_userfrom_unread2read - It is not necessary
for us to mark a message as read on user deletion.
2. message_get_blocked_users - Horrible logic used to
determine if a user is blocked via reference on some
randomly chosen 'isblocked' variable.
3. message_get_contacts - The same as above. This can be
done in a much nicer way.
4. message_mark_message_read - We want two functions to do
this to avoid confusing messages and notifications.
5. message_can_delete_message - This assumed the variable
$message contained the 'useridto' property, which
was present in the old table structure. We do not want
future usages where a query is done on the new table
and is simply passed as this won't contain this property.
6. message_delete_message - Same as above.
Previously we appended a dot at the end of IP addresses and domain names in the
cURL security helper, but it causes issues with Google OAuth so this patch removes it.
Freaky deaky Unicode/octal/hex domains can be resolved by cURL but are technically not valid.
This patch causes anything that Moodle does not consider to be a valid domain or IP to be blocked
by the cURL security helper.
There is an edge case whereby redis will fail
to accept connections on the first try but
retrying the connection seems to make it work
Included in this commit:
* Retry functionality in the session init
There were basically two problems, which are demostrated by
the new test users u7 and u8 in the unit test.
1. There was a problem if a role was overridden in a context above the
one where it was assigned. E.g. User has teacher role in a course, and
there is a role override in the course-category context. That was being
ignored.
2. Problems with the handling of PROHIBITs. It should be the case that
if there is a PROHIBIT in force, then it cannot be overridden by another
role, or by a role override. However, this was not working in all cases.
Also, I had to add comments to the unit test so that I could understand
it. Hopefully these will be hepful to other developers too.
The new recordset support for Postgres requires transactions and
will cause errors if recordsets are not closed correctly. This
commit fixes problems that were identified during unit tests, and
via some basic code analysis, across all core code. Most of these
are incorrect usage of recordset (forgetting to close them).
Triggering a fatal error in an adhoc task is bad. It will be retried indefinitely.
Even though we are not sure how to get a module instance without a course module record,
it is possible and should not kill the Moodle site.
