The merge is still missing a few hunks related to temporary tables and
InnoDB log file size. The associated code did not seem to exist in
10.0, so the merge of that needs more work. Until this is fixed, there
are a number of test failures as a result.
For compatibility purposes let InnoDB use DATA_INT for MYSQL_TYPE_ENUM and MYSQL_TYPE_SET.
Silence the warning for these types and let the index translation table to be built anyway.
Test case by Jeremy Cole.
In parallel replication, there are two kinds of events which are
executed in different ways.
Normal events that are part of event groups/transactions are executed
asynchroneously by being queued for a worker thread.
Other events like format description and rotate and such are executed
directly in the driver SQL thread.
If the direct execution of the other events were to update the old-style
position, then the position gets updated too far ahead, before the normal
events that have been queued for a worker thread have been executed. So
this patch adds some special cases to prevent such position updates ahead
of time, and instead queues dummy events for the worker threads, so that
they will at an appropriate time do the position updates instead.
(Also fix a race in a test case that happened to trigger while running
tests for this patch).
which makes it possible to add more world language collations
with very complex collation rules (e.g. Myanmar):
- Weight string for a single character in a user defined collation
was erroneously limited to 7 weights (instead of 8 weights).
Added an extra element in the user-defined weight arrays,
to fit 8 non-zero weights.
- Weight string limit for contractions was made two times longer (16 weights),
which allows longer contractions without affecting the performance
of filesort.
- A user-defined collation now refuses to initialize and reports an error
in case if a weight string gets longer than 8 weights for a single character,
or longer than 16 weights for a contraction. Previously weight strings
for such characters (and contractions) were cut, so a collation
could silently start with wrong rules.
- Fixed a bug in handling rules like "&a << b" in combination with
shift-after-method="expand". The primary weight for "b" was not
correctly calculated, which erroneously made "b" primary greater than "a"
instead of primary equal to "a".
Analysis:
st_select_lex_unit::prepare() computes can_skip_order_by as TRUE.
As a result join->prepare() gets called with order == NULL, and
doesn't do name resolution for the inner ORDER clause. Due to this
the prepare phase doesn't detect that the query references non-exiting
function and field.
Later join->optimize() calls update_used_tables() for a non-resolved
Item_field, which understandably has no Field object. This call results
in a crash.
Solution:
Resolve unnecessary ORDER BY clauses to detect if they reference non-exising
objects. Then remove such clauses from the JOIN object.
GRANT ... IDENTIFIED BY [ PASSWORD ] ...
GRANT ... IDENTIFIED VIA ... [ USING ... ]
GRANT ... REQUIRE ...
GRANT ... MAX_xxx ...
SET PASSWORD FOR ... = ...
- Backport MySQL's fix: do set ha_partition::m_pkey_is_clustered for ha_partition
objects created with handler->clone() call.
- Also, include a testcase.
Slackware install docs to different folder, which wasn't handled
by file_contents test properly. E.g. on Redhat/Debian docs are
expected to be in /usr/share/doc/mariadb-server-(version), but
on Slackware they are in /usr/doc/mariadb-(version).
For details see Slackware mariadb build script:
http://slackbuilds.org/slackbuilds/14.0/system/mariadb/mariadb.SlackBuild
Only allow NONE instead of a role name in SET ROLE.
Don't allow PUBLIC as a role name anywhere (to be fixed later)
Fix db_access calculations on SET ROLE
Reduce the size of role_grants and parent_grantee per-user/role arrays.
Fix the wording and specify the correct sqlstate for ER_INVALID_ROLE
functions for traversing the role graph in either direction.
merging of global, database, table, column, routine privileges.
debug status variables for counting number of privilege merges.
tests.
mainly to avoid the pattern of
* get username/hostname/rolename
* optionally find the corresponding ACL_USER and ACL_ROLE
* allocate memory, concatenate username/hostname/rolename
* call a function passing only this memory as an argument
** use concatenated username/etc to find ACL_USER and ACL_ROLE again
** do something
* free the object
Also to undo push_dynamic we use pop_dynamic now,
not a linear search/scan through the dynamic array.
as a bonus, role@ is now an invalid way to refer to a role.
