[TAP] New fix for the transaction management.

The transaction and Statement were closed too early before.
  - Fetching the row was not possible once the first bunch of fetched
rows was over.
  - The problem of "statement is aborted" preventing the re-use of
a same DB connection was apparently still there, but occurred less often.

  Now, any transaction potentially started in a DB connection is always
closed after one of the public functions of JDBCConnection is called ;
except executeQuery(ADQLQuery) whose the call MUST be wrapped inside a
try...catch block in which DBConnection.cancel(true) MUST be called
in case of error (in order to effectively end any started transaction).

src/tap/ADQLExecutor.java

@@ -16,7 +16,7 @@ package tap;
  * You should have received a copy of the GNU Lesser General Public License
  * along with TAPLibrary.  If not, see <http://www.gnu.org/licenses/>.
  * 
- * Copyright 2012-2015 - UDS/Centre de Données astronomiques de Strasbourg (CDS),
+ * Copyright 2012-2016 - UDS/Centre de Données astronomiques de Strasbourg (CDS),
  *                       Astronomisches Rechen Institut (ARI)
  */
 
@@ -25,6 +25,10 @@ import java.io.OutputStream;
 
 import javax.servlet.http.HttpServletResponse;
 
+import adql.parser.ADQLParser;
+import adql.parser.ADQLQueryFactory;
+import adql.parser.ParseException;
+import adql.query.ADQLQuery;
 import tap.data.DataReadException;
 import tap.data.TableIterator;
 import tap.db.DBConnection;
@@ -41,10 +45,6 @@ import uws.UWSToolBox;
 import uws.job.JobThread;
 import uws.job.Result;
 import uws.service.log.UWSLog.LogLevel;
-import adql.parser.ADQLParser;
-import adql.parser.ADQLQueryFactory;
-import adql.parser.ParseException;
-import adql.query.ADQLQuery;
 
 /**
  * <p>Let process completely an ADQL query.</p>
@@ -86,7 +86,7 @@ import adql.query.ADQLQuery;
  * <p><i>Note:
  * 	{@link #start()} is using the Template Method Design Pattern: it defines the skeleton/algorithm of the processing, and defers some steps
  * 	to other functions.
- * </i></p> 
+ * </i></p>
  * 
  * <p>
  * 	So, you are able to customize almost all individual steps of the ADQL query processing: {@link #parseADQL()}, {@link #executeADQL(ADQLQuery)} and
@@ -104,7 +104,7 @@ import adql.query.ADQLQuery;
  * </p>
  * 
  * @author Gr&eacute;gory Mantelet (CDS;ARI)
- * @version 2.1 (11/2015)
+ * @version 2.1 (04/2016)
  */
 public class ADQLExecutor {
 
@@ -260,7 +260,7 @@ public class ADQLExecutor {
 			dbConn = service.getFactory().getConnection(jobID);
 	}
 
-	/** 
+	/**
 	 * Cancel the current SQL query execution or result set fetching if any is currently running.
 	 * If no such process is on going, this function has no effect.
 	 * 
@@ -324,7 +324,7 @@ public class ADQLExecutor {
 	 * 	notify the user of the progression of the query execution. This parameter is removed at the end of the execution if it is successful.
 	 * </p>
 	 * 
-	 * <p>The "interrupted" flag of the associated thread is often tested so that stopping the execution as soon as possible.</p> 
+	 * <p>The "interrupted" flag of the associated thread is often tested so that stopping the execution as soon as possible.</p>
 	 * 
 	 * @return	The updated execution report.
 	 * 
@@ -419,6 +419,9 @@ public class ADQLExecutor {
 
 			// Free the connection (so that giving it back to a pool, if any, otherwise, just free resources):
 			if (dbConn != null){
+				// be sure no query is still processing and that any open transaction is rollbacked and ended:
+				dbConn.cancel(true);
+				// free the connection:
 				service.getFactory().freeConnection(dbConn);
 				dbConn = null;
 			}

src/tap/config/ConfigurableTAPFactory.java

@@ -49,6 +49,9 @@ import javax.naming.InitialContext;
 import javax.naming.NamingException;
 import javax.sql.DataSource;
 
