Bug#59410 read uncommitted: unlock row could not find a 3 mode lock
on the record
This bug is present only in 5.6 but I am adding the test case to earlier
versions to ensure it never appears in earlier versions too.
Bug#59410 read uncommitted: unlock row could not find a 3 mode lock
on the record
This bug is present only in 5.6 but I am adding the test case to earlier
versions to ensure it never appears in earlier versions too.
on lctn2 systems
There was a local variable in get_all_tables() to store the
"original" value of the database name as it can get lowercased
depending on the lower_case_table_name value.
get_all_tables() iterates over database names and for each
database iterates over the tables in it.
The "original" db name was assigned in the table names loop.
Thus the first table is ok, but the second and subsequent tables
get the lowercased name from processing the first table.
Fixed by moving the assignment of the original database name
from the inner (table name) to the outer (database name) loop.
Test suite added.
In the string context the MIN() and MAX() functions don't take
into account the unsignedness of the UNSIGNED BIGINT argument
column.
I.e.:
CREATE TABLE t1 (a BIGINT UNSIGNED);
INSERT INTO t1 VALUES (18446668621106209655);
SELECT CONCAT(MAX(a)) FROM t1;
returns -75452603341961.
mysql-test/r/func_group.result:
Test case for bug #11766094.
mysql-test/t/func_group.test:
Test case for bug #11766094.
sql/item.cc:
Bug #11766094 - 59132: MIN() AND MAX() REMOVE UNSIGNEDNESS
The Item_cache_int::val_str() method has been modified to
take into account the unsigned_flag value when converting
data to string.
Valgrind warning happens due to missing NULL value check in
Item::get_date. The fix is to add this check.
mysql-test/r/func_time.result:
test case
mysql-test/t/func_time.test:
test case
sql/item.cc:
added check for NULL value
Valgrind warning happens because null values check happens too late
in Item_func_month::val_str(after result string calculation).The fix
is to check null value before result string calculation.
mysql-test/r/func_time.result:
test case
mysql-test/t/func_time.test:
test case
sql/item_timefunc.h:
check null value before result string calculation.
ASSERTION TABLE->DB_STAT FAILED IN
SQL_BASE.CC::OPEN_TABLE() DURING I_S Q
This assert could be triggered if a statement requiring a name
lock on a table (e.g. DROP TRIGGER) executed concurrently
with an I_S query which also used the table.
One connection first started an I_S query that opened a given table.
Then another connection started a statement requiring a name lock
on the same table. This statement was blocked since the table was
in use by the I_S query. When the I_S query resumed and tried to
open the table again as part of get_all_tables(), it would encounter
a table instance with an old version number representing the pending
name lock. Since I_S queries ignore version checks and thus pending
name locks, it would try to continue. This caused it to encounter
the assert. The assert checked that the TABLE instance found with a
different version, was a real, open table. However, since this TABLE
instance instead represented a pending name lock, the check would
fail and trigger the assert.
This patch fixes the problem by removing the assert. It is ok for
TABLE::db_stat to be 0 in this case since the TABLE instance can
represent a pending name lock.
Test case added to lock_sync.test.
Issue:
======
Test case Correction for bug#11751148.
mysql-test/r/events_bugs.result:
Result file Correction for bug#11751148.
mysql-test/t/events_bugs.test:
Test case Correction for bug#11751148.
Valgrind warning happens due to missing NULL value check in
Item_func::val_decimal. The fix is to add this check.
mysql-test/r/func_time.result:
test case
mysql-test/t/func_time.test:
test case
sql/item_func.cc:
added check for NULL value
Valgrind warning happens due to uninitialized cached_format_type field
which is used later in Item_func_str_to_date::val_str method.
The fix is to init cached_format_type field.
mysql-test/r/func_time.result:
test case
mysql-test/t/func_time.test:
test case
sql/item_timefunc.cc:
init cached_format_type field
Assert fails due to overflow which happens in
Item_func_int_val::fix_num_length_and_dec() as
geometry functions have max_length value equal to
max_field_size(4294967295U). The fix is to skip
max_length calculation for some boundary cases.
mysql-test/r/func_math.result:
test case
mysql-test/t/func_math.test:
test case
sql/item_func.cc:
skip max_length calculation
if argument max_length is near max_field_size.
