Source code editor What Is Ajax
↑
This is a new Beta development release, fixing recently discovered bugs.
Functionality added or changed:
Setting the configuration property rewriteBatchedStatements
to true
will now cause the driver to rewrite batched prepared statements with more than 3 parameter sets in a batch into multi-statements (separated by ";") if they are not plain (i.e. without
or SELECT
or ON DUPLICATE KEY UPDATE
clauses) INSERTREPLACE
statements.
This is a new Alpha development release, adding new features and fixing recently discovered bugs.
Functionality added or changed:
Pulled vendor-extension methods of Connection
implementation out into an interface to support java.sql.Wrapper
functionality from ConnectionPoolDataSource
. The vendor extensions are javadoc'd in the com.mysql.jdbc.Connection
interface.
For those looking further into the driver implementation, it is not an API that is used for plugability of implementations inside our driver (which is why there are still references to ConnectionImpl
throughout the code).
We've also added server and client prepareStatement()
methods that cover all of the variants in the JDBC API.
Incompatible change: Connection.serverPrepare(String)
has been re-named to Connection.serverPrepareStatement()
for consistency with Connection.clientPrepareStatement()
.
Similar to Connection
, we pulled out vendor extensions to Statement
into an interface named com.mysql.Statement
, and moved the Statement
class into com.mysql.StatementImpl
. The two methods (javadoc'd in com.mysql.Statement
are enableStreamingResults()
, which already existed, and disableStreamingResults()
which sets the statement instance back to the fetch size and result set type it had before enableStreamingResults()
was called.
Added experimental support for statement "interceptors" via the com.mysql.jdbc.StatementInterceptor
interface, examples are in com/mysql/jdbc/interceptors
.
Implement this interface to be placed "in between" query execution, so that you can influence it. (currently experimental).
StatementInterceptors
are "chainable" when configured by the user, the results returned by the "current" interceptor will be passed on to the next on in the chain, from left-to-right order, as specified by the user in the JDBC configuration property statementInterceptors
.
See the sources (fully javadoc'd) for com.mysql.jdbc.StatementInterceptor
for more details until we iron out the API and get it documented in the manual.
Externalized the descriptions of connection properties.
The data (and how it's stored) for ResultSet
rows are now behind an interface which allows us (in some cases) to allocate less memory per row, in that for "streaming" result sets, we re-use the packet used to read rows, since only one row at a time is ever active.
Made it possible to retrieve prepared statement parameter bindings (to be used in StatementInterceptors
, primarily).
Row navigation now causes any streams/readers open on the result set to be closed, as in some cases we're reading directly from a shared network packet and it will be overwritten by the "next" row.
Setting rewriteBatchedStatements
to true
now causes CallableStatements
with batched arguments to be re-written in the form CALL (...); CALL (...); ...
to send the batch in as few client-server round trips as possible.
Driver now picks appropriate internal row representation (whole row in one buffer, or individual byte[]s for each column value) depending on heuristics, including whether or not the row has BLOB
or TEXT
types and the overall row-size. The threshold for row size that will cause the driver to use a buffer rather than individual byte[]s is configured by the configuration property largeRowSizeThreshold
, which has a default value of 2KB.
This is the first public alpha release of the current Connector/J 5.1 development branch, providing an insight to upcoming features. Although some of these are still under development, this release includes the following new features and changes (in comparison to the current Connector/J 5.0 production release):
Important change: Due to a number of issues with the use of server-side prepared statements, Connector/J 5.0.5 has disabled their use by default. The disabling of server-side prepared statements does not affect the operation of the connector in any way.
To enable server-side prepared statements you must add the following configuration property to your connector string:
useServerPrepStmts=true
The default value of this property is false
(that is, Connector/J does not use server-side prepared statements).
The disabling of server-side prepared statements does not affect the operation of the connector. However, if you use the useTimezone=true
connection option and use client-side prepared statements (instead of server-side prepared statements) you should also set useSSPSCompatibleTimezoneShift=true
.
Functionality added or changed:
Re-worked Ant buildfile to build JDBC-4.0 classes separately, as well as support building under Eclipse (since Eclipse can't mix/match JDKs).
To build, you must set JAVA_HOME
to J2SDK-1.4.2 or Java-5, and set the following properties on your Ant command line:
com.mysql.jdbc.java6.javac
— full path to your Java-6 javac executable
com.mysql.jdbc.java6.rtjar
— full path to your Java-6 rt.jar
file
New feature — driver will automatically adjust session variable net_write_timeout
when it determines it has been asked for a "streaming" result, and resets it to the previous value when the result set has been consumed. (configuration property is named netTimeoutForStreamingResults
value and has a unit of seconds, the value 0
means the driver will not try and adjust this value).
Added support for JDBC-4.0 categorized SQLExceptions
.
Refactored CommunicationsException
into a JDBC-3.0 version, and a JDBC-4.0 version (which extends SQLRecoverableException
, now that it exists).
This change means that if you were catching com.mysql.jdbc.CommunicationsException
in your applications instead of looking at the SQLState class of 08
, and are moving to Java 6 (or newer), you need to change your imports to that exception to be com.mysql.jdbc.exceptions.jdbc4.CommunicationsException
, as the old class will not be instantiated for communications link-related errors under Java 6.
Added support for JDBC-4.0's client information. The backend storage of information provided via Connection.setClientInfo()
and retrieved by Connection.getClientInfo()
is pluggable by any class that implements the com.mysql.jdbc.JDBC4ClientInfoProvider
interface and has a no-args constructor.
The implementation used by the driver is configured using the clientInfoProvider
configuration property (with a default of value of com.mysql.jdbc.JDBC4CommentClientInfoProvider
, an implementation which lists the client information as a comment prepended to every query sent to the server).
This functionality is only available when using Java-6 or newer.
Added support for JDBC-4.0's SQLXML interfaces.
Added support for JDBC-4.0's Wrapper
interface.
Added support for JDBC-4.0's NCLOB
, and NCHAR
/NVARCHAR
types.
Functionality added or changed:
Fixed issue where a failed-over connection would let an application call setReadOnly(false)
, when that call should be ignored until the connection is reconnected to a writable master unless failoverReadOnly
had been set to false
.
Driver will now use INSERT INTO ... VALUES (DEFAULT)
form of statement for updatable result sets for ResultSet.insertRow()
, rather than pre-populating the insert row with values from DatabaseMetaData.getColumns()
(which results in a SHOW FULL COLUMNS
on the server for every result set). If an application requires access to the default values before insertRow()
has been called, the JDBC URL should be configured with populateInsertRowWithDefaultValues
set to true
.
This fix specifically targets performance issues with ColdFusion and the fact that it seems to ask for updatable result sets no matter what the application does with them.
com.mysql.jdbc.[NonRegistering]Driver
now understands URLs of the format jdbc:mysql:replication://
and jdbc:mysql:loadbalance://
which will create a ReplicationConnection (exactly like when using [NonRegistering]ReplicationDriver
) and an experimental load-balanced connection designed for use with SQL nodes in a MySQL Cluster/NDB environment, respectively.
In an effort to simplify things, we're working on deprecating multiple drivers, and instead specifying different core behavior based upon JDBC URL prefixes, so watch for [NonRegistering]ReplicationDriver
to eventually disappear, to be replaced with com.mysql.jdbc[NonRegistering]Driver
with the new URL prefix.
Added an experimental load-balanced connection designed for use with SQL nodes in a MySQL Cluster/NDB environment (This is not for master-slave replication. For that, we suggest you look at ReplicationConnection
or lbpool
).
If the JDBC URL starts with jdbc:mysql:loadbalance://host-1,host-2,...host-n
, the driver will create an implementation of java.sql.Connection
that load balances requests across a series of MySQL JDBC connections to the given hosts, where the balancing takes place after transaction commit.
Therefore, for this to work (at all), you must use transactions, even if only reading data.
Physical connections to the given hosts will not be created until needed.
The driver will invalidate connections that it detects have had communication errors when processing a request. A new connection to the problematic host will be attempted the next time it is selected by the load balancing algorithm.
There are two choices for load balancing algorithms, which may be specified by the loadBalanceStrategy
JDBC URL configuration property:
random
— the driver will pick a random host for each request. This tends to work better than round-robin, as the randomness will somewhat account for spreading loads where requests vary in response time, while round-robin can sometimes lead to overloaded nodes if there are variations in response times across the workload.
bestResponseTime
— the driver will route the request to the host that had the best response time for the previous transaction.
When useLocalSessionState
is set to true
and connected to a MySQL-5.0 or later server, the JDBC driver will now determine whether an actual commit
or rollback
statement needs to be sent to the database when Connection.commit()
or Connection.rollback()
is called.
This is especially helpful for high-load situations with connection pools that always call Connection.rollback()
on connection check-in/check-out because it avoids a round-trip to the server.
Give better error message when "streaming" result sets, and the connection gets clobbered because of exceeding net_write_timeout
on the server.
New configuration property, enableQueryTimeouts
(default true
).
When enabled, query timeouts set via Statement.setQueryTimeout()
use a shared java.util.Timer
instance for scheduling. Even if the timeout doesn't expire before the query is processed, there will be memory used by the TimerTask
for the given timeout which won't be reclaimed until the time the timeout would have expired if it hadn't been cancelled by the driver. High-load environments might want to consider disabling this functionality. (this configuration property is part of the maxPerformance
configuration bundle).
Added configuration property padCharsWithSpace
(defaults to false
). If set to true
, and a result set column has the CHAR
type and the value does not fill the amount of characters specified in the DDL for the column, the driver will pad the remaining characters with space (for ANSI compliance).
Added configuration property useDynamicCharsetInfo
. If set to false
(the default), the driver will use a per-connection cache of character set information queried from the server when necessary, or when set to true
, use a built-in static mapping that is more efficient, but isn't aware of custom character sets or character sets implemented after the release of the JDBC driver.
This only affects the padCharsWithSpace
configuration property and the ResultSetMetaData.getColumnDisplayWidth()
method.
More intelligent initial packet sizes for the "shared" packets are used (512 bytes, rather than 16K), and initial packets used during handshake are now sized appropriately as to not require reallocation.
Bugs fixed:
More useful error messages are generated when the driver thinks a result set is not updatable. (Thanks to Ashley Martens for the patch). (Bug#28085)
Using ResultSet.get*()
with a column index less than 1 returns a misleading error message. (Bug#27317)
Client options not sent correctly when using SSL, leading to stored procedures not being able to return results. Thanks to Don Cohen for the bug report, testcase and patch. (Bug#25545)
PreparedStatement
is not closed in BlobFromLocator.getBytes()
. (Bug#26592)
Whitespace surrounding storage/size specifiers in stored procedure parameters declaration causes NumberFormatException
to be thrown when calling stored procedure on JDK-1.5 or newer, as the Number classes in JDK-1.5+ are whitespace intolerant. (Bug#25624)
When the configuration property useCursorFetch
was set to true
, sometimes server would return new, more exact metadata during the execution of the server-side prepared statement that enables this functionality, which the driver ignored (using the original metadata returned during prepare()
), causing corrupt reading of data due to type mismatch when the actual rows were returned. (Bug#26173)
Comments in DDL of stored procedures/functions confuse procedure parser, and thus metadata about them can not be created, leading to inability to retrieve said metadata, or execute procedures that have certain comments in them. (Bug#26959)
Fast date/time parsing doesn't take into account 00:00:00
as a legal value. (Bug#26789)
ResultSet.get*()
with a column index < 1 returns misleading error message. (Bug#27317)
Statement.setMaxRows()
is not effective on result sets materialized from cursors. (Bug#25517)
CALL /* ... */
doesn't work. As a side effect of this fix, you can now use some_proc
()/* */
and #
comments when preparing statements using client-side prepared statement emulation. (Bug#27400)
If the comments happen to contain parameter markers (?
), they will be treated as belonging to the comment (that is, not recognized) rather than being a parameter of the statement.
The statement when sent to the server will contain the comments as-is, they're not stripped during the process of preparing the PreparedStatement
or CallableStatement
.
BIT(> 1)
is returned as java.lang.String
from ResultSet.getObject()
rather than byte[]
. (Bug#25328)
CallableStatements
with OUT/INOUT
parameters that are "binary" (BLOB
, BIT
, (VAR)BINARY
, JAVA_OBJECT
) have extra 7 bytes. (Bug#25715)
Connection.getTransactionIsolation()
uses "SHOW VARIABLES LIKE
" which is very inefficient on MySQL-5.0+ servers. (Bug#27655)
Fixed issue where calling getGeneratedKeys()
on a prepared statement after calling execute()
didn't always return the generated keys (executeUpdate()
worked fine however).
Functionality added or changed:
Important change: Due to a number of issues with the use of server-side prepared statements, Connector/J 5.0.5 has disabled their use by default. The disabling of server-side prepared statements does not affect the operation of the connector in any way.
To enable server-side prepared statements you must add the following configuration property to your connector string:
useServerPrepStmts=true
The default value of this property is false
(that is, Connector/J does not use server-side prepared statements).
Improved speed of datetime
parsing for ResultSets that come from plain or non-server-side prepared statements. You can enable old implementation with useFastDateParsing=false
as a configuration parameter.
Added configuration property localSocketAddress
,which is the hostname or IP address given to explicitly configure the interface that the driver will bind the client side of the TCP/IP connection to when connecting.
The rewriteBatchedStatements
feature can now be used with server-side prepared statements.
We've added a new configuration option treatUtilDateAsTimestamp
, which is false
by default, as (1) We already had specific behavior to treat java.util.Date as a java.sql.Timestamp because it's useful to many folks, and (2) that behavior will very likely be required for drivers JDBC-post-4.0.
Fixed logging of XA commands sent to server, it's now configurable via logXaCommands
property (defaults to false
).
Usage Advisor now detects empty results sets and does not report on columns not referenced in those empty sets.
Usage Advisor will now issue warnings for result sets with large numbers of rows. You can configure the trigger value by using the resultSetSizeThreshold
parameter, which has a default value of 100.
Bugs fixed:
Fixed an issue where XADataSources
couldn't be bound into JNDI, as the DataSourceFactory
didn't know how to create instances of them.