+import adql.translator.JDBCTranslator;
+import adql.translator.PgSphereTranslator;
+import adql.translator.PostgreSQLTranslator;
 import tap.AbstractTAPFactory;
 import tap.ServiceConnection;
 import tap.TAPException;
@@ -60,9 +63,6 @@ import uws.UWSException;
 import uws.service.UWSService;
 import uws.service.backup.UWSBackupManager;
 import uws.service.log.UWSLog.LogLevel;
-import adql.translator.JDBCTranslator;
-import adql.translator.PgSphereTranslator;
-import adql.translator.PostgreSQLTranslator;
 
 /**
  * <p>Concrete implementation of a {@link TAPFactory} which is parameterized by a TAP configuration file.</p>
@@ -74,7 +74,7 @@ import adql.translator.PostgreSQLTranslator;
  * </p>
  * 
  * @author Gr&eacute;gory Mantelet (ARI)
- * @version 2.1 (02/2016)
+ * @version 2.1 (04/2016)
  * @since 2.0
  */
 public class ConfigurableTAPFactory extends AbstractTAPFactory {
@@ -282,7 +282,7 @@ public class ConfigurableTAPFactory extends AbstractTAPFactory {
 	public void freeConnection(DBConnection conn){
 		try{
 			// Cancel any possible query that could be running:
-			conn.cancel(false);
+			conn.cancel(true);
 			// Close the connection (if a connection pool is used, the connection is not really closed but is freed and kept in the pool for further usage):
 			((JDBCConnection)conn).getInnerConnection().close();
 		}catch(SQLException se){

src/tap/data/ResultSetTableIterator.java

@@ -16,7 +16,7 @@ package tap.data;
  * You should have received a copy of the GNU Lesser General Public License
  * along with TAPLibrary.  If not, see <http://www.gnu.org/licenses/>.
  * 
- * Copyright 2014-2015 - Astronomisches Rechen Institut (ARI)
+ * Copyright 2014-2016 - Astronomisches Rechen Institut (ARI)
  */
 
 import java.sql.ResultSet;
@@ -26,14 +26,14 @@ import java.sql.Statement;
 import java.sql.Timestamp;
 import java.util.NoSuchElementException;
 
-import tap.metadata.TAPColumn;
-import uws.ISO8601Format;
 import adql.db.DBColumn;
 import adql.db.DBType;
 import adql.db.DBType.DBDatatype;
 import adql.db.STCS.Region;
 import adql.parser.ParseException;
 import adql.translator.JDBCTranslator;
+import tap.metadata.TAPColumn;
+import uws.ISO8601Format;
 
