Denormalize the query that checks for unread messages

[brettp]

Because it looks like this and makes servers (and sysadmins) weep:

SELECT DISTINCT e.*
FROM elggentities e
JOIN elggmetadata n_table ON e.guid = n_table.entity_guid
JOIN elggmetadata n_table1 ON e.guid = n_table1.entity_guid
JOIN elggmetastrings msn1 ON n_table1.name_id = msn1.id
JOIN elggmetastrings msv1 ON n_table1.value_id = msv1.id
JOIN elggmetadata n_table2 ON e.guid = n_table2.entity_guid
JOIN elggmetastrings msn2 ON n_table2.name_id = msn2.id
JOIN elggmetastrings msv2 ON n_table2.value_id = msv2.id
JOIN elggmetadata n_table3 ON e.guid = n_table3.entity_guid
JOIN elggmetastrings msn3 ON n_table3.name_id = msn3.id
JOIN elggmetastrings msv3 ON n_table3.value_id = msv3.id
WHERE (((msn1.string = 'toId'
         AND BINARY msv1.string = 11814
         AND ((1 = 1)
              AND n_table1.enabled='yes'))
        AND (msn2.string = 'readYet'
             AND BINARY msv2.string = 0
             AND ((1 = 1)
                  AND n_table2.enabled='yes'))
        AND (msn3.string = 'msg'
             AND BINARY msv3.string = 1
             AND ((1 = 1)
                  AND n_table3.enabled='yes'))))
  AND (e.owner_guid IN (11814))
  AND (e.site_guid IN (1))
  AND ((1 = 1)
       AND e.enabled='yes')
ORDER BY e.time_created DESC LIMIT 0,
                                   10

Merge up to trunk.

[brettp]

(In [svn:8763]) Fixes #3164. Denormalized unread message counting query because it was making things so slow I couldn't test.

[brettp]

(In [svn:8764]) Refs #3164. Merged messages denormalization to 1.7 branch.