Assertion happens due to missing initialization of unsigned_flag
for Item_func_set_user_var object. It leads to incorrect
calculation of decimal field size.
The fix is to add initialization of unsigned_flag.
mysql-test/r/variables.result:
test case
mysql-test/t/variables.test:
test case
sql/item_func.cc:
add initialization of unsigned_flag.
Valgrind warining happens due to missing
'end of the string' check. The fix is to
check if we reached the end of the string.
mysql-test/r/func_time.result:
test case
mysql-test/t/func_time.test:
test case
sql/item_timefunc.cc:
check if we reached the end of
the string after leading spaces skipping.
Problem: mysqlbinlog --server-id may filter out Format_description_log_events.
If mysqlbinlog does not process the Format_description_log_event,
then mysqlbinlog cannot read the rest of the binary log correctly.
This can have the effect that mysqlbinlog crashes, generates an error,
or generates output that causes mysqld to crash, generate an error,
or corrupt data.
Fix: Never filter out Format_description_log_events. Also, never filter
out Rotate_log_events.
client/mysqlbinlog.cc:
Process Format_description_log_events even when the
server_id does not match the number given by --server-id.
mysql-test/t/mysqlbinlog.test:
Add test case.
ARE NOT BEING HONORED
max_allowed_packet works in conjunction with net_buffer_length.
max_allowed_packet is an upper bound of net_buffer_length.
So it doesn't make sense to set the upper limit lower than the value.
Added a warning (using ER_UNKNOWN_ERRROR and a specific message)
when this is done (in the log at startup and when setting either
max_allowed_packet or the net_buffer_length variables)
Added a test case.
Fixed several tests that broke the above rule.
The slave was not able to find the correct row in the innodb
table, because the row fetched from the innodb table would not
match the before image. This happened because the (don't care)
bytes in the NULLed fields would change once the row was stored
in the storage engine (from zero to the default value). This
would make bulk memory comparison (using memcmp) to fail.
We fix this by taking a preventing measure and avoiding memcmp
for tables that contain nullable fields. Therefore, we protect
the slave search routine from engines that return arbitrary
values for don't care bytes (in the nulled fields). Instead, the
slave thread will only check null_bits and those fields that are
not set to NULL when comparing the before image against the
storage engine row.
mysql-test/extra/rpl_tests/rpl_record_compare.test:
Added test case to the include file so that this is tested
with more than one engine.
mysql-test/suite/rpl/r/rpl_row_rec_comp_innodb.result:
Result update.
mysql-test/suite/rpl/r/rpl_row_rec_comp_myisam.result:
Result update.
mysql-test/suite/rpl/t/rpl_row_rec_comp_myisam.test:
Moved the include file last, so that the result from
BUG#11766865 is not intermixed with the result for
BUG#11760454.
sql/log_event.cc:
Skips memory comparison if the table has nullable
columns and compares only non-nulled fields in the
field comparison loop.
Issue:
------
Due to prefix match, database like 'k' was matching with 'ka' and events of 'ka' we getting displayed for 'show event' of 'k'.
Resolution:
-----------
Scan for listing of events in a schema is made to be done on exact match of database (schema) name instead of just prefix.
mysql-test/r/events_bugs.result:
modified expected file with the expected results.
mysql-test/t/events_bugs.test:
added a test case to reproduce the scenario.
sql/event_db_repository.cc:
Scan for schema name is made to be done on exact db name match.
There is a race between two threads: user thread and the dump
thread. The former sets a debug instruction that makes the latter wait
before processing an Xid event. There can be cases that the dump
thread has not yet processed the previous Xid event, causing it to
wait one Xid event too soon, thus causing sync_slave_with_master never
to resume.
We fix this by moving the instructions that set the debug variable
after calling sync_slave_with_master.
The problem was that server didn't check resulting size of prepared
statement argument which was set using mysql_send_long_data() API.
By calling mysql_send_long_data() several times it was possible
to create overly big string and thus force server to allocate
memory for it. There was no way to limit this allocation.
