Fixed a possible problem with reading of dynamic records
when a write cache is active. The cache must be flushed
whenever a part of the file in the write cache is to be
read.
Added a read optimization to _mi_read_dynamic_record().
No test case. This was a hypothetical but existing problem.
myisam/mi_dynrec.c:
Bug#20719 - Reading dynamic records with write buffer could fail
Fixed a possible problem with reading of dynamic records
when a write cache is active. The cache must be flushed
whenever a part of the file in the write cache is to be
read. This must be done before the read of the header
and before the read of the rest block.
Renamed the 'flag' and 'skipp_deleted_blocks' variables.
Added a read optimization to _mi_read_dynamic_record()
that was present in _mi_read_rnd_dynamic_record() already.
After _mi_get_block_info() we have some bytes of the record
in the header buffer already. No need to read them again.
into dl145k.mysql.com:/data0/mkindahl/bk/MERGE/mysql-5.0-merge
ndb/include/kernel/GlobalSignalNumbers.h:
Auto merged
ndb/src/kernel/blocks/dbdict/Dbdict.cpp:
Auto merged
ndb/src/kernel/blocks/dbdict/Dbdict.hpp:
Auto merged
ndb/src/kernel/blocks/dbdih/DbdihMain.cpp:
Auto merged
ndb/src/ndbapi/ndberror.c:
Auto merged
ndb/src/kernel/blocks/dbdict/DictLock.txt:
wait until SL_STARTED before sending DICT_UNLOCK_ORD
ndb/src/kernel/blocks/dbdih/Dbdih.hpp:
wait until SL_STARTED before sending DICT_UNLOCK_ORD
ndb/src/kernel/blocks/dbdih/DbdihMain.cpp:
wait until SL_STARTED before sending DICT_UNLOCK_ORD
ndb/src/kernel/vm/SimulatedBlock.cpp:
wait until SL_STARTED before sending DICT_UNLOCK_ORD
ndb/src/kernel/vm/SimulatedBlock.hpp:
wait until SL_STARTED before sending DICT_UNLOCK_ORD
ndb/test/run-test/daily-basic-tests.txt:
wait until SL_STARTED before sending DICT_UNLOCK_ORD
closing temp tables through end_thread
had a flaw in binlog-off branch of close_temporary_tables where
next table to close was reset via table->next
for (table= thd->temporary_tables; table; table= table->next)
which was wrong since the current table instance got destoyed at
close_temporary(table, 1);
The fix adapts binlog-on branch method to engage the loop's internal 'next' variable which holds table->next prior table's destoying.
sql/sql_base.cc:
no-binlog branch is fixed: scanning across temporary_tables must be careful to save next table since the current is being destroyed inside of close_temporary.
binlog-is-open case is ok.
a too large value": the bug was that if MySQL generated a value for an
auto_increment column, based on auto_increment_* variables, and this value
was bigger than the column's max possible value, then that max possible
value was inserted (after issuing a warning). But this didn't honour
auto_increment_* variables (and so could cause conflicts in a master-master
replication where one master is supposed to generated only even numbers,
and the other only odd numbers), so now we "round down" this max possible
value to honour auto_increment_* variables, before inserting it.
mysql-test/r/rpl_auto_increment.result:
result update. Before the fix, the result was that master inserted 127 in t1
(which didn't honour auto_increment_* variables!),
instead of failing with "duplicate key 125" like now.
mysql-test/t/rpl_auto_increment.test:
Test for BUG#20524 "auto_increment_* not observed when inserting
a too large value".
We also check the pathological case (table t2) where it's impossible to
"round down".
The fixer of BUG#20573 will be able to use table t2 for testing his fix.
sql/handler.cc:
If handler::update_auto_increment() generates a value larger than the field's
max possible value, we used to simply insert this max possible value
(after pushing a warning). Now we "round down" this max possible value to
honour auto_increment_* variables (if at all possible), before trying the
insertion.
auto_increment breaks binlog":
if slave's table had a higher auto_increment counter than master's (even
though all rows of the two tables were identical), then in some cases,
REPLACE and INSERT ON DUPLICATE KEY UPDATE failed to replicate
statement-based (it inserted different values on slave from on master).
write_record() contained a "thd->next_insert_id=0" to force an adjustment
of thd->next_insert_id after the update or replacement. But it is this
assigment introduced indeterminism of the statement on the slave, thus
the bug. For ON DUPLICATE, we replace that assignment by a call to
handler::adjust_next_insert_id_after_explicit_value() which is deterministic
(does not depend on slave table's autoinc counter). For REPLACE, this
assignment can simply be removed (as REPLACE can't insert a number larger
than thd->next_insert_id).
