This deadlock occurs when a client issues a HANDLER ... OPEN statement
that tries to open a table that has a pending name-lock on it by another
client that also needs a name-lock on some other table which is already
open and associated to a HANDLER instance owned by the first client.
The deadlock happens because the open_table() function will back-off
and wait until the name-lock goes away, causing a circular wait if some
other name-lock is also pending for one of the open HANDLER tables.
Such situation, for example, can be easily repeated by issuing a RENAME
TABLE command in such a way that the existing table is already open
as a HANDLER table by another client and this client tries to open
a HANDLER to the new table name.
The solution is to allow handler tables with older versions (marked for
flush) to be closed before waiting for the name-lock completion. This is
safe because no other name-lock can be issued between the flush and the
check for pending name-locks.
The test case for this bug is going to be committed into 5.1 because it
requires a test feature only avaiable in 5.1 (wait_condition).
sql/sql_base.cc:
Improve comments in the open_table() function, stating the importance
of the handler tables flushing for the back-off process.
sql/sql_handler.cc:
Allows handler tables flushes when opening new tables in order to avoid
potential deadlocks. Add comments explaining the importance of the flush.
Even though it returns NULL, the MAKETIME function did not have this property set,
causing a failed assertion (designed to catch exactly this).
Fixed by setting the nullability property of MAKETIME().
mysql-test/r/func_sapdb.result:
Bug#31160: Changed test result.
mysql-test/r/func_time.result:
Bug#31160: Test result.
mysql-test/t/func_time.test:
Bug#31160: Test case.
sql/item_timefunc.h:
Bug#31160: The fix: Initializing maybe_null to true
bitmap_is_set(table->write_set, fiel
Problem: creating a temporary table we allocate the group buffer if needed
followed by table bitmaps (see create_tmp_table()). Reserving less memory for
the group buffer than actually needed (used) for values retrieval may lead
to overlapping with followed bitmaps in the memory pool that in turn leads
to unpredictable consequences.
As we use Item->max_length sometimes to calculate group buffer size,
it must be set to proper value. In this particular case
Item_datetime_typecast::max_length is too small.
Another problem is that we use max_length to calculate the group buffer
key length for items represented as DATE/TIME fields which is superfluous.
Fix: set Item_datetime_typecast::max_length properly,
accurately calculate the group buffer key length for items
represented as DATE/TIME fields in the buffer.
mysql-test/r/type_datetime.result:
Fix for bug #31249: Assertion `!table || (!table->write_set ||
bitmap_is_set(table->write_set, fiel
- test result.
mysql-test/t/type_datetime.test:
Fix for bug #31249: Assertion `!table || (!table->write_set ||
bitmap_is_set(table->write_set, fiel
- test case.
sql/item_timefunc.h:
Fix for bug #31249: Assertion `!table || (!table->write_set ||
bitmap_is_set(table->write_set, fiel
- set Item_datetime_typecast::max_length properly.
sql/sql_select.cc:
Fix for bug #31249: Assertion `!table || (!table->write_set ||
bitmap_is_set(table->write_set, fiel
- the group buffer key length for items represented as
DATE/TIME fields in the buffer should be calculated using
the maximum pack length of such fields (== 8), using
max_length here is redundant.
Completion of previous patch. Negative number were denied
as the second argument to NAME_CONST.
mysql-test/r/func_misc.result:
Bug#30832 completion: test result
mysql-test/t/func_misc.test:
Bug#30832 completion: test case
added negative numbers
sql/item_func.h:
Bug#30832 completion
The function that represents unary minus is considered a constant if the argument is.
Item_sum_distinct::setup(THD*): Assertion
There was an assertion to detect a bug in ROLLUP
implementation. However the assertion is not true
when used in a subquery context with non-cacheable
statements.
Fixed by turning the assertion to accepted case
(just like it's done for the other aggregate functions).
mysql-test/r/func_group.result:
Bug #31156: test case
mysql-test/t/func_group.test:
Bug #31156: test case
sql/item_sum.cc:
Bug #31156: make it OK to call setup() several times:
done for (e.g.) scalar subquery
into bodhi.(none):/opt/local/work/mysql-5.0-runtime
mysql-test/r/sp.result:
Auto merged
mysql-test/t/sp.test:
Auto merged
sql/sql_parse.cc:
Auto merged
Our web server has been restructured several times, and references
to it in our source code has stayed the same. This patch from Paul
DuBois updates all URLs to modern semantics.