The solution is to add check for size of result string against
value of max_long_data_size start-up parameter. When intermediate
string exceeds max_long_data_size value an appropriate error message
is emitted.
We can't use existing max_allowed_packet parameter for this purpose
since its value is limited by 1GB and therefore using it as a limit
for data set through mysql_send_long_data() API would have been an
incompatible change. Newly introduced max_long_data_size parameter
gets value from max_allowed_packet parameter unless its value is
specified explicitly. This new parameter is marked as deprecated
and will be eventually replaced by max_allowed_packet parameter.
Value of max_long_data_size parameter can be set only at server
startup.
mysql-test/t/variables.test:
Added checking for new start-up parameter max_long_data_size.
sql/item.cc:
Added call to my_message() when accumulated string exceeds
max_long_data_size value. my_message() calls error handler
that was installed in mysql_stmt_get_longdata before call
to Item_param::set_longdata.
The error handler then sets state, last_error and last_errno
fields for current statement to values which correspond to
error which was caught.
sql/mysql_priv.h:
Added max_long_data_size variable declaration.
sql/mysqld.cc:
Added support for start-up parameter 'max_long_data_size'.
This parameter limits size of data which can be sent from
client to server using mysql_send_long_data() API.
sql/set_var.cc:
Added variable 'max_long_data_size' into list of variables
displayed by command 'show variables'.
sql/sql_prepare.cc:
Added error handler class Set_longdata_error_handler.
This handler is used to catch any errors that can be
generated during execution of Item_param::set_longdata().
Source code snippet that makes checking for statement's state
during statement execution is moved from Prepared_statement::execute()
to Prepared_statement::execute_loop() in order not to call
set_parameters() when statement has failed during
set_long_data() execution. If this hadn't been done
the call to set_parameters() would have failed.
tests/mysql_client_test.c:
A testcase for the bug #56976 was added.
Issue:
SSL_CIPHER set to a specific CIPHER name was not getting picked up by SHOW STATUS Command.
Solution:
If specific cipher name is specified, avoid overwriting of Cipher List with default Cipher names.
extra/yassl/src/yassl_int.cpp:
If user specified Cipher name is there, avoid populating default
cipher names' list.
mysql-test/r/ssl_cipher.result:
Expected file for ssl_cipher.test test case
mysql-test/t/ssl_cipher-master.opt:
Server option file for ssl_cipher.test test case.
mysql-test/t/ssl_cipher.test:
Test case to verify that user specified SSL cipher name is shown in SHOW STATUS Command.
Part 2. Function QUOTE() was not multi-byte safe.
@ mysql-test/r/ctype_ucs.result
@ mysql-test/t/ctype_ucs.test
Adding tests
@ sql/item_strfunc.cc
Fixing Item_func_quote::val_str to be multi-byte safe.
@ sql/item_strfunc.h
Multiple size needed for quote characters to mbmaxlen
Problem: wrong character set pointer was passed to my_strtoll10_mb2,
which led to DBUG_ASSERT failure in some cases.
@ mysql-test/r/func_encrypt_ucs2.result
@ mysql-test/t/func_encrypt_ucs2.test
@ mysql-test/r/ctype_ucs.result
@ mysql-test/t/ctype_ucs.test
Adding tests
@ sql/item_func.cc
"cs" initialization was wrong (res does not necessarily point to &str_value)
@ sql/item_strfunc.cc
Item_func_dec_encrypt::val_str() and Item_func_des_descrypt::val_str()
did not set character set for tmp_value (the returned value),
so the old value, which was previously copied from args[1]->val_str(),
was incorrectly returned with tmp_value.
ordered data from archive tables
Archive was using wrong memory address to check if field
is NULL (after filesort, when reading record again).
mysql-test/r/archive.result:
A test case for BUG#11764339.
mysql-test/t/archive.test:
A test case for BUG#11764339.
storage/archive/ha_archive.cc:
Null bytes are restored to "record" buffer, which may
or may not be equal to record buffer for field. Check
null bits in "record" buffer, instead of Field::null_ptr.
Problem: a byte behind the end of input string was read
in case of a broken XML not having a quote or doublequote
character closing a string value.