 /**
  * <p>{@link TableIterator} which lets iterate over a SQL {@link ResultSet}.</p>
@@ -43,7 +43,7 @@ import adql.translator.JDBCTranslator;
  * </i></p>
  * 
  * @author Gr&eacute;gory Mantelet (ARI)
- * @version 2.1 (11/2015)
+ * @version 2.1 (04/2016)
  * @since 2.0
  */
 public class ResultSetTableIterator implements TableIterator {
@@ -154,7 +154,7 @@ public class ResultSetTableIterator implements TableIterator {
 	 * 	The second parameter of this constructor is given as second parameter of {@link #convertType(int, String, String)}.
 	 * 	<b>This parameter is really used ONLY when the DBMS is SQLite ("sqlite").</b>
 	 * 	Indeed, SQLite has so many datatype restrictions that it is absolutely needed to know it is the DBMS from which the
-	 * 	ResultSet is coming. Without this information, type guessing will be unpredictable! 
+	 * 	ResultSet is coming. Without this information, type guessing will be unpredictable!
 	 * </i></p>
 	 * 
 	 * @param dataSet		Dataset over which this iterator must iterate.
@@ -191,7 +191,7 @@ public class ResultSetTableIterator implements TableIterator {
 	 * 	The second parameter of this constructor is given as second parameter of {@link #convertType(int, String, String)}.
 	 * 	<b>This parameter is really used ONLY when the DBMS is SQLite ("sqlite").</b>
 	 * 	Indeed, SQLite has so many datatype restrictions that it is absolutely needed to know it is the DBMS from which the
-	 * 	ResultSet is coming. Without this information, type guessing will be unpredictable! 
+	 * 	ResultSet is coming. Without this information, type guessing will be unpredictable!
 	 * </i></p>
 	 * 
 	 * @param dataSet		Dataset over which this iterator must iterate.
@@ -229,7 +229,7 @@ public class ResultSetTableIterator implements TableIterator {
 	 * 
 	 * @param dataSet		Dataset over which this iterator must iterate.
 	 * @param translator	The {@link JDBCTranslator} used to transform the ADQL query into SQL query. This translator is also able to convert
-	 *                  	JDBC types and to parse geometrical values. <i>note: MAY be NULL</i> 
+	 *                  	JDBC types and to parse geometrical values. <i>note: MAY be NULL</i>
 	 * 
 	 * @throws NullPointerException	If NULL is given in parameter.
 	 * @throws DataReadException	If the given ResultSet is closed or if the metadata (columns count and types) can not be fetched.
@@ -261,7 +261,7 @@ public class ResultSetTableIterator implements TableIterator {
 	 * 
 	 * @param dataSet		Dataset over which this iterator must iterate.
 	 * @param translator	The {@link JDBCTranslator} used to transform the ADQL query into SQL query. This translator is also able to convert
-	 *                  	JDBC types and to parse geometrical values. <i>note: MAY be NULL</i> 
+	 *                  	JDBC types and to parse geometrical values. <i>note: MAY be NULL</i>
 	 * 
 	 * @throws NullPointerException	If NULL is given in parameter.
 	 * @throws DataReadException	If the given ResultSet is closed or if the metadata (columns count and types) can not be fetched.
@@ -297,12 +297,12 @@ public class ResultSetTableIterator implements TableIterator {
 	 * 	The third parameter of this constructor is given as second parameter of {@link #convertType(int, String, String)}.
 	 * 	<b>This parameter is really used ONLY when the translator conversion failed and when the DBMS is SQLite ("sqlite").</b>
 	 * 	Indeed, SQLite has so many datatype restrictions that it is absolutely needed to know it is the DBMS from which the
-	 * 	ResultSet is coming. Without this information, type guessing will be unpredictable! 
+	 * 	ResultSet is coming. Without this information, type guessing will be unpredictable!
 	 * </i></p>
 	 * 
 	 * @param dataSet		Dataset over which this iterator must iterate.
 	 * @param translator	The {@link JDBCTranslator} used to transform the ADQL query into SQL query. This translator is also able to convert
-	 *                  	JDBC types and to parse geometrical values. <i>note: MAY be NULL</i> 
+	 *                  	JDBC types and to parse geometrical values. <i>note: MAY be NULL</i>
 	 * @param dbms			Lower-case string which indicates from which DBMS the given ResultSet is coming. <i>note: MAY be NULL.</i>
 	 * 
 	 * @throws NullPointerException	If NULL is given in parameter.
@@ -337,12 +337,12 @@ public class ResultSetTableIterator implements TableIterator {