We also move a too early restore_auto_increment() down to when we really know
that we can restore the value.
mysql-test/r/rpl_insert_id.result:
result update, without the bugfix, slave's "3 350" were "4 350".
mysql-test/t/rpl_insert_id.test:
test for BUG#20188 "REPLACE or ON DUPLICATE KEY UPDATE in
auto_increment breaks binlog".
There is, in this order:
- a test of the bug for the case of REPLACE
- a test of basic ON DUPLICATE KEY UPDATE functionality which was not
tested before
- a test of the bug for the case of ON DUPLICATE KEY UPDATE
sql/handler.cc:
the adjustment of next_insert_id if inserting a big explicit value, is
moved to a separate method to be used elsewhere.
sql/handler.h:
see handler.cc
sql/sql_insert.cc:
restore_auto_increment() means "I know I won't use this autogenerated
autoincrement value, you are free to reuse it for next row". But we were
calling restore_auto_increment() in the case of REPLACE: if write_row() fails
inserting the row, we don't know that we won't use the value, as we are going to
try again by doing internally an UPDATE of the existing row, or a DELETE
of the existing row and then an INSERT. So I move restore_auto_increment()
further down, when we know for sure we failed all possibilities for the row.
Additionally, in case of REPLACE, we don't need to reset THD::next_insert_id:
the value of thd->next_insert_id will be suitable for the next row.
In case of ON DUPLICATE KEY UPDATE, resetting thd->next_insert_id is also
wrong (breaks statement-based binlog), but cannot simply be removed, as
thd->next_insert_id must be adjusted if the explicit value exceeds it.
We now do the adjustment by calling
handler::adjust_next_insert_id_after_explicit_value() (which, contrary to
thd->next_insert_id=0, does not depend on the slave table's autoinc counter,
and so is deterministic).
into mysql.com:/home/mydev/mysql-5.0-ateam
myisam/mi_key.c:
Auto merged
mysql-test/r/gis-rtree.result:
Auto merged
mysql-test/t/gis-rtree.test:
Auto merged
myisam/mi_check.c:
SCCS merged
into mysql.com:/home/mydev/mysql-5.0-ateam
libmysqld/libmysqld.c:
Auto merged
myisam/mi_rkey.c:
Auto merged
mysql-test/r/func_sapdb.result:
Auto merged
mysql-test/r/symlink.result:
Auto merged
mysql-test/t/func_sapdb.test:
Auto merged
scripts/make_binary_distribution.sh:
Auto merged
sql/item_geofunc.h:
Auto merged
sql/item_timefunc.cc:
Auto merged
sql/sql_class.cc:
Auto merged
sql/sql_parse.cc:
Auto merged
libmysqld/lib_sql.cc:
Manual merge
mysql-test/r/func_time.result:
Manual merge
mysql-test/r/gis.result:
Manual merge
mysql-test/t/func_time.test:
Manual merge
mysql-test/t/gis.test:
Manual merge
sql-common/client.c:
Manual merge
into mysql.com:/home/mydev/mysql-5.0-ateam
myisam/mi_create.c:
Auto merged
mysql-test/r/ctype_utf8.result:
Auto merged
mysql-test/r/key.result:
Auto merged
mysql-test/r/myisam.result:
Auto merged
mysql-test/t/ctype_utf8.test:
Auto merged
mysql-test/t/key.test:
Auto merged
mysql-test/t/myisam.test:
Auto merged
sql/opt_sum.cc:
Auto merged
sql/table.cc:
Auto merged
support-files/mysql.spec.sh:
Auto merged
sql/field.cc:
Manual merge
into mysql.com:/home/mydev/mysql-4.1-bug14400
myisam/mi_rkey.c:
Bug#14400 - Query joins wrong rows from table which is subject of "concurrent insert"
Manual merge
sql/sql_class.cc:
Bug#14400 - Query joins wrong rows from table which is subject of "concurrent insert"
Manual merge
Enabling rpl_openssl.test for Windows to check that
currently it still hangs (because I can't reproduce
this on my machine).
mysql-test/t/rpl_openssl.test:
Enabling rpl_openssl.test for Windows
It was possible that fetching a record by an exact key value
(including the record pointer) could return a record with a
different key value. This happened only if a concurrent insert
added a record with the searched key value after the fetching
statement locked the table for read.