Fix: changing condition not to read behind the end of input string
@ mysql-test/r/xml.result
@ mysql-test/t/xml.test
Adding tests
@ strings/xml.c
When checking if the closing quote/doublequote was found,
using p->cur[0] us unsafe, as p->cur can point to the byte after the value.
Comparing p->cur to p->beg instead.
This is a backport of the patch for MySQL Bug#50574.
Adding a SPATIAL INDEX on non-geometrical columns caused a
segmentation fault when the table was subsequently
inserted into.
A test was added in mysql_prepare_create_table to explicitly
check whether non-geometrical columns are used in a
spatial index, and throw an error if so.
For MySQL 5.5 and later, a new and more meaningful error
message was introduced. For 5.1, we (re-)use an existing
error code.
("-") IN DATABASE NAMES IN ALTER DATABASE.
mysqldump did not quote database name in 'ALTER DATABASE'
statements in its output. This can further cause a failure
while loading if database name contains a hyphen '-'.
This happened as, while printing the 'ALTER DATABASE'
statements, the database name was not quoted.
Fixed by quoting the database name.
client/mysqldump.c:
Bug#11766310 : 59398: MYSQLDUMP 5.1 CAN'T HANDLE A DASH
("-") IN DATABASE NAMES IN ALTER DATABASE.
Modified the print statement in order to print the quoted
database name for 'ALTER DATABASE' statements.
mysql-test/r/mysqldump.result:
Added a test case for bug#11766310.
mysql-test/t/mysqldump.test:
Added a test case for bug#11766310.
The loop that was looping over subqueries' references to outer field used a
local boolean variable to tell whether the field was grouped or not. But the
implementor failed to reset the variable after each iteration. Thus a field
that was not directly aggregated appeared to be.
Fixed by resetting the variable upon each new iteration.
memory reference
There are two issues present here.
1) There is a possibility that we test a byte beyond the
allocated buffer
2) We compare a byte that might never have been
initalized to see if it's 0.
The first issue is not triggered by existing code, but an
ASSERT has been added to safe-guard against introducing
new code that triggers it.
The second issue is what triggers the Valgrind warnings
reported in the bug report. A buffer is allocated in
class String to hold the value. This buffer is populated
by the character data constituting the string, but is not
zero-terminated in most cases. Testing if it is indeed
zero-terminated means that we check a byte that has never
been explicitly set, thus causing Valgrind to trigger.
Note that issue 2 is not a serious problem. The variable
is read, and if it's not zero, we will set it to zero.
There are no further consequences.
Note that this patch does not fix the underlying problems
with issue 1, as it is deemed too risky to fix at this
point (as noted in the bug report). As discussed in
the report, the c_ptr() method should probably be
replaced, but this requires a thorough analysis of the
~200 calls to the method.
sql/set_var.cc:
These two cases have been reported to fail
with Valgrind.
attempt to create spatial index on char > 31 bytes".
Attempt to create spatial index on char field with length
greater than 31 byte led to assertion failure on server
compiled with safemutex support.
The problem occurred in mi_create() function which was called
to create a new version of table being altered. This function
failed since it detected an attempt to create a spatial key
on non-binary column and tried to return an error.
On its error path it tried to unlock THR_LOCK_myisam mutex
which has not been not locked at this point. Indeed such an
incorrect behavior was caught by safemutex wrapper and caused
assertion failure.
This patch fixes the problem by ensuring that mi_create()
doesn't releases THR_LOCK_myisam mutex on error path if it was
not acquired.
mysql-test/r/gis.result:
Added test for bug @59888 "debug assertion when attempt to
create spatial index on char > 31 bytes".
mysql-test/t/gis.test:
Added test for bug @59888 "debug assertion when attempt to
create spatial index on char > 31 bytes".
storage/myisam/mi_create.c:
Changed mi_create() not to release THR_LOCK_myisam mutex on
error path if it was not acquired.
rw_lock_create_func(): Initialize lock->writer_thread, so that Valgrind
will not complain even when Valgrind instrumentation is not enabled.
