Bugs fixed:
The reported milliseconds since the last server packets were received/sent was incorrect by a factor of 1000. For example, the following method call:
SQLError.createLinkFailureMessageBasedOnHeuristics( (ConnectionImpl) this.conn, System.currentTimeMillis() - 1000, System.currentTimeMillis() - 2000, e, false);
returned the following string:
The last packet successfully received from the server was 2 milliseconds ago. The last packet sent successfully to the server was 1 milliseconds ago.
Calling Connection.serverPrepareStatement()
variants that do not take result set type or concurrency
arguments returned statements that produced result sets with
incorrect defaults, namely
TYPE_SCROLL_SENSITIVE
.
(Bug#45171)
The result set returned by getIndexInfo()
did
not have the format defined in the JDBC API specifications. The
fourth column, DATA_TYPE
, of the result set
should be of type BOOLEAN
. Connector/J
however returns CHAR
.
(Bug#44869)
The result set returned by getTypeInfo()
did
not have the format defined in the JDBC API specifications. The
second column, DATA_TYPE
, of the result set
should be of type INTEGER
. Connector/J
however returns SMALLINT
.
(Bug#44868)
The DEFERRABILITY
column in database metadata
result sets was expected to be of type SHORT
.
However, Connector/J returned it as INTEGER
.
This affected the following methods:
getImportedKeys()
,
getExportedKeys()
,
getCrossReference()
.
(Bug#44867)
The result set returned by getColumns()
did
not have the format defined in the JDBC API specifications. The
fifth column, DATA_TYPE
, of the result set
should be of type INTEGER
. Connector/J
however returns SMALLINT
.
(Bug#44865)
The result set returned by
getVersionColumns()
did not have the format
defined in the JDBC API specifications. The third column,
DATA_TYPE
, of the result set should be of
type INTEGER
. Connector/J however returns
SMALLINT
.
(Bug#44863)
The result set returned by
getBestRowIdentifier()
did not have the
format defined in the JDBC API specifications. The third column,
DATA_TYPE
, of the result set should be of
type INTEGER
. Connector/J however returns
SMALLINT
.
(Bug#44862)
Connector/J contains logic to generate a message text
specifically for streaming result sets when there are
CommunicationsException
exceptions generated.
However, this code was never reached.
In the CommunicationsException
code:
private boolean streamingResultSetInPlay = false; public CommunicationsException(ConnectionImpl conn, long lastPacketSentTimeMs, long lastPacketReceivedTimeMs, Exception underlyingException) { this.exceptionMessage = SQLError.createLinkFailureMessageBasedOnHeuristics(conn, lastPacketSentTimeMs, lastPacketReceivedTimeMs, underlyingException, this.streamingResultSetInPlay);
streamingResultSetInPlay
was always false,
which in the following code in
SQLError.createLinkFailureMessageBasedOnHeuristics()
never being executed:
if (streamingResultSetInPlay) { exceptionMessageBuf.append( Messages.getString("CommunicationsException.ClientWasStreaming")); //$NON-NLS-1$ } else { ...
The
SQLError.createLinkFailureMessageBasedOnHeuristics()
method created a message text for communication link failures.
When certain conditions were met, this message included both
“last packet sent” and “last packet
received” information, but when those conditions were not
met, only “last packet sent” information was
provided.
Information about when the last packet was successfully received should be provided in all cases. (Bug#44587)
Statement.getGeneratedKeys()
retained result
set instances until the statement was closed. This caused memory
leaks for long-lived statements, or statements used in tight
loops.
(Bug#44056)
Using useInformationSchema
with
DatabaseMetaData.getExportedKeys()
generated
the following exception:
com.mysql.jdbc.exceptions.MySQLIntegrityConstraintViolationException: Column 'REFERENCED_TABLE_NAME' in where clause is ambiguous ... at com.mysql.jdbc.PreparedStatement.executeInternal(PreparedStatement.java:1772) at com.mysql.jdbc.PreparedStatement.executeQuery(PreparedStatement.java:1923) at com.mysql.jdbc.DatabaseMetaDataUsingInfoSchema.executeMetadataQuery( DatabaseMetaDataUsingInfoSchema.java:50) at com.mysql.jdbc.DatabaseMetaDataUsingInfoSchema.getExportedKeys( DatabaseMetaDataUsingInfoSchema.java:603)
LoadBalancingConnectionProxy.doPing()
did not
have blacklist awareness.