Calling Statement.cancel()
could result in a Null Pointer Exception (NPE). (Bug#24721)
Calendars and timezones are now lazily instantiated when required. (Bug#24351)
Client-side prepared statement parser gets confused by in-line comments /*...*/
and therefore cannot rewrite batch statements or reliably detect the type of statements when they are used. (Bug#25025)
When using a JDBC connection URL that is malformed, the NonRegisteringDriver.getPropertyInfo
method will throw a Null Pointer Exception (NPE). (Bug#22628)
Using DatabaseMetaData.getSQLKeywords()
does not return a all of the of the reserved keywords for the current MySQL version. Current implementation returns the list of reserved words for MySQL 5.1, and does not distinguish between versions. (Bug#24794)
Specifying US-ASCII
as the character set in a connection to a MySQL 4.1 or newer server does not map correctly. (Bug#24840)
Storing a java.util.Date
object in a BLOB
column would not be serialized correctly during setObject
. (Bug#25787)
A query execution which timed out did not always throw a MySQLTimeoutException
. (Bug#25836)
A connection error would occur when connecting to a MySQL server with certain character sets. Some collations/character sets reported as "unknown" (specifically cias
variants of existing character sets), and inability to override the detected server character set. (Bug#23645)
Using setFetchSize()
breaks prepared SHOW
and other commands. (Bug#24360)
Using DATETIME
columns would result in time shifts when useServerPrepStmts
was true. The reason was due to different behavior when using client-side compared to server-side prepared statements and the useJDBCCompliantTimezoneShift
option. This is now fixed if moving from server-side prepared statements to client-side prepared statements by setting useSSPSCompatibleTimezoneShift
to true
, as the driver can't tell if this is a new deployment that never used server-side prepared statements, or if it is an existing deployment that is switching to client-side prepared statements from server-side prepared statements. (Bug#24344)
Inconsistency between getSchemas
and INFORMATION_SCHEMA
. (Bug#23304)
When using the rewriteBatchedStatements
connection option with PreparedState.executeBatch()
an internal memory leak would occur. (Bug#25073)
Fixed issue where field-level for metadata from DatabaseMetaData
when using INFORMATION_SCHEMA
didn't have references to current connections, sometimes leading to Null Pointer Exceptions (NPEs) when introspecting them via ResultSetMetaData
.
Connector/J now returns a better error message when server doesn't return enough information to determine stored procedure/function parameter types. (Bug#24065)
When using server-side prepared statements and timestamp columns, value would be incorrectly populated (with nanoseconds, not microseconds). (Bug#21438)
Timer instance used for Statement.setQueryTimeout()
created per-connection, rather than per-VM, causing memory leak. (Bug#25514)
Results sets from UPDATE
statements that are part of multi-statement queries would cause an SQLException
error, "Result is from UPDATE". (Bug#25009)
StringUtils.indexOfIgnoreCaseRespectQuotes()
isn't case-insensitive on the first character of the target. This bug also affected rewriteBatchedStatements
functionality when prepared statements did not use uppercase for the VALUES
clause. (Bug#25047)
Some exceptions thrown out of StandardSocketFactory
were needlessly wrapped, obscuring their true cause, especially when using socket timeouts. (Bug#21480)
DatabaseMetaData.getSchemas()
doesn't return a TABLE_CATALOG
column. (Bug#23303)
EscapeProcessor
gets confused by multiple backslashes. We now push the responsibility of syntax errors back on to the server for most escape sequences. (Bug#25399)
INOUT
parameters in CallableStatements
get doubly-escaped. (Bug#25379)
Connection property socketFactory
wasn't exposed via correctly named mutator/accessor, causing data source implementations that use JavaBean naming conventions to set properties to fail to set the property (and in the case of SJAS, fail silently when trying to set this parameter). (Bug#26326)
ParameterMetaData
throws NullPointerException
when prepared SQL has a syntax error. Added generateSimpleParameterMetadata
configuration property, which when set to true
will generate metadata reflecting VARCHAR
for every parameter (the default is false
, which will cause an exception to be thrown if no parameter metadata for the statement is actually available). (Bug#21267)
Other changes:
Performance enhancement of initial character set configuration, driver will only send commands required to configure connection character set session variables if the current values on the server do not match what is required.
Re-worked stored procedure parameter parser to be more robust. Driver no longer requires BEGIN
in stored procedure definition, but does have requirement that if a stored function begins with a label directly after the "returns" clause, that the label is not a quoted identifier.
Reverted back to internal character conversion routines for single-byte character sets, as the ones internal to the JVM are using much more CPU time than our internal implementation.
Changed cached result set metadata (when using cacheResultSetMetadata=true
) to be cached per-connection rather than per-statement as previously implemented.
Use a java.util.TreeMap
to map column names to ordinal indexes for ResultSet.findColumn()
instead of a HashMap. This allows us to have case-insensitive lookups (required by the JDBC specification) without resorting to the many transient object instances needed to support this requirement with a normal HashMap
with either case-adjusted keys, or case-insensitive keys. (In the worst case scenario for lookups of a 1000 column result set, TreeMaps are about half as fast wall-clock time as a HashMap, however in normal applications their use gives many orders of magnitude reduction in transient object instance creation which pays off later for CPU usage in garbage collection).
Avoid static synchronized code in JVM class libraries for dealing with default timezones.
Fixed cases where ServerPreparedStatements
weren't using cached metadata when cacheResultSetMetadata=true
was used.
When using cached metadata, skip field-level metadata packets coming from the server, rather than reading them and discarding them without creating com.mysql.jdbc.Field
instances.
Throw exceptions encountered during timeout to thread calling Statement.execute*()
, rather than RuntimeException
.
Take localSocketAddress
property into account when creating instances of CommunicationsException
when the underyling exception is a java.net.BindException
, so that a friendlier error message is given with a little internal diagnostics.
Fixed some Null Pointer Exceptions (NPEs) when cached metadata was used with UpdatableResultSets
.
When extracting foreign key information from SHOW CREATE TABLE
in DatabaseMetaData
, ignore exceptions relating to tables being missing (which could happen for cross-reference or imported-key requests, as the list of tables is generated first, then iterated).
Bugs fixed:
Column names don't match metadata in cases where server doesn't return original column names (column functions) thus breaking compatibility with applications that expect 1-1 mappings between findColumn() and rsmd.getColumnName(), usually manifests itself as "Can't find column ('')" exceptions. (Bug#21379)
When using information_schema for metadata, COLUMN_SIZE for getColumns() is not clamped to range of java.lang.Integer as is the case when not using information_schema, thus leading to a truncation exception that isn't present when not using information_schema. (Bug#21544)
Newlines causing whitespace to span confuse procedure parser when getting parameter metadata for stored procedures. (Bug#22024)
Driver was using milliseconds for Statement.setQueryTimeout() when specification says argument is to be in seconds. (Bug#22359)
Workaround for server crash when calling stored procedures via a server-side prepared statement (driver now detects prepare(stored procedure) and substitutes client-side prepared statement). (Bug#22297)
Added new _ci collations to CharsetMapping - utf8_unicode_ci not working. (Bug#22456)
Driver issues truncation on write exception when it shouldn't (due to sending big decimal incorrectly to server with server-side prepared statement). (Bug#22290)
DBMD.getColumns() does not return expected COLUMN_SIZE for the SET type, now returns length of largest possible set disregarding whitespace or the "," delimitters to be consistent with the ODBC driver. (Bug#22613)
Other changes:
Fixed configuration property jdbcCompliantTruncation
was not being used for reads of result set values.
Driver now supports {call sp}
(without "()" if procedure has no arguments).
Driver now sends numeric 1 or 0 for client-prepared statement setBoolean()
calls instead of '1' or '0'.
DatabaseMetaData correctly reports true
for supportsCatalog*()
methods.
Fixed Statement.cancel()
causes NullPointerException
if underlying connection has been closed due to server failure. (Bug#20650)
Added configuration option noAccessToProcedureBodies
which will cause the driver to create basic parameter metadata for CallableStatements
when the user does not have access to procedure bodies via SHOW CREATE PROCEDURE
or selecting from mysql.proc
instead of throwing an exception. The default value for this option is false
Bugs fixed:
If the connection to the server has been closed due to a server failure, then the cleanup process will call Statement.cancel()
, triggering a NullPointerException
, even though there is no active connection. (Bug#20650)
Fixed can't use XAConnection
for local transactions when no global transaction is in progress. (Bug#17401)
Fixed driver fails on non-ASCII platforms. The driver was assuming that the platform character set would be a superset of MySQL's latin1
when doing the handshake for authentication, and when reading error messages. We now use Cp1252 for all strings sent to the server during the handshake phase, and a hard-coded mapping of the language
systtem variable to the character set that is used for error messages. (Bug#18086)
Fixed ConnectionProperties
(and thus some subclasses) are not serializable, even though some J2EE containers expect them to be. (Bug#19169)
Fixed MysqlValidConnectionChecker
for JBoss doesn't work with MySQLXADataSources
. (Bug#20242)
Better caching of character set converters (per-connection) to remove a bottleneck for multibyte character sets.
Added connection/datasource property pinGlobalTxToPhysicalConnection
(defaults to false
). When set to true
, when using XAConnections
, the driver ensures that operations on a given XID are always routed to the same physical connection. This allows the XAConnection
to support XA START ... JOIN
after XA END
has been called, and is also a workaround for transaction managers that don't maintain thread affinity for a global transaction (most either always maintain thread affinity, or have it as a configuration option).
MysqlXaConnection.recover(int flags)
now allows combinations of XAResource.TMSTARTRSCAN
and TMENDRSCAN
. To simulate the “scanning” nature of the interface, we return all prepared XIDs for TMSTARTRSCAN
, and no new XIDs for calls with TMNOFLAGS
, or TMENDRSCAN
when not in combination with TMSTARTRSCAN
. This change was made for API compliance, as well as integration with IBM WebSphere's transaction manager.
Not released due to a packaging error
XADataSource
implemented (ported from 3.2 branch which won't be released as a product). Use com.mysql.jdbc.jdbc2.optional.MysqlXADataSource
as your datasource class name in your application server to utilize XA transactions in MySQL-5.0.10 and newer.
PreparedStatement.setString()
didn't work correctly when sql_mode
on server contained NO_BACKSLASH_ESCAPES
and no characters that needed escaping were present in the string.
Attempt detection of the MySQL type BINARY
(it's an alias, so this isn't always reliable), and use the java.sql.Types.BINARY
type mapping for it.
Moved -bin-g.jar
file into separate debug
subdirectory to avoid confusion.
Don't allow .setAutoCommit(true)
, or .commit()
or .rollback()
on an XA-managed connection as per the JDBC specification.
If the connection useTimezone
is set to true
, then also respect time zone conversions in escape-processed string literals (for example, "{ts ...}"
and "{t ...}"
).
Return original column name for RSMD.getColumnName()
if the column was aliased, alias name for .getColumnLabel()
(if aliased), and original table name for .getTableName()
. Note this only works for MySQL-4.1 and newer, as older servers don't make this information available to clients.
Setting useJDBCCompliantTimezoneShift=true
(it's not the default) causes the driver to use GMT for all TIMESTAMP
/DATETIME
time zones, and the current VM time zone for any other type that refers to time zones. This feature can not be used when useTimezone=true
to convert between server and client time zones.
Add one level of indirection of internal representation of CallableStatement
parameter metadata to avoid class not found issues on JDK-1.3 for ParameterMetadata
interface (which doesn't exist prior to JDBC-3.0).
Added unit tests for XADatasource
, as well as friendlier exceptions for XA failures compared to the "stock" XAException
(which has no messages).
Idle timeouts cause XAConnections
to whine about rolling themselves back. (Bug#14729)
Added support for Connector/MXJ integration via url subprotocol jdbc:mysql:mxj://...
.
Moved all SQLException
constructor usage to a factory in SQLError
(ground-work for JDBC-4.0 SQLState
-based exception classes).
Removed Java5-specific calls to BigDecimal
constructor (when result set value is ''
, (int)0
was being used as an argument indirectly via method return value. This signature doesn't exist prior to Java5.)
Added service-provider entry to META-INF/services/java.sql.Driver
for JDBC-4.0 support.
Return "[VAR]BINARY" for RSMD.getColumnTypeName()
when that is actually the type, and it can be distinguished (MySQL-4.1 and newer).
When fix for Bug#14562 was merged from 3.1.12, added functionality for CallableStatement
's parameter metadata to return correct information for .getParameterClassName()
.
Fuller synchronization of Connection
to avoid deadlocks when using multithreaded frameworks that multithread a single connection (usually not recommended, but the JDBC spec allows it anyways), part of fix to Bug#14972).
Implementation of Statement.cancel()
and Statement.setQueryTimeout()
. Both require MySQL-5.0.0 or newer server, require a separate connection to issue the KILL QUERY
statement, and in the case of setQueryTimeout()
creates an additional thread to handle the timeout functionality.
Note: Failures to cancel the statement for setQueryTimeout()
may manifest themselves as RuntimeExceptions
rather than failing silently, as there is currently no way to unblock the thread that is executing the query being cancelled due to timeout expiration and have it throw the exception instead.
Important change: Due to a number of issues with the use of server-side prepared statements, Connector/J 5.0.5 has disabled their use by default. The disabling of server-side prepared statements does not affect the operation of the connector in any way.
To enable server-side prepared statements you must add the following configuration property to your connector string:
useServerPrepStmts=true
The default value of this property is false
(that is, Connector/J does not use server-side prepared statements).
Bugs fixed:
Specifying US-ASCII
as the character set in a connection to a MySQL 4.1 or newer server does not map correctly. (Bug#24840)
Fixed updatable result set throws ClassCastException when there is row data and moveToInsertRow() is called. (Bug#20479)
Fixed Updatable result set that contains a BIT column fails when server-side prepared statements are used. (Bug#20485)
Fixed memory leak with profileSQL=true. (Bug#16987)
Connection fails to localhost when using timeout and IPv6 is configured. (Bug#19726)
Fixed NullPointerException in MysqlDataSourceFactory due to Reference containing RefAddrs with null content. (Bug#16791)
Fixed ResultSet.getShort() for UNSIGNED TINYINT returns incorrect values when using server-side prepared statements. (Bug#20306)
Fixed can't pool server-side prepared statements, exception raised when re-using them. (Bug#20687)
ResultSet.getSomeInteger() doesn't work for BIT(>1). (Bug#21062)
ResultSet.getFloatFromString() can't retrieve values near Float.MIN/MAX_VALUE. (Bug#18880)
Escape of quotes in client-side prepared statements parsing not respected. Patch covers more than bug report, including NO_BACKSLASH_ESCAPES being set, and stacked quote characters forms of escaping (that is, '' or ""). (Bug#20888)
ReplicationDriver does not always round-robin load balance depending on URL used for slaves list. (Bug#19993)
Fixed calling toString() on ResultSetMetaData for driver-generated (that is, from DatabaseMetaData method calls, or from getGeneratedKeys()) result sets would raise a NullPointerException.
DDriver throws NPE when tracing prepared statements that have been closed (in asSQL()). (Bug#21207)
Removed logger autodetection altogether, must now specify logger explicitly if you want to use a logger other than one that logs to STDERR.