 	 * 	The third parameter of this constructor is given as second parameter of {@link #convertType(int, String, String)}.
 	 * 	<b>This parameter is really used ONLY when the translator conversion failed and when the DBMS is SQLite ("sqlite").</b>
 	 * 	Indeed, SQLite has so many datatype restrictions that it is absolutely needed to know it is the DBMS from which the
-	 * 	ResultSet is coming. Without this information, type guessing will be unpredictable! 
+	 * 	ResultSet is coming. Without this information, type guessing will be unpredictable!
 	 * </i></p>
 	 * 
 	 * @param dataSet		Dataset over which this iterator must iterate.
 	 * @param translator	The {@link JDBCTranslator} used to transform the ADQL query into SQL query. This translator is also able to convert
-	 *                  	JDBC types and to parse geometrical values. <i>note: MAY be NULL</i> 
+	 *                  	JDBC types and to parse geometrical values. <i>note: MAY be NULL</i>
 	 * @param dbms			Lower-case string which indicates from which DBMS the given ResultSet is coming. <i>note: MAY be NULL.</i>
 	 * 
 	 * @throws NullPointerException	If NULL is given in parameter.
@@ -395,12 +395,12 @@ public class ResultSetTableIterator implements TableIterator {
 	 * 	The third parameter of this constructor is given as second parameter of {@link #convertType(int, String, String)}.
 	 * 	<b>This parameter is really used ONLY when the translator conversion failed and when the DBMS is SQLite ("sqlite").</b>
 	 * 	Indeed, SQLite has so many datatype restrictions that it is absolutely needed to know it is the DBMS from which the
-	 * 	ResultSet is coming. Without this information, type guessing will be unpredictable! 
+	 * 	ResultSet is coming. Without this information, type guessing will be unpredictable!
 	 * </i></p>
 	 * 
 	 * @param dataSet		Dataset over which this iterator must iterate.
 	 * @param translator	The {@link JDBCTranslator} used to transform the ADQL query into SQL query. This translator is also able to convert
-	 *                  	JDBC types and to parse geometrical values. <i>note: MAY be NULL</i> 
+	 *                  	JDBC types and to parse geometrical values. <i>note: MAY be NULL</i>
 	 * @param dbms			Lower-case string which indicates from which DBMS the given ResultSet is coming. <i>note: MAY be NULL.</i>
 	 * @param resultMeta	List of expected columns. <i>note: these metadata are expected to be really {@link TAPColumn} objects ; MAY be NULL.</i>
 	 * 
@@ -452,12 +452,12 @@ public class ResultSetTableIterator implements TableIterator {
 	 * 	The third parameter of this constructor is given as second parameter of {@link #convertType(int, String, String)}.
 	 * 	<b>This parameter is really used ONLY when the translator conversion failed and when the DBMS is SQLite ("sqlite").</b>
 	 * 	Indeed, SQLite has so many datatype restrictions that it is absolutely needed to know it is the DBMS from which the
-	 * 	ResultSet is coming. Without this information, type guessing will be unpredictable! 
+	 * 	ResultSet is coming. Without this information, type guessing will be unpredictable!
 	 * </i></p>
 	 * 
 	 * @param dataSet		Dataset over which this iterator must iterate.
 	 * @param translator	The {@link JDBCTranslator} used to transform the ADQL query into SQL query. This translator is also able to convert
-	 *                  	JDBC types and to parse geometrical values. <i>note: MAY be NULL</i> 
+	 *                  	JDBC types and to parse geometrical values. <i>note: MAY be NULL</i>
 	 * @param dbms			Lower-case string which indicates from which DBMS the given ResultSet is coming. <i>note: MAY be NULL.</i>
 	 * @param resultMeta	List of expected columns. <i>note: these metadata are expected to be really {@link TAPColumn} objects ; MAY be NULL.</i>
 	 * 
@@ -688,7 +688,8 @@ public class ResultSetTableIterator implements TableIterator {
 			return new DBType(DBDatatype.UNKNOWN);
 
 		// Extract the type prefix and lower-case it:
-		int startParamIndex = dbmsTypeName.indexOf('('), endParamIndex = dbmsTypeName.indexOf(')');
+		int startParamIndex = dbmsTypeName.indexOf('('),
+				endParamIndex = dbmsTypeName.indexOf(')');
 		String dbmsTypePrefix = (startParamIndex <= 0) ? dbmsTypeName : dbmsTypeName.substring(0, endParamIndex);
 		dbmsTypePrefix = dbmsTypePrefix.trim().toLowerCase();
 		String[] typeParams = (startParamIndex <= 0) ? null : dbmsTypeName.substring(startParamIndex + 1, endParamIndex).split(",");