Currently the generator outputs comments like this:
{code}
comment on column T_JES_COMMANDATTEMPTS is 'Type:Integer,Anzahl der Retries';
{code}
Notice the missing dot / point to separate the table name (T_JES_COMMAND) and the column name (ATTEMPTS).
The problem can be fixed by overriding the applyComments method of the StandardTableExporter class as follows:
{code}
public class MobiOracle10gDialect extends Oracle10gDialect {
@Override
public boolean supportsUniqueConstraintInCreateAlterTable() {
return false;
}
@Override
public Exporter<Table> getTableExporter() {
return new StandardTableExporter(this){
protected void applyComments(Table table, QualifiedName tableName, List<String> sqlStrings) {
if ( dialect.supportsCommentOn() ) {
if ( table.getComment() != null ) {
sqlStrings.add( "comment on table " + tableName + " is '" + table.getComment() + "'" );
}
final Iterator iter = table.getColumnIterator();
while ( iter.hasNext() ) {
Column column = (Column) iter.next();
String columnComment = column.getComment();
if ( columnComment != null ) {
//this line is buggy in hibernate 5 :-(
//the point is missing between table name and column!
sqlStrings.add( "comment on column " + tableName + "." + column.getQuotedName( dialect ) + " is '" + columnComment + "'" );
}
}
}
}
};
}
}
{code}
Compare this line:
{code}
sqlStrings.add( "comment on column " + tableName + "." + column.getQuotedName( dialect ) + " is '" + columnComment + "'" );
{code}
with the original:
{code}
sqlStrings.add( "comment on column " + tableName + column.getQuotedName( dialect ) + " is '" + columnComment + "'" );
{code}
Notice the missing dot.
|
|