Driver issues truncation on write exception when it shouldn't (due to sending big decimal incorrectly to server with server-side prepared statement). (Bug#22290)
Driver now sends numeric 1 or 0 for client-prepared statement setBoolean() calls instead of '1' or '0'.
Fixed bug where driver would not advance to next host if roundRobinLoadBalance=true and the last host in the list is down.
Fixed Bug#18258 - DatabaseMetaData.getTables(), columns() with bad catalog parameter threw exception rather than return empty result set (as required by spec).
Check and store value for continueBatchOnError property in constructor of Statements, rather than when executing batches, so that Connections closed out from underneath statements don't cause NullPointerExceptions when it's required to check this property.
Fixed bug when calling stored functions, where parameters weren't numbered correctly (first parameter is now the return value, subsequent parameters if specified start at index "2").
INOUT
parameter does not store IN
value. (Bug#15464)
Exception thrown for new decimal type when using updatable result sets. (Bug#14609)
No "dos" character set in MySQL > 4.1.0. (Bug#15544)
PreparedStatement.setObject()
serializes BigInteger
as object, rather than sending as numeric value (and is thus not complementary to .getObject()
on an UNSIGNED LONG
type). (Bug#15383)
ResultSet.getShort()
for UNSIGNED TINYINT
returned wrong values. (Bug#11874)
lib-nodist
directory missing from package breaks out-of-box build. (Bug#15676)
DBMD.getColumns()
returns wrong type for BIT
. (Bug#15854)
Fixed issue where driver was unable to initialize character set mapping tables. Removed reliance on .properties
files to hold this information, as it turns out to be too problematic to code around class loader hierarchies that change depending on how an application is deployed. Moved information back into the CharsetMapping
class. (Bug#14938)
Fixed updatable result set doesn't return AUTO_INCREMENT
values for insertRow()
when multiple column primary keys are used. (the driver was checking for the existence of single-column primary keys and an autoincrement value > 0 instead of a straightforward isAutoIncrement()
check). (Bug#16841)
Fixed Statement.getGeneratedKeys()
throws NullPointerException
when no query has been processed. (Bug#17099)
Fixed driver trying to call methods that don't exist on older and newer versions of Log4j. The fix is not trying to auto-detect presence of log4j, too many different incompatible versions out there in the wild to do this reliably. (Bug#13469)
If you relied on autodetection before, you will need to add "logger=com.mysql.jdbc.log.Log4JLogger" to your JDBC URL to enable Log4J usage, or alternatively use the new "CommonsLogger" class to take care of this.
Added support for Apache Commons logging, use "com.mysql.jdbc.log.CommonsLogger" as the value for the "logger" configuration property.
LogFactory now prepends "com.mysql.jdbc.log" to log class name if it can't be found as-specified. This allows you to use "short names" for the built-in log factories, for example "logger=CommonsLogger" instead of "logger=com.mysql.jdbc.log.CommonsLogger".
Fixed issue with ReplicationConnection
incorrectly copying state, doesn't transfer connection context correctly when transitioning between the same read-only states. (Bug#15570)
Fixed issue where server-side prepared statements don't cause truncation exceptions to be thrown when truncation happens. (Bug#18041)
Added performance feature, re-writing of batched executes for Statement.executeBatch()
(for all DML statements) and PreparedStatement.executeBatch()
(for INSERTs with VALUE clauses only). Enable by using "rewriteBatchedStatements=true" in your JDBC URL.
Fixed CallableStatement.registerOutParameter()
not working when some parameters pre-populated. Still waiting for feedback from JDBC experts group to determine what correct parameter count from getMetaData()
should be, however. (Bug#17898)
Fixed calling clearParameters()
on a closed prepared statement causes NPE. (Bug#17587)
Map "latin1" on MySQL server to CP1252 for MySQL > 4.1.0.
Added additional accessor and mutator methods on ConnectionProperties so that DataSource users can use same naming as regular URL properties.
Fixed data truncation and getWarnings()
only returns last warning in set. (Bug#18740)
Improved performance of retrieving BigDecimal
, Time
, Timestamp
and Date
values from server-side prepared statements by creating fewer short-lived instances of Strings
when the native type is not an exact match for the requested type. (Bug#18496)
Fixed aliased column names where length of name > 251 are corrupted. (Bug#18554)
Fixed ResultSet.wasNull()
not always reset correctly for booleans when done via conversion for server-side prepared statements. (Bug#17450)
Fixed invalid classname returned for ResultSetMetaData.getColumnClassName()
for BIGINT type
. (Bug#19282)
Fixed case where driver wasn't reading server status correctly when fetching server-side prepared statement rows, which in some cases could cause warning counts to be off, or multiple result sets to not be read off the wire.
Driver now aware of fix for BIT
type metadata that went into MySQL-5.0.21 for server not reporting length consistently (Bug#13601).
Fixed PreparedStatement.setObject(int, Object, int)
doesn't respect scale of BigDecimals. (Bug#19615)
Fixed ResultSet.wasNull()
returns incorrect value when extracting native string from server-side prepared statement generated result set. (Bug#19282)
Fixed client-side prepared statement bug with embedded ?
characters inside quoted identifiers (it was recognized as a placeholder, when it was not).
Don't allow executeBatch()
for CallableStatements
with registered OUT
/INOUT
parameters (JDBC compliance).
Fall back to platform-encoding for URLDecoder.decode()
when parsing driver URL properties if the platform doesn't have a two-argument version of this method.
Java type conversion may be incorrect for MEDIUMINT
. (Bug#14562)
Added configuration property useGmtMillisForDatetimes
which when set to true
causes ResultSet.getDate()
, .getTimestamp()
to return correct millis-since GMT when .getTime()
is called on the return value (currently default is false
for legacy behavior).
Fixed DatabaseMetaData.stores*Identifiers()
:
If lower_case_table_names=0
(on server):
storesLowerCaseIdentifiers()
returns false
storesLowerCaseQuotedIdentifiers()
returns false
storesMixedCaseIdentifiers()
returns true
storesMixedCaseQuotedIdentifiers()
returns true
storesUpperCaseIdentifiers()
returns false
storesUpperCaseQuotedIdentifiers()
returns true
If lower_case_table_names=1
(on server):
storesLowerCaseIdentifiers()
returns true
storesLowerCaseQuotedIdentifiers()
returns true
storesMixedCaseIdentifiers()
returns false
storesMixedCaseQuotedIdentifiers()
returns false
storesUpperCaseIdentifiers()
returns false
storesUpperCaseQuotedIdentifiers()
returns true
DatabaseMetaData.getColumns()
doesn't return TABLE_NAME
correctly. (Bug#14815)
Escape processor replaces quote character in quoted string with string delimiter. (Bug#14909)
OpenOffice expects DBMD.supportsIntegrityEnhancementFacility()
to return true
if foreign keys are supported by the datasource, even though this method also covers support for check constraints, which MySQL doesn't have. Setting the configuration property overrideSupportsIntegrityEnhancementFacility
to true
causes the driver to return true
for this method. (Bug#12975)
Added com.mysql.jdbc.testsuite.url.default
system property to set default JDBC url for testsuite (to speed up bug resolution when I'm working in Eclipse).
Unable to initialize character set mapping tables (due to J2EE classloader differences). (Bug#14938)
Deadlock while closing server-side prepared statements from multiple threads sharing one connection. (Bug#14972)
logSlowQueries
should give better info. (Bug#12230)
Extraneous sleep on autoReconnect
. (Bug#13775)
Driver incorrectly closes streams passed as arguments to PreparedStatements
. Reverts to legacy behavior by setting the JDBC configuration property autoClosePStmtStreams
to true
(also included in the 3-0-Compat configuration “bundle”). (Bug#15024)
maxQuerySizeToLog
is not respected. Added logging of bound values for execute()
phase of server-side prepared statements when profileSQL=true
as well. (Bug#13048)
Usage advisor complains about unreferenced columns, even though they've been referenced. (Bug#15065)
Don't increase timeout for failover/reconnect. (Bug#6577)
Process escape tokens in Connection.prepareStatement(...)
. (Bug#15141) You can disable this behavior by setting the JDBC URL configuration property processEscapeCodesForPrepStmts
to false
.
Reconnect during middle of executeBatch()
should not occur if autoReconnect
is enabled. (Bug#13255)
Spurious !
on console when character encoding is utf8
. (Bug#11629)
Fixed statements generated for testcases missing ;
for “plain” statements.
Incorrect generation of testcase scripts for server-side prepared statements. (Bug#11663)
Fixed regression caused by fix for Bug#11552 that caused driver to return incorrect values for unsigned integers when those integers where within the range of the positive signed type.
Moved source code to Subversion repository.
Escape tokenizer doesn't respect stacked single quotes for escapes. (Bug#11797)
GEOMETRY
type not recognized when using server-side prepared statements.
ReplicationConnection
won't switch to slave, throws “Catalog can't be null” exception. (Bug#11879)
Properties shared between master and slave with replication connection. (Bug#12218)
Statement.getWarnings()
fails with NPE if statement has been closed. (Bug#10630)
Only get char[]
from SQL in PreparedStatement.ParseInfo()
when needed.
Geometry types not handled with server-side prepared statements. (Bug#12104)
StringUtils.getBytes()
doesn't work when using multi-byte character encodings and a length in characters is specified. (Bug#11614)
Pstmt.setObject(...., Types.BOOLEAN)
throws exception. (Bug#11798)
maxPerformance.properties
mis-spells “elideSetAutoCommits”. (Bug#11976)
DBMD.storesLower/Mixed/UpperIdentifiers()
reports incorrect values for servers deployed on Windows. (Bug#11575)
ResultSet.moveToCurrentRow()
fails to work when preceded by a call to ResultSet.moveToInsertRow()
. (Bug#11190)
VARBINARY
data corrupted when using server-side prepared statements and .setBytes()
. (Bug#11115)
explainSlowQueries
hangs with server-side prepared statements. (Bug#12229)
Escape processor didn't honor strings demarcated with double quotes. (Bug#11498)
Lifted restriction of changing streaming parameters with server-side prepared statements. As long as all
streaming parameters were set before execution, .clearParameters()
does not have to be called. (due to limitation of client/server protocol, prepared statements can not reset individual stream data on the server side).
Reworked Field
class, *Buffer
, and MysqlIO
to be aware of field lengths > Integer.MAX_VALUE
.
Updated DBMD.supportsCorrelatedQueries()
to return true
for versions > 4.1, supportsGroupByUnrelated()
to return true
and getResultSetHoldability()
to return HOLD_CURSORS_OVER_COMMIT
.
Handling of catalog argument in DatabaseMetaData.getIndexInfo()
, which also means changes to the following methods in DatabaseMetaData
: (Bug#12541)
getBestRowIdentifier()
getColumns()
getCrossReference()
getExportedKeys()
getImportedKeys()
getIndexInfo()
getPrimaryKeys()
getProcedures()
(and thus indirectly getProcedureColumns()
)
getTables()
The catalog
argument in all of these methods now behaves in the following way:
Specifying NULL
means that catalog will not be used to filter the results (thus all databases will be searched), unless you've set nullCatalogMeansCurrent=true
in your JDBC URL properties.
Specifying ""
means “current” catalog, even though this isn't quite JDBC spec compliant, it's there for legacy users.
Specifying a catalog works as stated in the API docs.
Made Connection.clientPrepare()
available from “wrapped” connections in the jdbc2.optional
package (connections built by ConnectionPoolDataSource
instances).
Added Connection.isMasterConnection()
for clients to be able to determine if a multi-host master/slave connection is connected to the first host in the list.
Tokenizer for =
in URL properties was causing sessionVariables=....
to be parameterized incorrectly. (Bug#12753)
Foreign key information that is quoted is parsed incorrectly when DatabaseMetaData
methods use that information. (Bug#11781)
The sendBlobChunkSize
property is now clamped to max_allowed_packet
with consideration of stream buffer size and packet headers to avoid PacketTooBigExceptions
when max_allowed_packet
is similar in size to the default sendBlobChunkSize
which is 1M.
CallableStatement.clearParameters()
now clears resources associated with INOUT
/OUTPUT
parameters as well as INPUT
parameters.
Connection.prepareCall()
is database name case-sensitive (on Windows systems). (Bug#12417)
cp1251
incorrectly mapped to win1251
for servers newer than 4.0.x. (Bug#12752)
java.sql.Types.OTHER
returned for BINARY
and VARBINARY
columns when using DatabaseMetaData.getColumns()
. (Bug#12970)
ServerPreparedStatement.getBinding()
now checks if the statement is closed before attempting to reference the list of parameter bindings, to avoid throwing a NullPointerException
.
ResultSetMetaData
from Statement.getGeneratedKeys()
caused a NullPointerException
to be thrown whenever a method that required a connection reference was called. (Bug#13277)
Backport of Field
class, ResultSetMetaData.getColumnClassName()
, and ResultSet.getObject(int)
changes from 5.0 branch to fix behavior surrounding VARCHAR BINARY
/VARBINARY
and related types.
Fixed NullPointerException
when converting catalog
parameter in many DatabaseMetaDataMethods
to byte[]
s (for the result set) when the parameter is null
. (null
isn't technically allowed by the JDBC specification, but we've historically allowed it).
Backport of VAR[BINARY|CHAR] [BINARY]
types detection from 5.0 branch.
Read response in MysqlIO.sendFileToServer()
, even if the local file can't be opened, otherwise next query issued will fail, because it's reading the response to the empty LOAD DATA INFILE
packet sent to the server.
Workaround for Bug#13374: ResultSet.getStatement()
on closed result set returns NULL
(as per JDBC 4.0 spec, but not backward-compatible). Set the connection property retainStatementAfterResultSetClose
to true
to be able to retrieve a ResultSet
's statement after the ResultSet
has been closed via .getStatement()
(the default is false
, to be JDBC-compliant and to reduce the chance that code using JDBC leaks Statement
instances).
URL configuration parameters don't allow ‘&
’ or ‘=
’ in their values. The JDBC driver now parses configuration parameters as if they are encoded using the application/x-www-form-urlencoded format as specified by java.net.URLDecoder
(http://java.sun.com/j2se/1.5.0/docs/api/java/net/URLDecoder.html). (Bug#13453)
If the ‘%
’ character is present in a configuration property, it must now be represented as %25
, which is the encoded form of ‘%
’ when using application/x-www-form-urlencoded encoding.
The configuration property sessionVariables
now allows you to specify variables that start with the ‘@
’ sign.
When gatherPerfMetrics
is enabled for servers older than 4.1.0, a NullPointerException
is thrown from the constructor of ResultSet
if the query doesn't use any tables. (Bug#13043)
Fixed connecting without a database specified raised an exception in MysqlIO.changeDatabaseTo()
.