Flag lock->writer_thread uninitialized, so that Valgrind can complain
when it is used uninitialized.
rw_lock_set_writer_id_and_recursion_flag(): Revert the bogus Valgrind
instrumentation that was pushed in the first attempt to fix this bug.
Test failed on a certain Linux platform in automated environment. It turns out that this platform has an old version of Perl modules DBI and DBD::mysql installed, as well as the OS itself being relatively old.
Allowing error code 11 to be returned from mysqlhotcopy on expected error seems harmless and will make the test pass also with older libraries.
Added --debug-server and use $opt_debug_server where appropriate
Let --debug imply --debug-server
When merging to 5.5, must adapt fix for 59148
Oops, set debug => debug-server too late, fixed
privileges".
The first problem was that DROP USER didn't properly remove privileges
on stored functions from in-memory structures. So the dropped user
could have called stored functions on which he had privileges before
being dropped while his connection was still around.
Even worse if a new user with the same name was created he would
inherit privileges on stored functions from the dropped user.
Similar thing happened with old user name and function privileges
during RENAME USER.
This problem stemmed from the fact that the handle_grant_data() function
which handled DROP/RENAME USER didn't take any measures to update
in-memory hash with information about function privileges after
updating them on disk.
This patch solves this problem by adding code doing just that.
The second problem was that RENAME USER didn't properly update in-memory
structures describing table-level privileges and privileges on stored
procedures. As result such privileges could have been lost after a rename
(i.e. not associated with the new name of user) and inherited by a new
user with the same name as the old name of the original user.
This problem was caused by code handling RENAME USER in
handle_grant_struct() which [sic!]:
a) tried to update wrong (tables) hash when updating stored procedure
privileges for new user name.
b) passed wrong arguments to function performing the hash update and
didn't take into account the way in which such update could have
changed the order of the hash elements.
This patch solves this problem by ensuring that a) the correct hash
is updated, b) correct arguments are used for the hash_update()
function and c) we take into account possible changes in the order
of hash elements.
mysql-test/r/grant.result:
Added test coverage for bug#36544 "DROP USER does not remove stored
function privileges".
mysql-test/suite/funcs_1/r/innodb_storedproc_06.result:
Since after fixing bug#36544 "DROP USER does not remove stored function
privileges" in-memory structures are correctly updated by DROP USER,
DROP FUNCTION performed after DROP USER for its definer no longer
produces unwarranted warning/error messages.
mysql-test/suite/funcs_1/r/memory_storedproc_06.result:
Since after fixing bug#36544 "DROP USER does not remove stored function
privileges" in-memory structures are correctly updated by DROP USER,
DROP FUNCTION performed after DROP USER for its definer no longer
produces unwarranted warning/error messages.
mysql-test/suite/funcs_1/r/myisam_storedproc_06.result:
Since after fixing bug#36544 "DROP USER does not remove stored function
privileges" in-memory structures are correctly updated by DROP USER,
DROP FUNCTION performed after DROP USER for its definer no longer
produces unwarranted warning/error messages.
mysql-test/t/grant.test:
Added test coverage for bug#36544 "DROP USER does not remove stored
function privileges".
sql/sql_acl.cc:
Changed handle_grant_data() to also update hash with function
privileges. This allows DROP/RENAME USER correctly keep this
in-memory structure up-to-date.
To do this extended handle_grant_struct() to support updating of this
hash. In addition fixed code in this function which is responsible for
handling of column and routine hashes during RENAME USER, ensured that
we correctly update these hashes after changing user name and that we
don't skip elements while iterating through the hash and doing updates.
Also fix bug#59110: Memory leak of QUICK_SELECT_I allocated memory.
Includes Jørgen Lølands review comments.
Root cause of these bugs are that test_if_skip_sort_order() decided to
revert the 'skip_sort_order' descision (and use filesort) after the
query plan has been updated to reflect a 'skip' of the sort order.
This might happen in 'check_reverse_order:' if we have a
select->quick which could not be made descending by appending
a QUICK_SELECT_DESC. ().
The original 'save_quick' was then restored after the QEP has been modified,
which caused:
- An incorrect 'precomputed_group_by= TRUE' may have been set,
and not reverted, as part of the already modifified QEP (Bug#59308)
- A 'select->quick' might have been created which we fail to delete (bug#59110).