The search succeded on the key value, but the record was
rejected as it was past the file length that was remembered
at start of the fetching statement. With other words it was
rejected as being a concurrently inserted record.
The action to recover from this problem was to fetch the
record that is pointed at by the next key of the index.
This was repeated until a record below the file length was
found.
I do now avoid this loop if an exact match was searched.
If this match is beyond the file length, it is now treated
as "key not found". There cannot be another key with the
same record pointer.
myisam/mi_rkey.c:
Bug#14400 - Query joins wrong rows from table which is subject of "concurrent insert"
Added a check for exact key match before searching for
the next key that was not concurrently inserted. If an
exact key match finds a concurrently inserted row, this
must be treated as "key not found".
sql/sql_class.cc:
Bug#14400 - Query joins wrong rows from table which is subject of "concurrent insert"
Fixed some DBUG_ENTER strings.
A corrupt table with dynamic record format can crash the
server when trying to select from it.
I fixed the crash that resulted from the particular type
of corruption that has been reported for this bug.
No test case. To test it, one needs a table with a very special
corruption. The bug report contains a file with such a table.
myisam/mi_dynrec.c:
Bug#19835 - Binary copy of corrupted tables crash the server when issuing a query
Added a protection against corrupted records. A dynamic
record header with invalid 'next' pointer could trigger
the assert in _mi_get_block_info(). Now I avoid this by
reporting a corruption error.
In the Windows build files, the "Max nt" configuration for some reason
had the mysql_client_test project disabled. Enable it.
VC++Files/mysql.sln:
The "Max nt" configuration for some reason had the mysql_client_test
project disabled. Enable it.
CHECK TABLE could complain about a fully intact spatial index.
A wrong comparison operator was used for table checking.
The result was that it checked for non-matching spatial keys.
This succeeded if at least two different keys were present,
but failed if only the matching key was present.
I fixed the key comparison.
myisam/mi_check.c:
Bug#17877 - Corrupted spatial index
Fixed the comparison operator for checking a spatial index.
Using MBR_EQUAL | MBR_DATA to compare for equality and
include the data pointer in the comparison. The latter
finds the index entry that points to the current record.
This is necessary for non-unique indexes.
The old operator, SEARCH_SAME, is unknown to the rtree
search functions and handled like MBR_DISJOINT.
myisam/mi_key.c:
Bug#17877 - Corrupted spatial index
Added a missing DBUG_RETURN.
myisam/rt_index.c:
Bug#17877 - Corrupted spatial index
Included the data pointer in the copy of the search key.
This is necessary for searching the index entry that points
to a specific record if the search_flag contains MBR_DATA.
myisam/rt_mbr.c:
Bug#17877 - Corrupted spatial index
Extended the RT_CMP() macro with an assert for an
unexpected comparison operator.
mysql-test/r/gis-rtree.result:
Bug#17877 - Corrupted spatial index
The test result.
mysql-test/t/gis-rtree.test:
Bug#17877 - Corrupted spatial index
The test case.
Improved definition of mysys configuration for -nt builds.
VC++Files/mysql.sln:
Use the name 'nt' instead of 'Release' for configuration.
VC++Files/mysys/mysys.vcproj:
Use the name 'nt' instead of 'Release' for configuration.
Use separate output files for NT and non-NT configurations.
Make sure for the mysys project that __NT__ is defined in *nt solution
configurations (but not in other configurations).
VC++Files/mysql.sln:
Define __NT__ in mysys for *nt configurations.
VC++Files/mysys/mysys.vcproj:
Add configurations with __NT__ defined.
mysql-test/mysql-test-run.pl:
Also allow testing a "Max nt" build.
and BUG#19208 "Test 'rpl000017' hangs on Windows".
Both bugs are caused by attempting to delete an opened
file and to create immediatedly a new one with the same
name. On Windows it can be supported only on NT-platforms
(by using FILE_SHARE_DELETE mode and with renaming the
file before deletion). Because deleting not-closed files
is not supported on all platforms (e.g. Win 98|ME) this
is to be considered harmful and should be eliminated by
a "code redesign".