Initial implemention of ParameterMetadata
for PreparedStatement.getParameterMetadata()
. Only works fully for CallableStatements
, as current server-side prepared statements return every parameter as a VARCHAR
type.
Overhaul of character set configuration, everything now lives in a properties file.
Driver now correctly uses CP932 if available on the server for Windows-31J, CP932 and MS932 java encoding names, otherwise it resorts to SJIS, which is only a close approximation. Currently only MySQL-5.0.3 and newer (and MySQL-4.1.12 or .13, depending on when the character set gets backported) can reliably support any variant of CP932.
com.mysql.jdbc.PreparedStatement.ParseInfo
does unnecessary call to toCharArray()
. (Bug#9064)
Memory leak in ServerPreparedStatement
if serverPrepare()
fails. (Bug#10144)
Actually write manifest file to correct place so it ends up in the binary jar file.
Added createDatabaseIfNotExist
property (default is false
), which will cause the driver to ask the server to create the database specified in the URL if it doesn't exist. You must have the appropriate privileges for database creation for this to work.
Unsigned SMALLINT
treated as signed for ResultSet.getInt()
, fixed all cases for UNSIGNED
integer values and server-side prepared statements, as well as ResultSet.getObject()
for UNSIGNED TINYINT
. (Bug#10156)
Double quotes not recognized when parsing client-side prepared statements. (Bug#10155)
Made enableStreamingResults()
visible on com.mysql.jdbc.jdbc2.optional.StatementWrapper
.
Made ServerPreparedStatement.asSql()
work correctly so auto-explain functionality would work with server-side prepared statements.
Made JDBC2-compliant wrappers public in order to allow access to vendor extensions.
Cleaned up logging of profiler events, moved code to dump a profiler event as a string to com.mysql.jdbc.log.LogUtils
so that third parties can use it.
DatabaseMetaData.supportsMultipleOpenResults()
now returns true
. The driver has supported this for some time, DBMD just missed that fact.
Driver doesn't support {?=CALL(...)}
for calling stored functions. This involved adding support for function retrieval to DatabaseMetaData.getProcedures()
and getProcedureColumns()
as well. (Bug#10310)
SQLException
thrown when retrieving YEAR(2)
with ResultSet.getString()
. The driver will now always treat YEAR
types as java.sql.Dates
and return the correct values for getString()
. Alternatively, the yearIsDateType
connection property can be set to false
and the values will be treated as SHORT
s. (Bug#10485)
The datatype returned for TINYINT(1)
columns when tinyInt1isBit=true
(the default) can be switched between Types.BOOLEAN
and Types.BIT
using the new configuration property transformedBitIsBoolean
, which defaults to false
. If set to false
(the default), DatabaseMetaData.getColumns()
and ResultSetMetaData.getColumnType()
will return Types.BOOLEAN
for TINYINT(1)
columns. If true
, Types.BOOLEAN
will be returned instead. Regardless of this configuration property, if tinyInt1isBit
is enabled, columns with the type TINYINT(1)
will be returned as java.lang.Boolean
instances from ResultSet.getObject(...)
, and ResultSetMetaData.getColumnClassName()
will return java.lang.Boolean
.
SQLException
is thrown when using property characterSetResults
with cp932
or eucjpms
. (Bug#10496)
Reorganized directory layout. Sources now are in src
folder. Don't pollute parent directory when building, now output goes to ./build
, distribution goes to ./dist
.
Added support/bug hunting feature that generates .sql
test scripts to STDERR
when autoGenerateTestcaseScript
is set to true
.
0-length streams not sent to server when using server-side prepared statements. (Bug#10850)
Setting cachePrepStmts=true
now causes the Connection
to also cache the check the driver performs to determine if a prepared statement can be server-side or not, as well as caches server-side prepared statements for the lifetime of a connection. As before, the prepStmtCacheSize
parameter controls the size of these caches.
Try to handle OutOfMemoryErrors
more gracefully. Although not much can be done, they will in most cases close the connection they happened on so that further operations don't run into a connection in some unknown state. When an OOM has happened, any further operations on the connection will fail with a “Connection closed” exception that will also list the OOM exception as the reason for the implicit connection close event.
Don't send COM_RESET_STMT
for each execution of a server-side prepared statement if it isn't required.
Driver detects if you're running MySQL-5.0.7 or later, and does not scan for LIMIT ?[,?]
in statements being prepared, as the server supports those types of queries now.
VARBINARY
data corrupted when using server-side prepared statements and ResultSet.getBytes()
. (Bug#11115)
Connection.setCatalog()
is now aware of the useLocalSessionState
configuration property, which when set to true
will prevent the driver from sending USE ...
to the server if the requested catalog is the same as the current catalog.
Added the following configuration bundles, use one or many via the useConfigs
configuration property:
maxPerformance
— maximum performance without being reckless
solarisMaxPerformance
— maximum performance for Solaris, avoids syscalls where it can
3-0-Compat
— Compatibility with Connector/J 3.0.x functionality
Added maintainTimeStats
configuration property (defaults to true
), which tells the driver whether or not to keep track of the last query time and the last successful packet sent to the server's time. If set to false
, removes two syscalls per query.
autoReconnect
ping causes exception on connection startup. (Bug#11259)
Connector/J dumping query into SQLException
twice. (Bug#11360)
Fixed PreparedStatement.setClob()
not accepting null
as a parameter.
Production package doesn't include JBoss integration classes. (Bug#11411)
Removed nonsensical “costly type conversion” warnings when using usage advisor.
Fixed DatabaseMetaData.getTables()
returning views when they were not asked for as one of the requested table types.
Added support for new precision-math DECIMAL
type in MySQL 5.0.3 and up.
Fixed ResultSet.getTime()
on a NULL
value for server-side prepared statements throws NPE.
Made Connection.ping()
a public method.
DATE_FORMAT()
queries returned as BLOB
s from getObject()
. (Bug#8868)
ServerPreparedStatements
now correctly “stream” BLOB
/CLOB
data to the server. You can configure the threshold chunk size using the JDBC URL property blobSendChunkSize
(the default is 1MB).
BlobFromLocator
now uses correct identifier quoting when generating prepared statements.
Server-side session variables can be preset at connection time by passing them as a comma-delimited list for the connection property sessionVariables
.
Fixed regression in ping()
for users using autoReconnect=true
.
PreparedStatement.addBatch()
doesn't work with server-side prepared statements and streaming BINARY
data. (Bug#9040)
DBMD.supportsMixedCase*Identifiers()
returns wrong value on servers running on case-sensitive filesystems. (Bug#8800)
Cannot use UTF-8
for characterSetResults configuration property. (Bug#9206)
A continuation of Bug#8868, where functions used in queries that should return non-string types when resolved by temporary tables suddenly become opaque binary strings (work-around for server limitation). Also fixed fields with type of CHAR(n) CHARACTER SET BINARY
to return correct/matching classes for RSMD.getColumnClassName()
and ResultSet.getObject()
. (Bug#9236)
DBMD.supportsResultSetConcurrency()
not returning true
for forward-only/read-only result sets (we obviously support this). (Bug#8792)
DATA_TYPE
column from DBMD.getBestRowIdentifier()
causes ArrayIndexOutOfBoundsException
when accessed (and in fact, didn't return any value). (Bug#8803)
Check for empty strings (''
) when converting CHAR
/VARCHAR
column data to numbers, throw exception if emptyStringsConvertToZero
configuration property is set to false
(for backward-compatibility with 3.0, it is now set to true
by default, but will most likely default to false
in 3.2).
PreparedStatement.getMetaData()
inserts blank row in database under certain conditions when not using server-side prepared statements. (Bug#9320)
Connection.canHandleAsPreparedStatement()
now makes “best effort” to distinguish LIMIT
clauses with placeholders in them from ones without in order to have fewer false positives when generating work-arounds for statements the server cannot currently handle as server-side prepared statements.
Fixed build.xml
to not compile log4j
logging if log4j
not available.
Added support for the c3p0 connection pool's (http://c3p0.sf.net/) validation/connection checker interface which uses the lightweight COM_PING
call to the server if available. To use it, configure your c3p0 connection pool's connectionTesterClassName
property to use com.mysql.jdbc.integration.c3p0.MysqlConnectionTester
.
Better detection of LIMIT
inside/outside of quoted strings so that the driver can more correctly determine whether a prepared statement can be prepared on the server or not.
Stored procedures with same name in different databases confuse the driver when it tries to determine parameter counts/types. (Bug#9319)
Added finalizers to ResultSet
and Statement
implementations to be JDBC spec-compliant, which requires that if not explicitly closed, these resources should be closed upon garbage collection.
Stored procedures with DECIMAL
parameters with storage specifications that contained ‘,
’ in them would fail. (Bug#9682)
PreparedStatement.setObject(int, Object, int type, int scale)
now uses scale value for BigDecimal
instances.
Statement.getMoreResults()
could throw NPE when existing result set was .close()
d. (Bug#9704)
The performance metrics feature now gathers information about number of tables referenced in a SELECT.
The logging system is now automatically configured. If the value has been set by the user, via the URL property logger
or the system property com.mysql.jdbc.logger
, then use that, otherwise, autodetect it using the following steps:
Log4j, if it's available,
Then JDK1.4 logging,
Then fallback to our STDERR
logging.
DBMD.getTables()
shouldn't return tables if views are asked for, even if the database version doesn't support views. (Bug#9778)
Fixed driver not returning true
for -1
when ResultSet.getBoolean()
was called on result sets returned from server-side prepared statements.
Added a Manifest.MF
file with implementation information to the .jar
file.
More tests in Field.isOpaqueBinary()
to distinguish opaque binary (that is, fields with type CHAR(n)
and CHARACTER SET BINARY
) from output of various scalar and aggregate functions that return strings.
Should accept null
for catalog (meaning use current) in DBMD methods, even though it's not JDBC-compliant for legacy's sake. Disable by setting connection property nullCatalogMeansCurrent
to false
(which will be the default value in C/J 3.2.x). (Bug#9917)
Should accept null
for name patterns in DBMD (meaning ‘%
’), even though it isn't JDBC compliant, for legacy's sake. Disable by setting connection property nullNamePatternMatchesAll
to false
(which will be the default value in C/J 3.2.x). (Bug#9769)
Timestamp key column data needed _binary
stripped for UpdatableResultSet.refreshRow()
. (Bug#7686)
Timestamps converted incorrectly to strings with server-side prepared statements and updatable result sets. (Bug#7715)
Detect new sql_mode
variable in string form (it used to be integer) and adjust quoting method for strings appropriately.
Added holdResultsOpenOverStatementClose
property (default is false
), that keeps result sets open over statement.close() or new execution on same statement (suggested by Kevin Burton).
Infinite recursion when “falling back” to master in failover configuration. (Bug#7952)
Disable multi-statements (if enabled) for MySQL-4.1 versions prior to version 4.1.10 if the query cache is enabled, as the server returns wrong results in this configuration.
Fixed duplicated code in configureClientCharset()
that prevented useOldUTF8Behavior=true
from working properly.
Removed dontUnpackBinaryResults
functionality, the driver now always stores results from server-side prepared statements as is from the server and unpacks them on demand.
Emulated locators corrupt binary data when using server-side prepared statements. (Bug#8096)
Fixed synchronization issue with ServerPreparedStatement.serverPrepare()
that could cause deadlocks/crashes if connection was shared between threads.
By default, the driver now scans SQL you are preparing via all variants of Connection.prepareStatement()
to determine if it is a supported type of statement to prepare on the server side, and if it is not supported by the server, it instead prepares it as a client-side emulated prepared statement. You can disable this by passing emulateUnsupportedPstmts=false
in your JDBC URL. (Bug#4718)
Remove _binary
introducer from parameters used as in/out parameters in CallableStatement
.
Always return byte[]
s for output parameters registered as *BINARY
.
Send correct value for “boolean” true
to server for PreparedStatement.setObject(n, "true", Types.BIT)
.
Fixed bug with Connection not caching statements from prepareStatement()
when the statement wasn't a server-side prepared statement.
Choose correct “direction” to apply time adjustments when both client and server are in GMT time zone when using ResultSet.get(..., cal)
and PreparedStatement.set(...., cal)
.
Added dontTrackOpenResources
option (default is false
, to be JDBC compliant), which helps with memory use for non-well-behaved apps (that is, applications that don't close Statement
objects when they should).
ResultSet.getString()
doesn't maintain format stored on server, bug fix only enabled when noDatetimeStringSync
property is set to true
(the default is false
). (Bug#8428)
Fixed NPE in ResultSet.realClose()
when using usage advisor and result set was already closed.
PreparedStatements
not creating streaming result sets. (Bug#8487)
Don't pass NULL
to String.valueOf()
in ResultSet.getNativeConvertToString()
, as it stringifies it (that is, returns null
), which is not correct for the method in question.
ResultSet.getBigDecimal()
throws exception when rounding would need to occur to set scale. The driver now chooses a rounding mode of “half up” if non-rounding BigDecimal.setScale()
fails. (Bug#8424)
Added useLocalSessionState
configuration property, when set to true
the JDBC driver trusts that the application is well-behaved and only sets autocommit and transaction isolation levels using the methods provided on java.sql.Connection
, and therefore can manipulate these values in many cases without incurring round-trips to the database server.
Added enableStreamingResults()
to Statement
for connection pool implementations that check Statement.setFetchSize()
for specification-compliant values. Call Statement.setFetchSize(>=0)
to disable the streaming results for that statement.
Added support for BIT
type in MySQL-5.0.3. The driver will treat BIT(1-8)
as the JDBC standard BIT
type (which maps to java.lang.Boolean
), as the server does not currently send enough information to determine the size of a bitfield when < 9 bits are declared. BIT(>9)
will be treated as VARBINARY
, and will return byte[]
when getObject()
is called.
Fixed hang on SocketInputStream.read()
with Statement.setMaxRows()
and multiple result sets when driver has to truncate result set directly, rather than tacking a LIMIT
on the end of it.n
DBMD.getProcedures()
doesn't respect catalog parameter. (Bug#7026)
Fix comparisons made between string constants and dynamic strings that are converted with either toUpperCase()
or toLowerCase()
to use Locale.ENGLISH
, as some locales “override” case rules for English. Also use StringUtils.indexOfIgnoreCase()
instead of .toUpperCase().indexOf()
, avoids creating a very short-lived transient String
instance.
Server-side prepared statements did not honor zeroDateTimeBehavior
property, and would cause class-cast exceptions when using ResultSet.getObject()
, as the all-zero string was always returned. (Bug#5235)
Fixed batched updates with server prepared statements weren't looking if the types had changed for a given batched set of parameters compared to the previous set, causing the server to return the error “Wrong arguments to mysql_stmt_execute()”.
Handle case when string representation of timestamp contains trailing ‘.
’ with no numbers following it.