This fix is a refactorication of test_if_skip_sort_order() where all logic
related to modification of QEP (controlled by argument 'bool no_changes'), is
moved to the end of test_if_skip_sort_order(), and done after *all* 'test_if_skip'
checks has been performed - including the 'check_reverse_order:' checks.
The refactorication above contains now intentional changes to the logic which
has been moved to the end of the function.
Furthermore, a smaller part of the fix address the handling of the
select->quick objects which may already exists when we call
'test_if_skip_sort_order()' (save_quick) -and
new select->quick's created during test_if_skip_sort_order():
- Before new select->quick may be created by calling ::test_quick_select(), we
set 'select->quick= 0' to avoid that ::test_quick_select() prematurely
delete the save_quick's. (After this call we may have both a 'save_quick'
and 'select->quick')
- All returns from ::test_if_skip_sort_order() where we may have both a
'save_quick' and a 'select->quick' has been changed to goto's to the
exit points 'skiped_sort_order:' or 'need_filesort:' where we
decide which of the QUICK_SELECT's to keep, and delete the other.
if the standard input is a directory.
The problem is that mysql monitor try to read from stdin without
checking input source type.
The solution is to stop reading data from standard input if a call
to read(2) failed.
A new test case was added into mysql.test.
client/my_readline.h:
Data members error and truncated was added to LINE_BUFFER structure.
These data members used instead of out parameters in functions
batch_readline, intern_read_line.
client/mysql.cc:
read_and_execute() was modified: set status.exit_status to 1
when the error occured while reading the next command line in
non-interactive mode. Also the value of the truncated attribute
of structure LINE_BUFF is taken into account only for non-iteractive mode.
client/readline.cc:
intern_read_line() was modified: cancel reading from input if
fill_buffer() returns -1, e.g. if call to read failed.
batch_readline was modified: set the error data member of LINE_BUFFER
structure to value of my_errno when system error happened during call
to my_read/my_realloc.
mysql-test/t/mysql.test:
Test for bug#57450 was added.
handling.
The problem was that parsing of nested regular expression involved
recursive calls. Such recursion didn't take into account the amount of
available stack space, which ended up leading to stack overflow crashes.
mysql-test/t/not_embedded_server.test:
Added test for bug#58026.
regex/my_regex.h:
added pointer to function as last argument of my_regex_init() for check
enough memory in stack.
regex/regcomp.c:
p_ere() was modified: added call to function for check enough memory
in stack. Function for check available stack space specified by
global variable my_regex_enough_mem_in_stack. This variable set to
NULL for embedded mysqld and to a pointer to function
check_enough_stack_size otherwise.
regex/reginit.c:
my_regex_init was modified: pass a pointer to a function for check
enough memory in stack space. Reset this pointer to NULL in my_regex_end.
sql/mysqld.cc:
Added function check_enough_stack_size() for check enough memory in stack.
Passed this function as second argument to my_regex_init. For embedded
mysqld passed NULL as second argument.
There is one part of the test case that needs to break
and re-establish the circular topology. For this the test
stops the slave threads on a couple of servers and restarts
them with START SLAVE. However, no check is done on the
status of the IO or SQL threads before proceeding with
the subsequent commands.
Because rpl_only_running_threads is set to 1 this can lead
to silently not syncing all slave threads as expected,
ultimately resulting in unexpected results (and consequently
on a failing test run).
We fix this by replacing the START SLAVE instructions with
calls to --source include/start_slave.inc, which will wait
for the slave threads to be running (show 'Yes' in
Slave_IO|SQL_Running fields of SHOW SLAVE STATUS) before
proceeding. Additionally, we change rpl_sync.inc to make the
IO thread report that it is running when its running status
is any other than 'No'.
Bug #55755 : Date STD variable signness breaks server on FreeBSD and OpenBSD
* Added a check to configure on the size of time_t
* Created a macro to check for a valid time_t that is safe to use with datetime
functions and store in TIMESTAMP columns.