debian/po/ca.po:
Change URLs.
debian/po/cs.po:
Change URLs.
debian/po/da.po:
Change URLs.
debian/po/gl.po:
Change URLs.
debian/po/ja.po:
Change URLs.
debian/po/pt_BR.po:
Change URLs.
debian/po/sv.po:
Change URLs.
debian/po/tr.po:
Change URLs.
mysql-test/lib/mtr_report.pl:
Change URLs.
mysql-test/mysql-test-run-shell.sh:
Change URLs.
ndb/include/ndbapi/Ndb.hpp:
Change URLs.
netware/mysql_test_run.c:
Change URLs.
scripts/mysqld_safe.sh:
Change URLs.
sql/mysqld.cc:
Change URLs.
convert(<invalid time>, datetime) in WHERE caused crash as function
returned (void*)NULL, but did not flag SQL NULL. It does now.
mysql-test/r/type_datetime.result:
show that convert() to datetime in a WHERE-clause will no longer crash
the server on invalid input.
mysql-test/t/type_datetime.test:
show that convert() to datetime in a WHERE-clause will no longer crash
the server on invalid input.
sql/item.cc:
When failing to process time values, do not just return (void*)NULL,
but flag SQL NULL as well so no caller will try to process the 0x0L.
This makes behaviour for Item::Item::val_decimal_from_date() and
Item::Item::val_decimal_from_time() identical.
Reserve the space for the leading 0 (before the decimal point) in DECIMAL(a,a) ZEROFILL.
mysql-test/r/type_decimal.result:
show that we allocate a large enough buffer for output of DECIMAL(a,a) [ZEROFILL].
mysql-test/t/type_decimal.test:
show that we allocate a large enough buffer for output of DECIMAL(a,a) [ZEROFILL].
without patch for bug#31227, valgrind will complain here; so will a
debug build.
sql/my_decimal.cc:
Reserve the space for the leading 0 (before the decimal point) in DECIMAL(a,a) ZEROFILL.
Doxygenise preamble of my_decimal2string().
This bug is a symptom of the way handler's tables are managed. The
most different aspect, compared to the conventional behavior, is that
the handler's tables are long lived, meaning that their lifetimes are
not bounded by the duration of the command that opened them. For this
effect the handler code uses its own list (handler_tables instead of
open_tables) to hold open handler tables so that the tables won't be
closed at the end of the command/statement. Besides the handler_tables
list, there is a hash (handler_tables_hash) which is used to associate
handler aliases to tables and to refresh the tables upon demand (flush
tables).
The current implementation doesn't work properly with refreshed tables
-- more precisely when flush commands are issued by other initiators.
This happens because when a handler open or read statement is being
processed, the associated table has to be opened or locked and, for this
matter, the open_tables and handler_tables lists are swapped so that the
new table being opened is inserted into the handler_tables list. But when
opening or locking the table, if the refresh version is different from the
thread refresh version then all used tables in the open_tables list (now
handler_tables) are refreshed. In the "refreshing" process the handler
tables are flushed (closed) without being properly unlinked from the
handler hash.
The current implementation also fails to properly discard handlers of
dropped tables, but this and other problems are going to be addressed
in the fixes for bugs 31397 and 31409.
The chosen approach tries to properly save and restore the table state
so that no table is flushed during the table open and lock operations.
The logic is almost the same as before with the list swapping, but with
a working glue code.
The test case for this bug is going to be committed into 5.1 because it
requires a test feature only avaiable in 5.1 (wait_condition).
sql/sql_handler.cc:
Backup and reset the open_tables and handler_table lists when opening
a new handler table and merge the opened table into the handler_tables
list afterwards. When locking, do the same but possibly close the table
if a refresh is pending.
This actually, fix for the patch for bug-27354. The problem with
the patch was that Item_func_sp::used_tables() was updated, but
Item_func_sp::const_item() was not. So, for Item_func_sp, we had
the following inconsistency:
- used_tables() returned RAND_TABLE, which means that the item
can produce "random" results;
- but const_item() returned TRUE, which means that the item is
a constant one.