Inefficient detection of pre-existing string instances in ResultSet.getNativeString()
. (Bug#5706)
Don't throw exceptions for Connection.releaseSavepoint()
.
Use a per-session Calendar
instance by default when decoding dates from ServerPreparedStatements
(set to old, less performant behavior by setting property dynamicCalendars=true
).
Added experimental configuration property dontUnpackBinaryResults
, which delays unpacking binary result set values until they're asked for, and only creates object instances for non-numerical values (it is set to false
by default). For some usecase/jvm combinations, this is friendlier on the garbage collector.
UNSIGNED BIGINT
unpacked incorrectly from server-side prepared statement result sets. (Bug#5729)
ServerSidePreparedStatement
allocating short-lived objects unnecessarily. (Bug#6225)
Removed unwanted new Throwable()
in ResultSet
constructor due to bad merge (caused a new object instance that was never used for every result set created). Found while profiling for Bug#6359.
Fixed too-early creation of StringBuffer
in EscapeProcessor.escapeSQL()
, also return String
when escaping not needed (to avoid unnecessary object allocations). Found while profiling for Bug#6359.
Use null-safe-equals for key comparisons in updatable result sets.
SUM()
on DECIMAL
with server-side prepared statement ignores scale if zero-padding is needed (this ends up being due to conversion to DOUBLE
by server, which when converted to a string to parse into BigDecimal
, loses all “padding” zeros). (Bug#6537)
Use DatabaseMetaData.getIdentifierQuoteString()
when building DBMD queries.
Use 1MB packet for sending file for LOAD DATA LOCAL INFILE
if that is < max_allowed_packet
on server.
ResultSetMetaData.getColumnDisplaySize()
returns incorrect values for multi-byte charsets. (Bug#6399)
Make auto-deserialization of java.lang.Objects
stored in BLOB
columns configurable via autoDeserialize
property (defaults to false
).
Re-work Field.isOpaqueBinary()
to detect CHAR(
to support fixed-length binary fields for n
) CHARACTER SET BINARYResultSet.getObject()
.
Use our own implementation of buffered input streams to get around blocking behavior of java.io.BufferedInputStream
. Disable this with useReadAheadInput=false
.
Failing to connect to the server when one of the addresses for the given host name is IPV6 (which the server does not yet bind on). The driver now loops through all IP addresses for a given host, and stops on the first one that accepts()
a socket.connect()
. (Bug#6348)
Connector/J 3.1.3 beta does not handle integers correctly (caused by changes to support unsigned reads in Buffer.readInt()
-> Buffer.readShort()
). (Bug#4510)
Added support in DatabaseMetaData.getTables()
and getTableTypes()
for views, which are now available in MySQL server 5.0.x.
ServerPreparedStatement.execute*()
sometimes threw ArrayIndexOutOfBoundsException
when unpacking field metadata. (Bug#4642)
Optimized integer number parsing, enable “old” slower integer parsing using JDK classes via useFastIntParsing=false
property.
Added useOnlyServerErrorMessages
property, which causes message text in exceptions generated by the server to only contain the text sent by the server (as opposed to the SQLState's “standard” description, followed by the server's error message). This property is set to true
by default.
ResultSet.wasNull()
does not work for primatives if a previous null
was returned. (Bug#4689)
Track packet sequence numbers if enablePacketDebug=true
, and throw an exception if packets received out-of-order.
ResultSet.getObject()
returns wrong type for strings when using prepared statements. (Bug#4482)
Calling MysqlPooledConnection.close()
twice (even though an application error), caused NPE. Fixed.
ServerPreparedStatements
dealing with return of DECIMAL
type don't work. (Bug#5012)
ResultSet.getObject()
doesn't return type Boolean
for pseudo-bit types from prepared statements on 4.1.x (shortcut for avoiding extra type conversion when using binary-encoded result sets obscured test in getObject()
for “pseudo” bit type). (Bug#5032)
You can now use URLs in LOAD DATA LOCAL INFILE
statements, and the driver will use Java's built-in handlers for retreiving the data and sending it to the server. This feature is not enabled by default, you must set the allowUrlInLocalInfile
connection property to true
.
The driver is more strict about truncation of numerics on ResultSet.get*()
, and will throw an SQLException
when truncation is detected. You can disable this by setting jdbcCompliantTruncation
to false
(it is enabled by default, as this functionality is required for JDBC compliance).
Added three ways to deal with all-zero datetimes when reading them from a ResultSet
: exception
(the default), which throws an SQLException
with an SQLState of S1009
; convertToNull
, which returns NULL
instead of the date; and round
, which rounds the date to the nearest closest value which is '0001-01-01'
.
Fixed ServerPreparedStatement
to read prepared statement metadata off the wire, even though it's currently a placeholder instead of using MysqlIO.clearInputStream()
which didn't work at various times because data wasn't available to read from the server yet. This fixes sporadic errors users were having with ServerPreparedStatements
throwing ArrayIndexOutOfBoundExceptions
.
Use com.mysql.jdbc.Message
's classloader when loading resource bundle, should fix sporadic issues when the caller's classloader can't locate the resource bundle.
Mangle output parameter names for CallableStatements
so they will not clash with user variable names.
Added support for INOUT
parameters in CallableStatements
.
Null bitmask sent for server-side prepared statements was incorrect. (Bug#4119)
Use SQL Standard SQL states by default, unless useSqlStateCodes
property is set to false
.
Added packet debuging code (see the enablePacketDebug
property documentation).
Added constants for MySQL error numbers (publicly accessible, see com.mysql.jdbc.MysqlErrorNumbers
), and the ability to generate the mappings of vendor error codes to SQLStates that the driver uses (for documentation purposes).
Externalized more messages (on-going effort).
Error in retrieval of mediumint
column with prepared statements and binary protocol. (Bug#4311)
Support new time zone variables in MySQL-4.1.3 when useTimezone=true
.
Support for unsigned numerics as return types from prepared statements. This also causes a change in ResultSet.getObject()
for the bigint unsigned
type, which used to return BigDecimal
instances, it now returns instances of java.lang.BigInteger
.
Fixed stored procedure parameter parsing info when size was specified for a parameter (for example, char()
, varchar()
).
Enabled callable statement caching via cacheCallableStmts
property.
Fixed case when no output parameters specified for a stored procedure caused a bogus query to be issued to retrieve out parameters, leading to a syntax error from the server.
Fixed case when no parameters could cause a NullPointerException
in CallableStatement.setOutputParameters()
.
Removed wrapping of exceptions in MysqlIO.changeUser()
.
Fixed sending of split packets for large queries, enabled nio ability to send large packets as well.
Added .toString()
functionality to ServerPreparedStatement
, which should help if you're trying to debug a query that is a prepared statement (it shows SQL as the server would process).
Added gatherPerformanceMetrics
property, along with properties to control when/where this info gets logged (see docs for more info).
ServerPreparedStatements
weren't actually de-allocating server-side resources when .close()
was called.
Added logSlowQueries
property, along with slowQueriesThresholdMillis
property to control when a query should be considered “slow.”
Correctly map output parameters to position given in prepareCall()
versus. order implied during registerOutParameter()
. (Bug#3146)
Correctly detect initial character set for servers >= 4.1.0.
Cleaned up detection of server properties.
Support placeholder for parameter metadata for server >= 4.1.2.
getProcedures()
does not return any procedures in result set. (Bug#3539)
getProcedureColumns()
doesn't work with wildcards for procedure name. (Bug#3540)
DBMD.getSQLStateType()
returns incorrect value. (Bug#3520)
Added connectionCollation
property to cause driver to issue set collation_connection=...
query on connection init if default collation for given charset is not appropriate.
Fixed DatabaseMetaData.getProcedures()
when run on MySQL-5.0.0 (output of SHOW PROCEDURE STATUS
changed between 5.0.0 and 5.0.1.
getWarnings()
returns SQLWarning
instead of DataTruncation
. (Bug#3804)
Don't enable server-side prepared statements for server version 5.0.0 or 5.0.1, as they aren't compatible with the '4.1.2+' style that the driver uses (the driver expects information to come back that isn't there, so it hangs).
Fixed bug with UpdatableResultSets
not using client-side prepared statements.
Fixed character encoding issues when converting bytes to ASCII when MySQL doesn't provide the character set, and the JVM is set to a multi-byte encoding (usually affecting retrieval of numeric values).
Unpack “unknown” data types from server prepared statements as Strings
.
Implemented long data (Blobs, Clobs, InputStreams, Readers) for server prepared statements.
Implemented Statement.getWarnings()
for MySQL-4.1 and newer (using SHOW WARNINGS
).
Default result set type changed to TYPE_FORWARD_ONLY
(JDBC compliance).
Centralized setting of result set type and concurrency.
Refactored how connection properties are set and exposed as DriverPropertyInfo
as well as Connection
and DataSource
properties.
Support for NIO. Use useNIO=true
on platforms that support NIO.
Support for transaction savepoints (MySQL >= 4.0.14 or 4.1.1).
Support for mysql_change_user()
. See the changeUser()
method in com.mysql.jdbc.Connection
.
Reduced number of methods called in average query to be more efficient.
Prepared Statements
will be re-prepared on auto-reconnect. Any errors encountered are postponed until first attempt to re-execute the re-prepared statement.
Ensure that warnings are cleared before executing queries on prepared statements, as-per JDBC spec (now that we support warnings).
Support “old” profileSql
capitalization in ConnectionProperties
. This property is deprecated, you should use profileSQL
if possible.
Optimized Buffer.readLenByteArray()
to return shared empty byte array when length is 0.
Allow contents of PreparedStatement.setBlob()
to be retained between calls to .execute*()
.
Deal with 0-length tokens in EscapeProcessor
(caused by callable statement escape syntax).
Check for closed connection on delete/update/insert row operations in UpdatableResultSet
.
Fix support for table aliases when checking for all primary keys in UpdatableResultSet
.
Removed useFastDates
connection property.
Correctly initialize datasource properties from JNDI Refs, including explicitly specified URLs.
DatabaseMetaData
now reports supportsStoredProcedures()
for MySQL versions >= 5.0.0
Fixed stack overflow in Connection.prepareCall()
(bad merge).
Fixed IllegalAccessError
to Calendar.getTimeInMillis()
in DateTimeValue
(for JDK < 1.4).
DatabaseMetaData.getColumns()
is not returning correct column ordinal info for non-'%'
column name patterns. (Bug#1673)
Merged fix of datatype mapping from MySQL type FLOAT
to java.sql.Types.REAL
from 3.0 branch.
Detect collation of column for RSMD.isCaseSensitive()
.
Fixed sending of queries larger than 16M.
Added named and indexed input/output parameter support to CallableStatement
. MySQL-5.0.x or newer.
Fixed NullPointerException
in ServerPreparedStatement.setTimestamp()
, as well as year and month descrepencies in ServerPreparedStatement.setTimestamp()
, setDate()
.
Added ability to have multiple database/JVM targets for compliance and regression/unit tests in build.xml
.
Fixed NPE and year/month bad conversions when accessing some datetime functionality in ServerPreparedStatements
and their resultant result sets.
Display where/why a connection was implicitly closed (to aid debugging).
CommunicationsException
implemented, that tries to determine why communications was lost with a server, and displays possible reasons when .getMessage()
is called.
NULL
values for numeric types in binary encoded result sets causing NullPointerExceptions
. (Bug#2359)
Implemented Connection.prepareCall()
, and DatabaseMetaData
. getProcedures()
and getProcedureColumns()
.
Reset long binary
parameters in ServerPreparedStatement
when clearParameters()
is called, by sending COM_RESET_STMT
to the server.
Merged prepared statement caching, and .getMetaData()
support from 3.0 branch.
Fixed off-by-1900 error in some cases for years in TimeUtil.fastDate
/TimeCreate()
when unpacking results from server-side prepared statements.
Fixed charset conversion issue in getTables()
. (Bug#2502)
Implemented multiple result sets returned from a statement or stored procedure.
Server-side prepared statements were not returning datatype YEAR
correctly. (Bug#2606)
Enabled streaming of result sets from server-side prepared statements.
Class-cast exception when using scrolling result sets and server-side prepared statements. (Bug#2623)
Merged unbuffered input code from 3.0.
Fixed ConnectionProperties
that weren't properly exposed via accessors, cleaned up ConnectionProperties
code.
NULL
fields were not being encoded correctly in all cases in server-side prepared statements. (Bug#2671)
Fixed rare buffer underflow when writing numbers into buffers for sending prepared statement execution requests.
Use DocBook version of docs for shipped versions of drivers.
Added requireSSL
property.
Added useServerPrepStmts
property (default false
). The driver will use server-side prepared statements when the server version supports them (4.1 and newer) when this property is set to true
. It is currently set to false
by default until all bind/fetch functionality has been implemented. Currently only DML prepared statements are implemented for 4.1 server-side prepared statements.
Track open Statements
, close all when Connection.close()
is called (JDBC compliance).
Timestamp
/Time
conversion goes in the wrong “direction” when useTimeZone=true
and server time zone differs from client time zone. (Bug#5874)
DatabaseMetaData.getIndexInfo()
ignored unique
parameter. (Bug#7081)
Support new protocol type MYSQL_TYPE_VARCHAR
.
Added useOldUTF8Behavior
' configuration property, which causes JDBC driver to act like it did with MySQL-4.0.x and earlier when the character encoding is utf-8
when connected to MySQL-4.1 or newer.
Statements created from a pooled connection were returning physical connection instead of logical connection when getConnection()
was called. (Bug#7316)
PreparedStatements
don't encode Big5 (and other multi-byte) character sets correctly in static SQL strings. (Bug#7033)
Connections starting up failed-over (due to down master) never retry master. (Bug#6966)
PreparedStatement.fixDecimalExponent()
adding extra +
, making number unparseable by MySQL server. (Bug#7061)
Timestamp key column data needed _binary
stripped for UpdatableResultSet.refreshRow()
. (Bug#7686)
Backported SQLState codes mapping from Connector/J 3.1, enable with useSqlStateCodes=true
as a connection property, it defaults to false
in this release, so that we don't break legacy applications (it defaults to true
starting with Connector/J 3.1).
PreparedStatement.fixDecimalExponent()
adding extra +
, making number unparseable by MySQL server. (Bug#7601)
Escape sequence {fn convert(..., type)} now supports ODBC-style types that are prepended by SQL_
.
Fixed duplicated code in configureClientCharset()
that prevented useOldUTF8Behavior=true
from working properly.
Handle streaming result sets with more than 2 billion rows properly by fixing wraparound of row number counter.
