Re: Unit test patches

Kris Jurka wrote:
>
>
> On Fri, 1 May 2009, John Lister wrote:
>
>> The first occurs in DatabaseMetaDataTest.testTables line 104:
>>
>> It fetches all the columns for any table beginning with test - i
>> happen to have other tables named test... which causes this to fail.
>> I've modified it to look for the testmetadata table, alternatively
>> the docs should be altered to state that the test database should be
>> clean.
>
> It is expected that the test database is empty.
OK, maybe we should alter the docs, it only mentions that tables are
created/dropped and shouldn't contain production tables, but nothing
about it being empty. I'd forgotten i had some other tables in there...
>
>> The second in TimeTest.timeTest : line 269 & 283:
>> The timezone tests fail because the as it happens local daylight
>> savings is in effect, the code tries to test for this by submitting
>> the time being tested, but this actually works out daylight savings
>> for the time at the epoch (1/1/1970). I've modified it to use
>> DST_OFFSET instead (note -  this may fail if ran during the
>> switchover period).
>
> What timezone are you running the tests in?  I don't see them in
> US/Pacific.  Reproducing the failure would help.
>
GMT with DST in effect.

JOHN