LoadBalancingConnectionProxy
implemented
doPing()
to ping all underlying connections,
but it threw any exceptions it encountered during this process.
With the global blacklist enabled, it catches these exceptions, adds the host to the global blacklist, and only throws an exception if all hosts are down. (Bug#43421)
The method Statement.getGeneratedKeys()
did
not return values for UNSIGNED BIGINTS
with
values greater than Long.MAX_VALUE
.
Unfortunately, because the server does not tell clients what
TYPE the auto increment value is, the driver cannot consistently
return BigIntegers for the result set returned from
getGeneratedKeys()
, it will only return them
if the value is greater than Long.MAX_VALUE
.
If your application needs this consistency, it will need to
check the class of the return value from
.getObject()
on the ResultSet returned by
Statement.getGeneratedKeys()
and if it is not
a BigInteger, create one based on the
java.lang.Long
that is returned.
(Bug#43196)
When the MySQL Server was upgraded from 4.0 to 5.0, the Connector/J application then failed to connect to the server. This was because authentication failed when the application ran from EBCDIC platforms such as z/OS. (Bug#43071)
When connecting with traceProtocol=true
, no
trace data was generated for the server greeting or login
request.
(Bug#43070)
Connector/J generated an unhandled
StringIndexOutOfBoundsException
:
java.lang.StringIndexOutOfBoundsException: String index out of range: -1 at java.lang.String.substring(String.java:1938) at com.mysql.jdbc.EscapeProcessor.processTimeToken(EscapeProcessor.java:353) at com.mysql.jdbc.EscapeProcessor.escapeSQL(EscapeProcessor.java:257) at com.mysql.jdbc.StatementImpl.executeUpdate(StatementImpl.java:1546) at com.mysql.jdbc.StatementImpl.executeUpdate(StatementImpl.java:1524)
A ConcurrentModificationException
was
generated in LoadBalancingConnectionProxy
:
java.util.ConcurrentModificationException at java.util.HashMap$HashIterator.nextEntry(Unknown Source) at java.util.HashMap$KeyIterator.next(Unknown Source) at com.mysql.jdbc.LoadBalancingConnectionProxy.getGlobalBlacklist(LoadBalancingConnectionProxy.java:520) at com.mysql.jdbc.RandomBalanceStrategy.pickConnection(RandomBalanceStrategy.java:55) at com.mysql.jdbc.LoadBalancingConnectionProxy.pickNewConnection(LoadBalancingConnectionProxy.java:414) at com.mysql.jdbc.LoadBalancingConnectionProxy.invoke(LoadBalancingConnectionProxy.java:390)
SQL injection was possible when using a string containing U+00A5 in a client-side prepared statement, and the character set being used was SJIS/Windows-31J. (Bug#41730)
If there was an apostrophe in a comment in a statement that was
being sent through Connector/J, the apostrophe was still
recognized as a quote and put the state machine in
EscapeTokenizer
into the
inQuotes
state. This led to further parse
errors.
For example, consider the following statement:
String sql = "-- Customer's zip code will be fixed\n" + "update address set zip_code = 99999\n" + "where not regexp '^[0-9]{5}([[.-.]])?([0-9]{4})?$'";
When passed through Connector/J, the
EscapeTokenizer
did not recognize that the
first apostrophe was in a comment and thus set
inQuotes
to true. When that happened, the
quote count was incorrect and thus the regular expression did
not appear to be in quotes. With the parser not detecting that
the regular expression was in quotes, the curly braces were
recognized as escape sequences and were removed from the regular
expression, breaking it. The server thus received SQL such as:
-- Customer's zip code will be fixed update address set zip_code = '99999' where not regexp '^[0-9]([[.-.]])?([0-9])?$'
MySQL Connector/J 5.1.7 was slower than previous versions when
the rewriteBatchedStatements
option was set
to true
.