MS932
, SHIFT_JIS
, and Windows_31J
not recognized as aliases for sjis
. (Bug#7607)
Adding CP943
to aliases for sjis
. (Bug#6549, fixed while fixing Bug#7607)
Which requires hex escaping of binary data when using multi-byte charsets with prepared statements. (Bug#8064)
NON_UNIQUE
column from DBMD.getIndexInfo()
returned inverted value. (Bug#8812)
Workaround for server Bug#9098: Default values of CURRENT_*
for DATE
, TIME
, DATETIME
, and TIMESTAMP
columns can't be distinguished from string
values, so UpdatableResultSet.moveToInsertRow()
generates bad SQL for inserting default values.
EUCKR
charset is sent as SET NAMES euc_kr
which MySQL-4.1 and newer doesn't understand. (Bug#8629)
DatabaseMetaData.supportsSelectForUpdate()
returns correct value based on server version.
Use hex escapes for PreparedStatement.setBytes()
for double-byte charsets including “aliases” Windows-31J
, CP934
, MS932
.
Added support for the EUC_JP_Solaris
character encoding, which maps to a MySQL encoding of eucjpms
(backported from 3.1 branch). This only works on servers that support eucjpms
, namely 5.0.3 or later.
Re-issue character set configuration commands when re-using pooled connections and/or Connection.changeUser()
when connected to MySQL-4.1 or newer.
Fixed ResultSetMetaData.isReadOnly()
to detect non-writable columns when connected to MySQL-4.1 or newer, based on existence of “original” table and column names.
ResultSet.updateByte()
when on insert row throws ArrayOutOfBoundsException
. (Bug#5664)
Fixed DatabaseMetaData.getTypes()
returning incorrect (this is, non-negative) scale for the NUMERIC
type.
Off-by-one bug in Buffer.readString(
. (Bug#5664)string
)
Made TINYINT(1)
-> BIT
/Boolean
conversion configurable via tinyInt1isBit
property (default true
to be JDBC compliant out of the box).
Only set character_set_results
during connection establishment if server version >= 4.1.1.
Fixed regression where useUnbufferedInput
was defaulting to false
.
ResultSet.getTimestamp()
on a column with TIME
in it fails. (Bug#5664)
StringUtils.escapeEasternUnicodeByteStream
was still broken for GBK. (Bug#4010)
Failover for autoReconnect
not using port numbers for any hosts, and not retrying all hosts. (Warning: This required a change to the SocketFactory
connect()
method signature, which is now public Socket connect(String host, int portNumber, Properties props)
; therefore, any third-party socket factories will have to be changed to support this signature. (Bug#4334)
Logical connections created by MysqlConnectionPoolDataSource
will now issue a rollback()
when they are closed and sent back to the pool. If your application server/connection pool already does this for you, you can set the rollbackOnPooledClose
property to false
to avoid the overhead of an extra rollback()
.
Removed redundant calls to checkRowPos()
in ResultSet
.
DOUBLE
mapped twice in DBMD.getTypeInfo()
. (Bug#4742)
Added FLOSS license exemption.
Calling .close()
twice on a PooledConnection
causes NPE. (Bug#4808)
DBMD.getColumns()
returns incorrect JDBC type for unsigned columns. This affects type mappings for all numeric types in the RSMD.getColumnType()
and RSMD.getColumnTypeNames()
methods as well, to ensure that “like” types from DBMD.getColumns()
match up with what RSMD.getColumnType()
and getColumnTypeNames()
return. (Bug#4138, Bug#4860)
“Production” is now “GA” (General Availability) in naming scheme of distributions.
RSMD.getPrecision()
returning 0 for non-numeric types (should return max length in chars for non-binary types, max length in bytes for binary types). This fix also fixes mapping of RSMD.getColumnType()
and RSMD.getColumnTypeName()
for the BLOB
types based on the length sent from the server (the server doesn't distinguish between TINYBLOB
, BLOB
, MEDIUMBLOB
or LONGBLOB
at the network protocol level). (Bug#4880)
ResultSet
should release Field[]
instance in .close()
. (Bug#5022)
ResultSet.getMetaData()
should not return incorrectly initialized metadata if the result set has been closed, but should instead throw an SQLException
. Also fixed for getRow()
and getWarnings()
and traversal methods by calling checkClosed()
before operating on instance-level fields that are nullified during .close()
. (Bug#5069)
Parse new time zone variables from 4.1.x servers.
Use _binary
introducer for PreparedStatement.setBytes()
and set*Stream()
when connected to MySQL-4.1.x or newer to avoid misinterpretation during character conversion.
Add unsigned attribute to DatabaseMetaData.getColumns()
output in the TYPE_NAME
column.
Added failOverReadOnly
property, to allow end-user to configure state of connection (read-only/writable) when failed over.
Backported “change user” and “reset server state” functionality from 3.1 branch, to allow clients of MysqlConnectionPoolDataSource
to reset server state on getConnection()
on a pooled connection.
Don't escape SJIS/GBK/BIG5 when using MySQL-4.1 or newer.
Allow url
parameter for MysqlDataSource
and MysqlConnectionPool
DataSource
so that passing of other properties is possible from inside appservers.
Map duplicate key and foreign key errors to SQLState of 23000
.
Backport documentation tooling from 3.1 branch.
Return creating statement for ResultSets
created by getGeneratedKeys()
. (Bug#2957)
Allow java.util.Date
to be sent in as parameter to PreparedStatement.setObject()
, converting it to a Timestamp
to maintain full precision. (Bug#103).
Don't truncate BLOB
or CLOB
values when using setBytes()
and/or setBinary/CharacterStream()
. (Bug#2670).
Dynamically configure character set mappings for field-level character sets on MySQL-4.1.0 and newer using SHOW COLLATION
when connecting.
Map binary
character set to US-ASCII
to support DATETIME
charset recognition for servers >= 4.1.2.
Use SET character_set_results
during initialization to allow any charset to be returned to the driver for result sets.
Use charsetnr
returned during connect to encode queries before issuing SET NAMES
on MySQL >= 4.1.0.
Add helper methods to ResultSetMetaData
(getColumnCharacterEncoding()
and getColumnCharacterSet()
) to allow end-users to see what charset the driver thinks it should be using for the column.
Only set character_set_results
for MySQL >= 4.1.0.
StringUtils.escapeSJISByteStream()
not covering all eastern double-byte charsets correctly. (Bug#3511)
Renamed StringUtils.escapeSJISByteStream()
to more appropriate escapeEasternUnicodeByteStream()
.
Not specifying database in URL caused MalformedURL
exception. (Bug#3554)
Auto-convert MySQL encoding names to Java encoding names if used for characterEncoding
property.
Added encoding names that are recognized on some JVMs to fix case where they were reverse-mapped to MySQL encoding names incorrectly.
Use junit.textui.TestRunner
for all unit tests (to allow them to be run from the command line outside of Ant or Eclipse).
UpdatableResultSet
not picking up default values for moveToInsertRow()
. (Bug#3557)
Inconsistent reporting of data type. The server still doesn't return all types for *BLOBs *TEXT correctly, so the driver won't return those correctly. (Bug#3570)
DBMD.getSQLStateType()
returns incorrect value. (Bug#3520)
Fixed regression in PreparedStatement.setString()
and eastern character encodings.
Made StringRegressionTest
4.1-unicode aware.
Trigger a SET NAMES utf8
when encoding is forced to utf8
or utf-8
via the characterEncoding
property. Previously, only the Java-style encoding name of utf-8
would trigger this.
AutoReconnect
time was growing faster than exponentially. (Bug#2447)
Fixed failover always going to last host in list. (Bug#2578)
Added useUnbufferedInput
parameter, and now use it by default (due to JVM issue http://developer.java.sun.com/developer/bugParade/bugs/4401235.html)
Detect on
/off
or 1
, 2
, 3
form of lower_case_table_names
value on server.
Return java.lang.Integer
for TINYINT
and SMALLINT
types from ResultSetMetaData.getColumnClassName()
. (Bug#2852)
Return java.lang.Double
for FLOAT
type from ResultSetMetaData.getColumnClassName()
. (Bug#2855)
Return [B
instead of java.lang.Object
for BINARY
, VARBINARY
and LONGVARBINARY
types from ResultSetMetaData.getColumnClassName()
(JDBC compliance).
Issue connection events on all instances created from a ConnectionPoolDataSource
.
Don't count quoted IDs when inside a 'string' in PreparedStatement
parsing. (Bug#1511)
“Friendlier” exception message for PacketTooLargeException
. (Bug#1534)
Backported fix for aliased tables and UpdatableResultSets
in checkUpdatability()
method from 3.1 branch.
Fix for ArrayIndexOutOfBounds
exception when using Statement.setMaxRows()
. (Bug#1695)
Barge blobs and split packets not being read correctly. (Bug#1576)
Fixed regression of Statement.getGeneratedKeys()
and REPLACE
statements.
Subsequent call to ResultSet.updateFoo()
causes NPE if result set is not updatable. (Bug#1630)
Fix for 4.1.1-style authentication with no password.
Foreign Keys column sequence is not consistent in DatabaseMetaData.getImported/Exported/CrossReference()
. (Bug#1731)
DatabaseMetaData.getSystemFunction()
returning bad function VResultsSion
. (Bug#1775)
Cross-database updatable result sets are not checked for updatability correctly. (Bug#1592)
DatabaseMetaData.getColumns()
should return Types.LONGVARCHAR
for MySQL LONGTEXT
type.
ResultSet.getObject()
on TINYINT
and SMALLINT
columns should return Java type Integer
. (Bug#1913)
Added alwaysClearStream
connection property, which causes the driver to always empty any remaining data on the input stream before each query.
Added more descriptive error message Server Configuration Denies Access to DataSource
, as well as retrieval of message from server.
Autoreconnect code didn't set catalog upon reconnect if it had been changed.
Implement ResultSet.updateClob()
.
ResultSetMetaData.isCaseSensitive()
returned wrong value for CHAR
/VARCHAR
columns.
Connection property maxRows
not honored. (Bug#1933)
Statements being created too many times in DBMD.extractForeignKeyFromCreateTable()
. (Bug#1925)
Support escape sequence {fn convert ... }. (Bug#1914)
ArrayIndexOutOfBounds
when parameter number == number of parameters + 1. (Bug#1958)
ResultSet.findColumn()
should use first matching column name when there are duplicate column names in SELECT
query (JDBC-compliance). (Bug#2006)
Removed static synchronization bottleneck from PreparedStatement.setTimestamp()
.
Removed static synchronization bottleneck from instance factory method of SingleByteCharsetConverter
.
Enable caching of the parsing stage of prepared statements via the cachePrepStmts
, prepStmtCacheSize
, and prepStmtCacheSqlLimit
properties (disabled by default).
Speed up parsing of PreparedStatements
, try to use one-pass whenever possible.
Fixed security exception when used in Applets (applets can't read the system property file.encoding
which is needed for LOAD DATA LOCAL INFILE
).
Use constants for SQLStates.
Map charset ko18_ru
to ko18r
when connected to MySQL-4.1.0 or newer.
Ensure that Buffer.writeString()
saves room for the \0
.
Fixed exception Unknown character set 'danish'
on connect with JDK-1.4.0
Fixed mappings in SQLError to report deadlocks with SQLStates of 41000
.
maxRows
property would affect internal statements, so check it for all statement creation internal to the driver, and set to 0 when it is not.
Faster date handling code in ResultSet
and PreparedStatement
(no longer uses Date
methods that synchronize on static calendars).
Fixed test for end of buffer in Buffer.readString()
.
Fixed ResultSet.previous()
behavior to move current position to before result set when on first row of result set. (Bug#496)
Fixed Statement
and PreparedStatement
issuing bogus queries when setMaxRows()
had been used and a LIMIT
clause was present in the query.
refreshRow
didn't work when primary key values contained values that needed to be escaped (they ended up being doubly escaped). (Bug#661)
Support InnoDB
contraint names when extracting foreign key information in DatabaseMetaData
(implementing ideas from Parwinder Sekhon). (Bug#517, Bug#664)
Backported 4.1 protocol changes from 3.1 branch (server-side SQL states, new field information, larger client capability flags, connect-with-database, and so forth).
Fix UpdatableResultSet
to return values for get
when on insert row. (Bug#675)XXX
()
The insertRow
in an UpdatableResultSet
is now loaded with the default column values when moveToInsertRow()
is called. (Bug#688)
DatabaseMetaData.getColumns()
wasn't returning NULL
for default values that are specified as NULL
.
Change default statement type/concurrency to TYPE_FORWARD_ONLY
and CONCUR_READ_ONLY
(spec compliance).
Don't try and reset isolation level on reconnect if MySQL doesn't support them.
Don't wrap SQLExceptions
in RowDataDynamic
.
Don't change timestamp TZ twice if useTimezone==true
. (Bug#774)
Fixed regression in large split-packet handling. (Bug#848)
Better diagnostic error messages in exceptions for “streaming” result sets.
Issue exception on ResultSet.get
on empty result set (wasn't caught in some cases).XXX
()
Don't hide messages from exceptions thrown in I/O layers.
Don't fire connection closed events when closing pooled connections, or on PooledConnection.getConnection()
with already open connections. (Bug#884)
Clip +/- INF (to smallest and largest representative values for the type in MySQL) and NaN (to 0) for setDouble
/setFloat()
, and issue a warning on the statement when the server does not support +/- INF or NaN.
Double-escaping of '\'
when charset is SJIS or GBK and '\'
appears in non-escaped input. (Bug#879)
When emptying input stream of unused rows for “streaming” result sets, have the current thread yield()
every 100 rows in order to not monopolize CPU time.
DatabaseMetaData.getColumns()
getting confused about the keyword “set” in character columns. (Bug#1099)
Fixed deadlock issue with Statement.setMaxRows()
.
Fixed CLOB.truncate()
. (Bug#1130)
Optimized CLOB.setChracterStream()
. (Bug#1131)
Made databaseName
, portNumber
, and serverName
optional parameters for MysqlDataSourceFactory
. (Bug#1246)
ResultSet.get/setString
mashing char 127. (Bug#1247)
Backported authentication changes for 4.1.1 and newer from 3.1 branch.
Added com.mysql.jdbc.util.BaseBugReport
to help creation of testcases for bug reports.
Added property to “clobber” streaming results, by setting the clobberStreamingResults
property to true
(the default is false
). This will cause a “streaming” ResultSet
to be automatically closed, and any oustanding data still streaming from the server to be discarded if another query is executed before all the data has been read from the server.
Allow bogus URLs in Driver.getPropertyInfo()
.
Return list of generated keys when using multi-value INSERTS
with Statement.getGeneratedKeys()
.
Use JVM charset with filenames and LOAD DATA [LOCAL] INFILE
.
Fix infinite loop with Connection.cleanup()
.
Changed Ant target compile-core
to compile-driver
, and made testsuite compilation a separate target.
