Review request for 7145913 CachedRowSetSwriter.insertNewRow() throws SQLException

Ulf Zibis Ulf.Zibis at gmx.de
Wed Jun 13 22:03:38 UTC 2012


Am 13.06.2012 13:04, schrieb Lance Andersen - Oracle:
> Hi Paul,
>
> Thank you for taking  the time to review the code.
>
>
> I made the change you suggested below
>
> http://cr.openjdk.java.net/~lancea/7145913/webrev.02
Typos:
  912                 * Hence we cannot exactly identify why the insertion failed
  913                 * Present the current row as a null row to the user.
Please add a ',', use lowercase 'p' and maybe rename user to caller
  912                 * Hence we cannot exactly identify why the insertion failed,
  913                 * present the current row as a null row to the caller.

In isPKNameValid(...) you could reuse icolCount from line 843.

Another additional thought:
Isn't it likely, that the internal imlementation of ResultSetMetaData methods decrements the column 
index like following:
     public String getColumnClassName(int column) {
         return columns[column - 1].className;
In this case, iterating over the range of 0..cols-1 could be a small performance gain after JIT 
optimization:
1469         for(int i = 0; i< cols; i++) {
1470             String colName = rsmd.getColumnClassName(i + 1);
If not, the offset could be anyway treated by single opcode using complex addressing.

Same for iterating over ResultSet collumns.

-Ulf




More information about the core-libs-dev mailing list