The performance regression in
indexOfIgnoreCaseRespectMarker()
has been
fixed. It has also been made possible for the driver to
rewrite INSERT
statements with ON
DUPLICATE KEY UPDATE
clauses in them, as long as the
UPDATE
clause contains no reference to
LAST_INSERT_ID()
, as that would cause the
driver to return bogus values for
getGeneratedKeys()
invocations. This has
resulted in improved performance over version 5.1.7.
When accessing a result set column by name using
ResultSetImpl.findColumn()
an exception was
generated:
java.lang.NullPointerException at com.mysql.jdbc.ResultSetImpl.findColumn(ResultSetImpl.java:1103) at com.mysql.jdbc.ResultSetImpl.getShort(ResultSetImpl.java:5415) at org.apache.commons.dbcp.DelegatingResultSet.getShort(DelegatingResultSet.java:219) at com.zimbra.cs.db.DbVolume.constructVolume(DbVolume.java:297) at com.zimbra.cs.db.DbVolume.get(DbVolume.java:197) at com.zimbra.cs.db.DbVolume.create(DbVolume.java:95) at com.zimbra.cs.store.Volume.create(Volume.java:227) at com.zimbra.cs.store.Volume.create(Volume.java:189) at com.zimbra.cs.service.admin.CreateVolume.handle(CreateVolume.java:48) at com.zimbra.soap.SoapEngine.dispatchRequest(SoapEngine.java:428) at com.zimbra.soap.SoapEngine.dispatch(SoapEngine.java:285)
The RETURN_GENERATED_KEYS
flag was being
ignored. For example, in the following code the
RETURN_GENERATED_KEYS
flag was ignored:
PreparedStatement ps = connection.prepareStatement("INSERT INTO table values(?,?)",PreparedStatement.RETURN_GENERATED_KEYS);
When using Connector/J 5.1.7 to connect to MySQL Server 4.1.18 the following error message was generated:
Thu Dec 11 17:38:21 PST 2008 WARN: Invalid value {1} for server variable named {0}, falling back to sane default of {2}
This occurred with MySQL Server version that did not support
auto_increment_increment
. The error message
should not have been generated.
(Bug#41416)
When DatabaseMetaData.getProcedureColumns()
was called, the value for LENGTH
was always
returned as 65535, regardless of the column type (fixed or
variable) or the actual length of the column.
However, if you obtained the PRECISION
value,
this was correct for both fixed and variable length columns.
(Bug#41269)
PreparedStatement.addBatch()
did not check
for all parameters being set, which led to inconsistent behavior
in executeBatch()
, especially when rewriting
batched statements into multi-value INSERT
s.
(Bug#41161)
Error message strings contained variable values that were not expanded. For example:
Mon Nov 17 11:43:18 JST 2008 WARN: Invalid value {1} for server variable named {0}, falling back to sane default of {2}
When using rewriteBatchedStatements=true
with:
INSERT INTO table_name_values (...) VALUES (...)
Query rewriting failed because “values” at the end of the table name was mistaken for the reserved keyword. The error generated was as follows:
testBug40439(testsuite.simple.TestBug40439)java.sql.BatchUpdateException: You have an error in your SQL syntax; check the manual that corresponds to your MySQL server version for the right syntax to use near 'values (2,'toto',2),(id,data, ordr) values (3,'toto',3),(id,data, ordr) values (' at line 1 at com.mysql.jdbc.PreparedStatement.executeBatchedInserts(PreparedStatement.java:1495) at com.mysql.jdbc.PreparedStatement.executeBatch(PreparedStatement.java:1097) at testsuite.simple.TestBug40439.testBug40439(TestBug40439.java:42) at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method) at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:39) at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25) at testsuite.simple.TestBug40439.main(TestBug40439.java:57)
A statement interceptor received the incorrect parameters when used with a batched statement. (Bug#39426)
Using Connector/J 5.1.6 the method
ResultSet.getObject
returned a
BYTE[]
for following:
SELECT TRIM(rowid) FROM tbl
Where rowid
had a type of INT(11)
PRIMARY KEY AUTO_INCREMENT
.
The expected return type was one of CHAR
,
VARCHAR
, CLOB
, however, a
BYTE[]
was returned.
Further, adding
functionsNeverReturnBlobs=true
to the
connection string did not have any effect on the return type.
(Bug#38387)
User Comments
http://forums.mysql.com/read.php?39,353773,353773
Add your own comment.