* Used the macro consistently instead of the ad-hoc checks introduced by 52315
* Fixed compliation warnings on platforms where the size of time_t is smaller than
the size of a long (e.g. OpenBSD 4.8 64 amd64).
Bug #52315: utc_date() crashes when system time > year 2037
* Added a correct check for the timestamp range instead of just variable size check to
SET TIMESTAMP.
* Added overflow checking before converting to time_t.
* Using a correct localized error message in this case instead of the generic error.
* Added a test suite.
* fixed the checks so that they check for unsigned time_t as well. Used the checks
consistently across the source code.
* fixed the original test case to expect the new error code.
primary_key_no == 0".
Attempt to create InnoDB table with non-nullable column of
geometry type having an unique key with length 12 on it and
with some other candidate key led to server crash due to
assertion failure in both non-debug and debug builds.
The problem was that such a non-candidate key could have
been sorted as the first key in table/.FRM, before any legit
candidate keys. This resulted in assertion failure in InnoDB
engine which assumes that primary key should either be the
first key in table/.FRM or should not exist at all.
The reason behind such an incorrect sorting was an wrong
value of Create_field::key_length member for geometry field
(which was set to its pack_length == 12) which confused code
in mysql_prepare_create_table(), so it would skip marking
such key as a key with partial segments.
This patch fixes the problem by ensuring that this member
gets the same value of Create_field::key_length member as
for other blob fields (from which geometry field class is
inherited), and as result unique keys on geometry fields
are correctly marked as having partial segments.
mysql-test/include/gis_keys.inc:
Added test case for bug #58650 "Failing assertion:
primary_key_no == -1 || primary_key_no == 0".
mysql-test/r/gis.result:
Added test case for bug #58650 "Failing assertion:
primary_key_no == -1 || primary_key_no == 0".
mysql-test/suite/innodb/r/innodb_gis.result:
Added test case for bug #58650 "Failing assertion:
primary_key_no == -1 || primary_key_no == 0".
mysql-test/suite/innodb_plugin/r/innodb_gis.result:
Added test case for bug #58650 "Failing assertion:
primary_key_no == -1 || primary_key_no == 0".
sql/field.cc:
Changed Create_field::create_length_to_internal_length() to
correctly set Create_field::key_length member for geometry
fields. Similar to the blob types key_length for such fields
should be the same as length and not field's packed length
(which is always 12 for geometry).
As result of this change code handling table creation now
always correctly identifies btree/unique keys on geometry
fields as partial keys, so such keys can't be erroneously
treated as candidate keys and sorted in keys array in .FRM
before legit candidate keys.
This fixes bug #58650 "Failing assertion: primary_key_no ==
-1 || primary_key_no == 0" in which incorrect candidate key
sorting led to assertion failure in InnoDB code.
Root cause for this bug is that the optimizer try to detect&
optimize the special case:
'<field> BETWEEN c1 AND c1' and handle this as the condition '<field> = c1'
This was implemented inside add_key_field(.. *field, *value[]...)
which assumed field to refer key Field, and value[] to refer a [low...high]
constant pair. value[0] and value[1] was then compared for equality.
In a 'normal' BETWEEN condition of the form '<field> BETWEEN val1 and val2' the
BETWEEN operation is represented with an argementlist containing the
values [<field>, val1, val2] - add_key_field() is then called with
parameters field=<field>, *value=val1.
However, if the BETWEEN predicate specified:
1) '<const1> BETWEEN<const2> AND<field>
the 'field' and 'value' arguments to add_key_field() had to be swapped.
This was implemented by trying to cheat add_key_field() to handle it like:
2) '<const1> GE<const2> AND<const1> LE<field>'
As we didn't really replace the BETWEEN operation with 'ge' and 'le',
add_key_field() still handled it as a 'BETWEEN' and compared the (swapped)
arguments<const1> and<const2> for equality. If they was equal, the
condition 1) was incorrectly 'optimized' to:
3) '<field> EQ <const1>'
This fix moves this optimization of '<field> BETWEEN c1 AND c1' into
add_key_fields() which then calls add_key_equal_fields() to collect
key equality / comparison for the key fields in the BETWEEN condition.