VC++Files/mysys/mysys.vcproj:
To be sure that __NT__ is defined for Win configurations.
Temporary, to be changed in more appropriate way.
include/my_sys.h:
Adding my_delete_allow_opened to be invoked to delete
a (possibly) not closed file on Windows NT-platforms.
mysys/my_delete.c:
Adding nt_share_delete() function implementing
a (possibly) not closed file deletion on Windows NT.
sql/log.cc:
MYSQL_LOG::reset_logs(): Deleting usually not
closed binlog files.
Produce a warning if DATA/INDEX DIRECTORY is specified in
ALTER TABLE statement.
Ignoring of these options is documented in the symbolic links
section of the manual.
mysql-test/r/symlink.result:
Modified test result according to fix for BUG#1662.
sql/sql_parse.cc:
Produce a warning if DATA/INDEX DIRECTORY is specified in
ALTER TABLE statement.
mysql-test/r/func_sapdb.result:
test cases for date range edge cases added
mysql-test/r/func_time.result:
test cases for date range edge cases added
mysql-test/t/func_sapdb.test:
test cases for date range edge cases added
mysql-test/t/func_time.test:
test cases for date range edge cases added
Very complex select statements can create temporary tables
that are too big to be represented as a MyISAM table.
This was not checked at table creation time, but only at
open time. The result was an attempt to delete the
"impossible" table.
But if the server is built --with-raid, MyISAM tries to
open the table before deleting the files. It needs to find
out if the table uses the raid support and how many raid
chunks there are. This is done with an open "for repair",
which will almost always succeed.
But in this case we have an "impossible" table. The open
failed. Hence the files were not deleted. Also the error
message was a bit unspecific.
I turned an open error in this situation into the assumption
of having no raid support on the table. Thus the normal data
file is tried to be deleted. This may however leave existing
raid chunks behind.
I also added a check in mi_create() to prevent the creation
of an "impossible" table. A more decriptive error message is
given in this case.
No test case. The required select statement is way too
large for the test suite. I added a test script to the
bug report.
myisam/mi_create.c:
Bug#11824 - internal /tmp/*.{MYD,MYI} files remain, causing subsequent queries to fail
Added a check to mi_create() that the table description
header of the index file does not exceed 64KB. The header
has only 16 bits to encode its length.
myisam/mi_delete_table.c:
Bug#11824 - internal /tmp/*.{MYD,MYI} files remain, causing subsequent queries to fail
Interpret error in table open as not having a raid
configuration on the tbale. Thus try to delete the
normal data file, but leave behind raid chunks if
they exist.
- make sure to allocate just enough pages in the fragments by using the actual
row count from the backup, to avoid over allocation of pages to fragments, and
thus avoid the bug
ndb/include/kernel/GlobalSignalNumbers.h:
Bug #19852 Restoring backup made from cluster with full data memory fails
- distribute fragment complete to all participants to update row count
ndb/include/kernel/signaldata/BackupContinueB.hpp:
Bug #19852 Restoring backup made from cluster with full data memory fails
- time slica writing of fragment info to ctl file
ndb/include/kernel/signaldata/BackupImpl.hpp:
Bug #19852 Restoring backup made from cluster with full data memory fails
- 32 -> 64 bit on bytes and records
- new signal fragment complete to all participants
ndb/include/kernel/signaldata/BackupSignalData.hpp:
Bug #19852 Restoring backup made from cluster with full data memory fails
- 32 -> 64 bit on bytes and records
ndb/include/kernel/signaldata/DictTabInfo.hpp:
Bug #19852 Restoring backup made from cluster with full data memory fails
- add min and max rows to dict tab info
ndb/include/kernel/signaldata/LqhFrag.hpp:
Bug #19852 Restoring backup made from cluster with full data memory fails
- added min and max rows to add frag req
ndb/include/kernel/signaldata/TupFrag.hpp:
Bug #19852 Restoring backup made from cluster with full data memory fails
- added min and max rows to add frag req
ndb/include/ndbapi/NdbDictionary.hpp:
Bug #19852 Restoring backup made from cluster with full data memory fails
- added get/set of min max rows
ndb/src/common/debugger/signaldata/BackupImpl.cpp:
Bug #19852 Restoring backup made from cluster with full data memory fails
- 32 -> 64 bit on bytes and records
ndb/src/common/debugger/signaldata/BackupSignalData.cpp:
Bug #19852 Restoring backup made from cluster with full data memory fails
- 32 -> 64 bit on bytes and records
ndb/src/common/debugger/signaldata/DictTabInfo.cpp:
Bug #19852 Restoring backup made from cluster with full data memory fails
- added min and max rows to dict tab info
ndb/src/common/debugger/signaldata/LqhFrag.cpp:
Bug #19852 Restoring backup made from cluster with full data memory fails
- added min and max rows to frag req
ndb/src/kernel/blocks/backup/Backup.cpp:
Bug #19852 Restoring backup made from cluster with full data memory fails
- new section in backup with per fragment info in ctl file
- 32 -> 64 bit on bytes and records
ndb/src/kernel/blocks/backup/Backup.hpp:
Bug #19852 Restoring backup made from cluster with full data memory fails
- new section in backup with per fragment info in ctl file
- 32 -> 64 bit on bytes and records
ndb/src/kernel/blocks/backup/BackupFormat.hpp:
Bug #19852 Restoring backup made from cluster with full data memory fails
- new section in backup with per fragment info in ctl file
- 32 -> 64 bit on bytes and records
ndb/src/kernel/blocks/backup/BackupInit.cpp:
Bug #19852 Restoring backup made from cluster with full data memory fails
- new signal fragment complete to all participants
ndb/src/kernel/blocks/dbdict/Dbdict.cpp:
Bug #19852 Restoring backup made from cluster with full data memory fails
- added max and min rows to dict table object
ndb/src/kernel/blocks/dbdict/Dbdict.hpp:
Bug #19852 Restoring backup made from cluster with full data memory fails
- added max and min rows to dict table object
ndb/src/kernel/blocks/dblqh/Dblqh.hpp:
Bug #19852 Restoring backup made from cluster with full data memory fails
- added min and max rows to frag req
ndb/src/kernel/blocks/dblqh/DblqhMain.cpp:
Bug #19852 Restoring backup made from cluster with full data memory fails
- added min and max rows to frag req
ndb/src/kernel/blocks/dbtup/Dbtup.hpp:
Bug #19852 Restoring backup made from cluster with full data memory fails
- added min and max rows to frag req
ndb/src/kernel/blocks/dbtup/DbtupMeta.cpp:
Bug #19852 Restoring backup made from cluster with full data memory fails
- added min and max rows to frag req
- move memory allocation to fragment to after adding of attributes to get correct headsize
- allocate pages to fragments according to min rows setting
ndb/src/kernel/blocks/dbtup/DbtupPageMap.cpp:
Bug #19852 Restoring backup made from cluster with full data memory fails
- grow page allocation starting from 2 irrespective of first page allocation
ndb/src/mgmsrv/MgmtSrvr.cpp:
Bug #19852 Restoring backup made from cluster with full data memory fails
- 32 -> 64 bits on bytes and records
ndb/src/mgmsrv/MgmtSrvr.hpp:
Bug #19852 Restoring backup made from cluster with full data memory fails
- 32 -> 64 bits on bytes and records
ndb/src/ndbapi/NdbDictionary.cpp:
Bug #19852 Restoring backup made from cluster with full data memory fails
- min and max rows in dict
ndb/src/ndbapi/NdbDictionaryImpl.cpp:
Bug #19852 Restoring backup made from cluster with full data memory fails
- min and max rows in dict
ndb/src/ndbapi/NdbDictionaryImpl.hpp:
Bug #19852 Restoring backup made from cluster with full data memory fails
- min and max rows in dict
ndb/tools/restore/Restore.cpp:
Bug #19852 Restoring backup made from cluster with full data memory fails
- add retrieval of fragment info
ndb/tools/restore/Restore.hpp:
Bug #19852 Restoring backup made from cluster with full data memory fails
- add retrieval of fragment info
ndb/tools/restore/consumer_restore.cpp:
Bug #19852 Restoring backup made from cluster with full data memory fails
- set min in restore to the actual row count (this is the actual bug fix)
sql/ha_ndbcluster.cc:
Bug #19852 Restoring backup made from cluster with full data memory fails
- set min and max rows according to sql definition
