On 5/20/08, Thomas Nielsen <[EMAIL PROTECTED]> wrote: > +1 to Knuts suggestion to get predicatePushdown.sql converted to junit :) > > Cheers, > Thomas > > > Knut Anders Hatlen wrote: > > Narayanan <[EMAIL PROTECTED]> writes: > > > > > > > Narayanan wrote: > > > > > > > Svein Erik Reknes Løvland wrote: > > > > > > > > > Hi, > > > > > > > > > > I want to convert a test from the old harness to JUnit. Any > > > > > suggestions on which test I might work on? > > > > > > > > > > Regards > > > > > Svein Erik > > > > > > > > > > > > > > > > > > > > > > > > http://issues.apache.org/jira/browse/DERBY-2514 > > > > > > > > tagged as newcomer, seems fairly easy for starters, > > > > > > > > You might also want to look into closely at the comments by Myrna, > > > > > > > > Narayanan > > > > > > > You can also take a look at > http://issues.apache.org/jira/browse/DERBY-1903 > > > > > > This is a bigger one than 2514, I guess if you want to look at a > > > simple one first 2514 is better. > > > > > > > Hi Svein Erik, > > > > Thanks for volunteering to do this! If you feel that you need more > > things to do, I'm sure everyone would be happy to see > > lang/predicatePushdown.sql converted to JUnit. I don't think that the > > conversion should be too difficult, but it's a fairly large SQL script > > so it may take some time. This test causes some noise in the nightly > > regression tests (DERBY-1720, DERBY-1902) because it sometimes prints > > different runtime statistics and therefore fails, even though the > > differences are not relevant to what is tested. In JUnit, it would be > > easier to check only the relevant parts of the statistics, and we could > > hopefully reduce the noise from that test. > > > > > > -- > Thomas Nielsen > lang/subquery.sql is another one that fails (in)frequently, apparently sometimes the optimizer picks a different path...The automated signature matching often attributes these failures to DERBY-1902 also.
Myrna