Fixed result set not getting set for Statement.executeUpdate()
, which affected getGeneratedKeys()
and getUpdateCount()
in some cases.
Unicode character 0xFFFF in a string would cause the driver to throw an ArrayOutOfBoundsException
. (Bug#378).
Return correct number of generated keys when using REPLACE
statements.
Fix problem detecting server character set in some cases.
Fix row data decoding error when using very large packets.
Optimized row data decoding.
Issue exception when operating on an already closed prepared statement.
Fixed SJIS encoding bug, thanks to Naoto Sato.
Optimized usage of EscapeProcessor
.
Allow multiple calls to Statement.close()
.
Fixed MysqlPooledConnection.close()
calling wrong event type.
Fixed StringIndexOutOfBoundsException
in PreparedStatement.setClob()
.
4.1 Column Metadata fixes.
Remove synchronization from Driver.connect()
and Driver.acceptsUrl()
.
IOExceptions
during a transaction now cause the Connection
to be closed.
Fixed missing conversion for YEAR
type in ResultSetMetaData.getColumnTypeName()
.
Don't pick up indexes that start with pri
as primary keys for DBMD.getPrimaryKeys()
.
Throw SQLExceptions
when trying to do operations on a forcefully closed Connection
(that is, when a communication link failure occurs).
You can now toggle profiling on/off using Connection.setProfileSql(boolean)
.
Fixed charset issues with database metadata (charset was not getting set correctly).
Updatable ResultSets
can now be created for aliased tables/columns when connected to MySQL-4.1 or newer.
Fixed LOAD DATA LOCAL INFILE
bug when file > max_allowed_packet
.
Fixed escaping of 0x5c ('\'
) character for GBK and Big5 charsets.
Fixed ResultSet.getTimestamp()
when underlying field is of type DATE
.
Ensure that packet size from alignPacketSize()
does not exceed max_allowed_packet
(JVM bug)
Don't reset Connection.isReadOnly()
when autoReconnecting.
Fixed ResultSetMetaData
to return ""
when catalog not known. Fixes NullPointerExceptions
with Sun's CachedRowSet
.
Fixed DBMD.getTypeInfo()
and DBMD.getColumns()
returning different value for precision in TEXT
and BLOB
types.
Allow ignoring of warning for “non transactional tables” during rollback (compliance/usability) by setting ignoreNonTxTables
property to true
.
Fixed SQLExceptions
getting swallowed on initial connect.
Fixed Statement.setMaxRows()
to stop sending LIMIT
type queries when not needed (performance).
Clean up Statement
query/method mismatch tests (that is, INSERT
not allowed with .executeQuery()
).
More checks added in ResultSet
traversal method to catch when in closed state.
Fixed ResultSetMetaData.isWritable()
to return correct value.
Add “window” of different NULL
sorting behavior to DBMD.nullsAreSortedAtStart
(4.0.2 to 4.0.10, true; otherwise, no).
Implemented Blob.setBytes()
. You still need to pass the resultant Blob
back into an updatable ResultSet
or PreparedStatement
to persist the changes, because MySQL does not support “locators”.
Backported 4.1 charset field info changes from Connector/J 3.1.
Fixed Buffer.fastSkipLenString()
causing ArrayIndexOutOfBounds
exceptions with some queries when unpacking fields.
Implemented an empty TypeMap
for Connection.getTypeMap()
so that some third-party apps work with MySQL (IBM WebSphere 5.0 Connection pool).
Added missing LONGTEXT
type to DBMD.getColumns()
.
Retrieve TX_ISOLATION
from database for Connection.getTransactionIsolation()
when the MySQL version supports it, instead of an instance variable.
Quote table names in DatabaseMetaData.getColumns()
, getPrimaryKeys()
, getIndexInfo()
, getBestRowIdentifier()
.
Greatly reduce memory required for setBinaryStream()
in PreparedStatements
.
Fixed ResultSet.isBeforeFirst()
for empty result sets.
Added update options for foreign key metadata.
Added quoted identifiers to database names for Connection.setCatalog
.
Added support for quoted identifiers in PreparedStatement
parser.
Streamlined character conversion and byte[]
handling in PreparedStatements
for setByte()
.
Reduce memory footprint of PreparedStatements
by sharing outbound packet with MysqlIO
.
Added strictUpdates
property to allow control of amount of checking for “correctness” of updatable result sets. Set this to false
if you want faster updatable result sets and you know that you create them from SELECT
statements on tables with primary keys and that you have selected all primary keys in your query.
Added support for 4.0.8-style large packets.
Fixed PreparedStatement.executeBatch()
parameter overwriting.
Changed charsToByte
in SingleByteCharConverter
to be non-static.
Changed SingleByteCharConverter
to use lazy initialization of each converter.
Fixed charset handling in Fields.java
.
Implemented Connection.nativeSQL()
.
More robust escape tokenizer: Recognize --
comments, and allow nested escape sequences (see testsuite.EscapeProcessingTest
).
DBMD.getImported/ExportedKeys()
now handles multiple foreign keys per table.
Fixed ResultSetMetaData.getPrecision()
returning incorrect values for some floating-point types.
Fixed ResultSetMetaData.getColumnTypeName()
returning BLOB
for TEXT
and TEXT
for BLOB
types.
Fixed Buffer.isLastDataPacket()
for 4.1 and newer servers.
Added CLIENT_LONG_FLAG
to be able to get more column flags (isAutoIncrement()
being the most important).
Because of above, implemented ResultSetMetaData.isAutoIncrement()
to use Field.isAutoIncrement()
.
Honor lower_case_table_names
when enabled in the server when doing table name comparisons in DatabaseMetaData
methods.
Some MySQL-4.1 protocol support (extended field info from selects).
Use non-aliased table/column names and database names to fullly qualify tables and columns in UpdatableResultSet
(requires MySQL-4.1 or newer).
Allow user to alter behavior of Statement
/ PreparedStatement.executeBatch()
via continueBatchOnError
property (defaults to true
).
Check for connection closed in more Connection
methods (createStatement
, prepareStatement
, setTransactionIsolation
, setAutoCommit
).
More robust implementation of updatable result sets. Checks that all primary keys of the table have been selected.
LOAD DATA LOCAL INFILE ...
now works, if your server is configured to allow it. Can be turned off with the allowLoadLocalInfile
property (see the README
).
Substitute '?'
for unknown character conversions in single-byte character sets instead of '\0'
.
NamedPipeSocketFactory
now works (only intended for Windows), see README
for instructions.
Fixed issue with updatable result sets and PreparedStatements
not working.
Fixed ResultSet.setFetchDirection(FETCH_UNKNOWN)
.
Fixed issue when calling Statement.setFetchSize()
when using arbitrary values.
Fixed incorrect conversion in ResultSet.getLong()
.
Implemented ResultSet.updateBlob()
.
Removed duplicate code from UpdatableResultSet
(it can be inherited from ResultSet
, the extra code for each method to handle updatability I thought might someday be necessary has not been needed).
Fixed UnsupportedEncodingException
thrown when “forcing” a character encoding via properties.
Fixed various non-ASCII character encoding issues.
Added driver property useHostsInPrivileges
. Defaults to true
. Affects whether or not @hostname
will be used in DBMD.getColumn/TablePrivileges
.
All DBMD
result set columns describing schemas now return NULL
to be more compliant with the behavior of other JDBC drivers for other database systems (MySQL does not support schemas).
Added SSL support. See README
for information on how to use it.
Properly restore connection properties when autoReconnecting or failing-over, including autoCommit
state, and isolation level.
Use SHOW CREATE TABLE
when possible for determining foreign key information for DatabaseMetaData
. Also allows cascade options for DELETE
information to be returned.
Escape 0x5c
character in strings for the SJIS charset.
Fixed start position off-by-1 error in Clob.getSubString()
.
Implemented Clob.truncate()
.
Implemented Clob.setString()
.
Implemented Clob.setAsciiStream()
.
Implemented Clob.setCharacterStream()
.
Added com.mysql.jdbc.MiniAdmin
class, which allows you to send shutdown
command to MySQL server. This is intended to be used when “embedding” Java and MySQL server together in an end-user application.
Added connectTimeout
parameter that allows users of JDK-1.4 and newer to specify a maxium time to wait to establish a connection.
Failover and autoReconnect
work only when the connection is in an autoCommit(false)
state, in order to stay transaction-safe.
Added queriesBeforeRetryMaster
property that specifies how many queries to issue when failed over before attempting to reconnect to the master (defaults to 50).
Fixed DBMD.supportsResultSetConcurrency()
so that it returns true
for ResultSet.TYPE_SCROLL_INSENSITIVE
and ResultSet.CONCUR_READ_ONLY
or ResultSet.CONCUR_UPDATABLE
.
Fixed ResultSet.isLast()
for empty result sets (should return false
).
PreparedStatement
now honors stream lengths in setBinary/Ascii/Character Stream() unless you set the connection property useStreamLengthsInPrepStmts
to false
.
Removed some not-needed temporary object creation by smarter use of Strings
in EscapeProcessor
, Connection
and DatabaseMetaData
classes.
Fixed ResultSet.getRow()
off-by-one bug.
Fixed RowDataStatic.getAt()
off-by-one bug.
Added limited Clob
functionality (ResultSet.getClob()
, PreparedStatemtent.setClob()
, PreparedStatement.setObject(Clob)
.
Added socketTimeout
parameter to URL.
Connection.isClosed()
no longer “pings” the server.
Connection.close()
issues rollback()
when getAutoCommit()
is false
.
Added paranoid
parameter, which sanitizes error messages by removing “sensitive” information from them (such as hostnames, ports, or usernames), as well as clearing “sensitive” data structures when possible.
Fixed ResultSetMetaData.isSigned()
for TINYINT
and BIGINT
.
Charsets now automatically detected. Optimized code for single-byte character set conversion.
Implemented ResultSet.getCharacterStream()
.
Added LOCAL TEMPORARY
to table types in DatabaseMetaData.getTableTypes()
.
Massive code clean-up to follow Java coding conventions (the time had come).
!!! LICENSE CHANGE !!! The driver is now GPL. If you need non-GPL licenses, please contact me <mark@mysql.com>
.
JDBC-3.0 functionality including Statement/PreparedStatement.getGeneratedKeys()
and ResultSet.getURL()
.
Performance enchancements: Driver is now 50–100% faster in most situations, and creates fewer temporary objects.
Repackaging: New driver name is com.mysql.jdbc.Driver
, old name still works, though (the driver is now provided by MySQL-AB).
Better checking for closed connections in Statement
and PreparedStatement
.
Support for streaming (row-by-row) result sets (see README
) Thanks to Doron.
Support for large packets (new addition to MySQL-4.0 protocol), see README
for more information.
JDBC Compliance: Passes all tests besides stored procedure tests.
Fix and sort primary key names in DBMetaData
(SF bugs 582086 and 582086).
Float types now reported as java.sql.Types.FLOAT
(SF bug 579573).
ResultSet.getTimestamp()
now works for DATE
types (SF bug 559134).
ResultSet.getDate/Time/Timestamp
now recognizes all forms of invalid values that have been set to all zeros by MySQL (SF bug 586058).
Testsuite now uses Junit (which you can get from http://www.junit.org.
The driver now only works with JDK-1.2 or newer.
Added multi-host failover support (see README
).
General source-code cleanup.
Overall speed improvements via controlling transient object creation in MysqlIO
class when reading packets.
Performance improvements in string handling and field metadata creation (lazily instantiated) contributed by Alex Twisleton-Wykeham-Fiennes.
More code cleanup.
PreparedStatement
now releases resources on .close()
. (SF bug 553268)
Quoted identifiers not used if server version does not support them. Also, if server started with --ansi
or --sql-mode=ANSI_QUOTES
, ‘"
’ will be used as an identifier quote character, otherwise ‘'
’ will be used.
ResultSet.getDouble()
now uses code built into JDK to be more precise (but slower).
LogicalHandle.isClosed()
calls through to physical connection.
Added SQL profiling (to STDERR
). Set profileSql=true
in your JDBC URL. See README
for more information.
Fixed typo for relaxAutoCommit
parameter.
More code cleanup.
Fixed unicode chars being read incorrectly. (SF bug 541088)
Faster blob escaping for PrepStmt
.
Added set
/getPortNumber()
to DataSource(s)
. (SF bug 548167)
Added setURL()
to MySQLXADataSource
. (SF bug 546019)
PreparedStatement.toString()
fixed. (SF bug 534026)
ResultSetMetaData.getColumnClassName()
now implemented.
Rudimentary version of Statement.getGeneratedKeys()
from JDBC-3.0 now implemented (you need to be using JDK-1.4 for this to work, I believe).
DBMetaData.getIndexInfo()
- bad PAGES fixed. (SF BUG 542201)
General code cleanup.
Added getIdleFor()
method to Connection
and MysqlLogicalHandle
.
Relaxed synchronization in all classes, should fix 520615 and 520393.
Added getTable/ColumnPrivileges()
to DBMD (fixes 484502).
Added new types to getTypeInfo()
, fixed existing types thanks to Al Davis and Kid Kalanon.
Added support for BIT
types (51870) to PreparedStatement
.
Fixed getRow()
bug (527165) in ResultSet
.
Fixes for ResultSet
updatability in PreparedStatement
.
Fixed time zone off-by-1-hour bug in PreparedStatement
(538286, 528785).
ResultSet
: Fixed updatability (values being set to null
if not updated).
DataSources
- fixed setUrl
bug (511614, 525565), wrong datasource class name (532816, 528767).
Added identifier quoting to all DatabaseMetaData
methods that need them (should fix 518108).
Added support for YEAR
type (533556).
ResultSet.insertRow()
should now detect auto_increment fields in most cases and use that value in the new row. This detection will not work in multi-valued keys, however, due to the fact that the MySQL protocol does not return this information.
ResultSet.refreshRow()
implemented.
Fixed testsuite.Traversal
afterLast()
bug, thanks to Igor Lastric.
Fixed missing DELETE_RULE
value in DBMD.getImported/ExportedKeys()
and getCrossReference()
.
Full synchronization of Statement.java
.
More changes to fix Unexpected end of input stream
errors when reading BLOB
values. This should be the last fix.
Fixed spurious Unexpected end of input stream
errors in MysqlIO
(bug 507456).
Fixed null-pointer-exceptions when using MysqlConnectionPoolDataSource
with Websphere 4 (bug 505839).
Ant build was corrupting included jar
files, fixed (bug 487669).
Fixed extra memory allocation in MysqlIO.readPacket()
(bug 488663).
Implementation of DatabaseMetaData.getExported/ImportedKeys()
and getCrossReference()
.
Full synchronization on methods modifying instance and class-shared references, driver should be entirely thread-safe now (please let me know if you have problems).