The fix is to change Item_func_sp::const_item() behaviour: it must
return TRUE (an item is a constant one) only if a stored function
is deterministic and each of its arguments (if any) is a constant
item.
mysql-test/r/sp.result:
Update result file.
mysql-test/t/sp.test:
Add test cases for BUG-31035, BUG-31191, BUG-31226
sql/item_func.cc:
Make Item::const_item() be consistent with Item::used_tables().
added check for hidden I_S tables for 'show columns|keys' commands
mysql-test/r/information_schema.result:
test result
mysql-test/t/information_schema.test:
test case
sql/sql_parse.cc:
added check for hidden I_S tables for 'show columns|keys' commands
Bug#28878: InnoDB tables with UTF8 character set and indexes cause wrong result for DML
When making key reference buffers over CHAR fields whitespace (0x20) must be used to fill in the remaining space in the field's buffer. This is what Field_string::store() does. Fixed Field_string::get_key_image() to do the same.
mysql-test/r/innodb_mysql.result:
Bug#28878: test case
mysql-test/t/innodb_mysql.test:
Bug#28878: test case
sql/field.cc:
Bug#28878: Fill with space instead of binary zeros.
This is for bug #29446 "Specifying a myisam_sort_buffer > 4GB on 64 bit machines not possible". Support for myisam_sort_buffer_size > 4 GB on 64-bit Windows will be looked at later in 5.2.
mysql-test/r/variables.result:
Fixed the test.
mysql-test/t/variables.test:
Fixed the test.
mysys/my_getopt.c:
Print a warning to the error log if a user sets an option to a value greater than the option's maximum value.
sql/set_var.cc:
Issue an SQL warning if a user assignes a value greater than the variable's maximum value.
The function str_to_date has a field to say whether it's invoked constant
arguments. But this member was not initialized, causing the function to
think that it could use a cache of the format type when said cache was in
fact not initialized.
Fixed by initializing the field to false.
mysql-test/r/type_date.result:
Bug#30942: Test result
mysql-test/t/type_date.test:
Bug#30942: Test case
sql/item_timefunc.h:
Bug#30942: Initialized const_item to false
We use get_geometry_type() call to decide the exact type
of a geometry field to be created (POINT, POLYGON etc)
Though this function was only implemented for few items.
In the bug's case we need to call this function for the
Item_sum instance, where it was not implemented, what is
the reason of the crash.
Fixed by implementing virtual Item::get_geometry_type(),
so it can be called for any Item.
sql/item.h:
Bug #31155 gis types in union'd select cause crash.
virtual Item::geometry_type() added instead of
various geometry_type() fucntions.
sql/item.cc:
Bug #31155 gis types in union'd select cause crash.
Unified virtual ::get_geometry_type() function used
sql/item_geofunc.cc:
Bug #31155 gis types in union'd select cause crash.
virtual Item::geometry_type() implemented for geo-Items.
Mostly previous ::get_geometry_type() implementation changed
sql/item_geofunc.h:
Bug #31155 gis types in union'd select cause crash.
get_geometry_type() declarations unified
end-of-line check missed in Gis_read_stream::get_next_word,
what can lead to crashes (expecially with NULL strings).
End-of-line check added
sql/gstream.cc:
Bug #30955 geomfromtext() crasher
mysql-test/r/gis.result:
Bug #30955 geomfromtext() crasher.
test result
mysql-test/t/gis.test:
Bug #30955 geomfromtext() crasher.
test case
Add --skip-innodb-adaptive-hash-index option, which is a way to
work around the bug (by disabling the adaptive hash feature entirely).
This may be useful even once the bug is fixed, for benchmarking purposes.
There are some workloads for which the adaptive hash index is not effective.
sql/ha_innodb.cc:
Add --skip-innodb-adaptive-hash-index option.
sql/ha_innodb.h:
Add --skip-innodb-adaptive-hash-index option.
sql/mysqld.cc:
Add --skip-innodb-adaptive-hash-index option.
sql/set_var.cc:
Add --skip-innodb-adaptive-hash-index option.
The Item_func_rollup_const class is used for wrapping constants to avoid
wrong result for ROLLUP queries with DISTINCT and a constant in the select
list. This class is also used to wrap up a NULL constant but its null_value
wasn't set accordingly. This led to a server crash.
Now the null_value of an object of the Item_func_rollup_const class is set
by its fix_length_and_dec member function.
mysql-test/t/olap.test:
Added a test case for the bug#31095: Unexpected NULL constant caused server crash.
mysql-test/r/olap.result:
Added a test case for the bug#31095: Unexpected NULL constant caused server crash.
sql/item_func.h:
Bug#31095: Unexpected NULL constant caused server crash.
Now the null_value of an object of the Item_func_rollup_const class is set
by its fix_length_and_dec member function.