DataSource
implementations moved to org.gjt.mm.mysql.jdbc2.optional
package, and (initial) implementations of PooledConnectionDataSource
and XADataSource
are in place (thanks to Todd Wolff for the implementation and testing of PooledConnectionDataSource
with IBM WebSphere 4).
Added detection of network connection being closed when reading packets (thanks to Todd Lizambri).
Fixed quoting error with escape processor (bug 486265).
Report batch update support through DatabaseMetaData
(bug 495101).
Fixed off-by-one-hour error in PreparedStatement.setTimestamp()
(bug 491577).
Removed concatenation support from driver (the ||
operator), as older versions of VisualAge seem to be the only thing that use it, and it conflicts with the logical ||
operator. You will need to start mysqld with the --ansi
flag to use the ||
operator as concatenation (bug 491680).
Fixed casting bug in PreparedStatement
(bug 488663).
Batch updates now supported (thanks to some inspiration from Daniel Rall).
XADataSource
/ConnectionPoolDataSource
code (experimental)
PreparedStatement.setAnyNumericType()
now handles positive exponents correctly (adds +
so MySQL can understand it).
DatabaseMetaData.getPrimaryKeys()
and getBestRowIdentifier()
are now more robust in identifying primary keys (matches regardless of case or abbreviation/full spelling of Primary Key
in Key_type
column).
PreparedStatement.setCharacterStream()
now implemented
Fixed dangling socket problem when in high availability (autoReconnect=true
) mode, and finalizer for Connection
will close any dangling sockets on GC.
Fixed ResultSetMetaData.getPrecision()
returning one less than actual on newer versions of MySQL.
ResultSet.getBlob()
now returns null
if column value was null
.
Character sets read from database if useUnicode=true
and characterEncoding
is not set. (thanks to Dmitry Vereshchagin)
Initial transaction isolation level read from database (if avaialable). (thanks to Dmitry Vereshchagin)
Fixed DatabaseMetaData.supportsTransactions()
, and supportsTransactionIsolationLevel()
and getTypeInfo()
SQL_DATETIME_SUB
and SQL_DATA_TYPE
fields not being readable.
Fixed PreparedStatement
generating SQL that would end up with syntax errors for some queries.
Fixed ResultSet.isAfterLast()
always returning false
.
Fixed time zone issue in PreparedStatement.setTimestamp()
. (thanks to Erik Olofsson)
Captialize type names when captializeTypeNames=true
is passed in URL or properties (for WebObjects. (thanks to Anjo Krank)
Updatable result sets now correctly handle NULL
values in fields.
PreparedStatement.setDouble() now uses full-precision doubles (reverting a fix made earlier to truncate them).
PreparedStatement.setBoolean() will use 1/0 for values if your MySQL version is 3.21.23 or higher.
Fixed PreparedStatement
parameter checking.
Fixed case-sensitive column names in ResultSet.java
.
Fixed ResultSet.getBlob()
ArrayIndex
out-of-bounds.
Fixed ResultSetMetaData.getColumnTypeName
for TEXT
/BLOB
.
Fixed ArrayIndexOutOfBounds
when sending large BLOB
queries. (Max size packet was not being set)
Added ISOLATION
level support to Connection.setIsolationLevel()
Fixed NPE on PreparedStatement.executeUpdate()
when all columns have not been set.
Fixed data parsing of TIMESTAMP
values with 2-digit years.
Added Byte
to PreparedStatement.setObject()
.
ResultSet.getBoolean()
now recognizes -1
as true
.
ResultSet
has +/-Inf/inf support.
ResultSet.insertRow()
works now, even if not all columns are set (they will be set to NULL
).
DataBaseMetaData.getCrossReference()
no longer ArrayIndexOOB
.
getObject()
on ResultSet
correctly does TINYINT
->Byte
and SMALLINT
->Short
.
Implemented getBigDecimal()
without scale component for JDBC2.
Fixed composite key problem with updatable result sets.
Added detection of -/+INF for doubles.
Faster ASCII string operations.
Fixed incorrect detection of MAX_ALLOWED_PACKET
, so sending large blobs should work now.
Fixed off-by-one error in java.sql.Blob
implementation code.
Added ultraDevHack
URL parameter, set to true
to allow (broken) Macromedia UltraDev to use the driver.
Fixed RSMD.isWritable()
returning wrong value. Thanks to Moritz Maass.
Cleaned up exception handling when driver connects.
Columns that are of type TEXT
now return as Strings
when you use getObject()
.
DatabaseMetaData.getPrimaryKeys()
now works correctly with respect to key_seq
. Thanks to Brian Slesinsky.
No escape processing is done on PreparedStatements
anymore per JDBC spec.
Fixed many JDBC-2.0 traversal, positioning bugs, especially with respect to empty result sets. Thanks to Ron Smits, Nick Brook, Cessar Garcia and Carlos Martinez.
Fixed some issues with updatability support in ResultSet
when using multiple primary keys.
Fixes to ResultSet for insertRow() - Thanks to Cesar Garcia
Fix to Driver to recognize JDBC-2.0 by loading a JDBC-2.0 class, instead of relying on JDK version numbers. Thanks to John Baker.
Fixed ResultSet to return correct row numbers
Statement.getUpdateCount() now returns rows matched, instead of rows actually updated, which is more SQL-92 like.
10-29-99
Statement/PreparedStatement.getMoreResults() bug fixed. Thanks to Noel J. Bergman.
Added Short as a type to PreparedStatement.setObject(). Thanks to Jeff Crowder
Driver now automagically configures maximum/preferred packet sizes by querying server.
Autoreconnect code uses fast ping command if server supports it.
Fixed various bugs with respect to packet sizing when reading from the server and when alloc'ing to write to the server.
Now compiles under JDK-1.2. The driver supports both JDK-1.1 and JDK-1.2 at the same time through a core set of classes. The driver will load the appropriate interface classes at runtime by figuring out which JVM version you are using.
Fixes for result sets with all nulls in the first row. (Pointed out by Tim Endres)
Fixes to column numbers in SQLExceptions in ResultSet (Thanks to Blas Rodriguez Somoza)
The database no longer needs to specified to connect. (Thanks to Christian Motschke)
Better Documentation (in progress), in doc/mm.doc/book1.html
DBMD now allows null for a column name pattern (not in spec), which it changes to '%'.
DBMD now has correct types/lengths for getXXX().
ResultSet.getDate(), getTime(), and getTimestamp() fixes. (contributed by Alan Wilken)
EscapeProcessor now handles \{ \} and { or } inside quotes correctly. (thanks to Alik for some ideas on how to fix it)
Fixes to properties handling in Connection. (contributed by Juho Tikkala)
ResultSet.getObject() now returns null for NULL columns in the table, rather than bombing out. (thanks to Ben Grosman)
ResultSet.getObject() now returns Strings for types from MySQL that it doesn't know about. (Suggested by Chris Perdue)
Removed DataInput/Output streams, not needed, 1/2 number of method calls per IO operation.
Use default character encoding if one is not specified. This is a work-around for broken JVMs, because according to spec, EVERY JVM must support "ISO8859_1", but they don't.
Fixed Connection to use the platform character encoding instead of "ISO8859_1" if one isn't explicitly set. This fixes problems people were having loading the character- converter classes that didn't always exist (JVM bug). (thanks to Fritz Elfert for pointing out this problem)
Changed MysqlIO to re-use packets where possible to reduce memory usage.
Fixed escape-processor bugs pertaining to {} inside quotes.
Fixed character-set support for non-Javasoft JVMs (thanks to many people for pointing it out)
Fixed ResultSet.getBoolean() to recognize 'y' & 'n' as well as '1' & '0' as boolean flags. (thanks to Tim Pizey)
Fixed ResultSet.getTimestamp() to give better performance. (thanks to Richard Swift)
Fixed getByte() for numeric types. (thanks to Ray Bellis)
Fixed DatabaseMetaData.getTypeInfo() for DATE type. (thanks to Paul Johnston)
Fixed EscapeProcessor for "fn" calls. (thanks to Piyush Shah at locomotive.org)
Fixed EscapeProcessor to not do extraneous work if there are no escape codes. (thanks to Ryan Gustafson)
Fixed Driver to parse URLs of the form "jdbc:mysql://host:port" (thanks to Richard Lobb)
Fixed Timestamps for PreparedStatements
Fixed null pointer exceptions in RSMD and RS
Re-compiled with jikes for valid class files (thanks ms!)
Fixed escape processor to deal with unmatched { and } (thanks to Craig Coles)
Fixed escape processor to create more portable (between DATETIME and TIMESTAMP types) representations so that it will work with BETWEEN clauses. (thanks to Craig Longman)
MysqlIO.quit() now closes the socket connection. Before, after many failed connections some OS's would run out of file descriptors. (thanks to Michael Brinkman)
Fixed NullPointerException in Driver.getPropertyInfo. (thanks to Dave Potts)
Fixes to MysqlDefs to allow all *text fields to be retrieved as Strings. (thanks to Chris at Leverage)
Fixed setDouble in PreparedStatement for large numbers to avoid sending scientific notation to the database. (thanks to J.S. Ferguson)
Fixed getScale() and getPrecision() in RSMD. (contrib'd by James Klicman)
Fixed getObject() when field was DECIMAL or NUMERIC (thanks to Bert Hobbs)
DBMD.getTables() bombed when passed a null table-name pattern. Fixed. (thanks to Richard Lobb)
Added check for "client not authorized" errors during connect. (thanks to Hannes Wallnoefer)
Result set rows are now byte arrays. Blobs and Unicode work bidriectonally now. The useUnicode and encoding options are implemented now.
Fixes to PreparedStatement to send binary set by setXXXStream to be sent untouched to the MySQL server.
Fixes to getDriverPropertyInfo().
Changed all ResultSet fields to Strings, this should allow Unicode to work, but your JVM must be able to convert between the character sets. This should also make reading data from the server be a bit quicker, because there is now no conversion from StringBuffer to String.
Changed PreparedStatement.streamToString() to be more efficient (code from Uwe Schaefer).
URL parsing is more robust (throws SQL exceptions on errors rather than NullPointerExceptions)
PreparedStatement now can convert Strings to Time/Date values via setObject() (code from Robert Currey).
IO no longer hangs in Buffer.readInt(), that bug was introduced in 1.1d when changing to all byte-arrays for result sets. (Pointed out by Samo Login)
Fixes to DatabaseMetaData to allow both IBM VA and J-Builder to work. Let me know how it goes. (thanks to Jac Kersing)
Fix to ResultSet.getBoolean() for NULL strings (thanks to Barry Lagerweij)
Beginning of code cleanup, and formatting. Getting ready to branch this off to a parallel JDBC-2.0 source tree.
Added "final" modifier to critical sections in MysqlIO and Buffer to allow compiler to inline methods for speed.
9-29-98
If object references passed to setXXX() in PreparedStatement are null, setNull() is automatically called for you. (Thanks for the suggestion goes to Erik Ostrom)
setObject() in PreparedStatement will now attempt to write a serialized representation of the object to the database for objects of Types.OTHER and objects of unknown type.
Util now has a static method readObject() which given a ResultSet and a column index will re-instantiate an object serialized in the above manner.
Got rid of "ugly hack" in MysqlIO.nextRow(). Rather than catch an exception, Buffer.isLastDataPacket() was fixed.
Connection.getCatalog() and Connection.setCatalog() should work now.
Statement.setMaxRows() works, as well as setting by property maxRows. Statement.setMaxRows() overrides maxRows set via properties or url parameters.
Automatic re-connection is available. Because it has to "ping" the database before each query, it is turned off by default. To use it, pass in "autoReconnect=true" in the connection URL. You may also change the number of reconnect tries, and the initial timeout value via "maxReconnects=n" (default 3) and "initialTimeout=n" (seconds, default 2) parameters. The timeout is an exponential backoff type of timeout; for example, if you have initial timeout of 2 seconds, and maxReconnects of 3, then the driver will timeout 2 seconds, 4 seconds, then 16 seconds between each re-connection attempt.
Fixed handling of blob data in Buffer.java
Fixed bug with authentication packet being sized too small.
The JDBC Driver is now under the LPGL
8-14-98
Fixed Buffer.readLenString() to correctly read data for BLOBS.
Fixed PreparedStatement.stringToStream to correctly read data for BLOBS.
Fixed PreparedStatement.setDate() to not add a day. (above fixes thanks to Vincent Partington)
Added URL parameter parsing (?user=... and so forth).
Big news! New package name. Tim Endres from ICE Engineering is starting a new source tree for GNU GPL'd Java software. He's graciously given me the org.gjt.mm package directory to use, so now the driver is in the org.gjt.mm.mysql package scheme. I'm "legal" now. Look for more information on Tim's project soon.
Now using dynamically sized packets to reduce memory usage when sending commands to the DB.
Small fixes to getTypeInfo() for parameters, and so forth.
DatabaseMetaData is now fully implemented. Let me know if these drivers work with the various IDEs out there. I've heard that they're working with JBuilder right now.
Added JavaDoc documentation to the package.
Package now available in .zip or .tar.gz.
Implemented getTypeInfo(). Connection.rollback() now throws an SQLException per the JDBC spec.
Added PreparedStatement that supports all JDBC API methods for PreparedStatement including InputStreams. Please check this out and let me know if anything is broken.
Fixed a bug in ResultSet that would break some queries that only returned 1 row.
Fixed bugs in DatabaseMetaData.getTables(), DatabaseMetaData.getColumns() and DatabaseMetaData.getCatalogs().
Added functionality to Statement that allows executeUpdate() to store values for IDs that are automatically generated for AUTO_INCREMENT fields. Basically, after an executeUpdate(), look at the SQLWarnings for warnings like "LAST_INSERTED_ID = 'some number', COMMAND = 'your SQL query'". If you are using AUTO_INCREMENT fields in your tables and are executing a lot of executeUpdate()s on one Statement, be sure to clearWarnings() every so often to save memory.
Split MysqlIO and Buffer to separate classes. Some ClassLoaders gave an IllegalAccess error for some fields in those two classes. Now mm.mysql works in applets and all classloaders. Thanks to Joe Ennis <jce@mail.boone.com> for pointing out the problem and working on a fix with me.
Fixed DatabaseMetadata problems in getColumns() and bug in switch statement in the Field constructor. Thanks to Costin Manolache <costin@tdiinc.com> for pointing these out.
Incorporated efficiency changes from Richard Swift <Richard.Swift@kanatek.ca> in MysqlIO.java
and ResultSet.java
:
We're now 15% faster than gwe's driver.
Started working on DatabaseMetaData
.
The following methods are implemented:
getTables()
getTableTypes()
getColumns
getCatalogs()
Source code editor What Is Ajax
→