Erick the test problems havent gone away (see latest hudson failure), but as you can see they appear to be unrelated to your patch, but existing problems.
I am glad Uwe and Hudson are now seeing the failures I saw, but unfortunately I am unable to reproduce this problem. I think it is very wierd that it only recently has started happening and the Junit 4 jar file is looking very suspicious to me
The next step is to force another test failure (with existing trunk code), except with an additional assertion in TestQueryParser.getParser(), that asserts that the created queryparser has the same Locale as LocalizedTestcase.locale
If this assertion is not triggered, and a failure happens anyway, then there is no problem with LocalizedTestCase and we should commit your patch. This would mean the problem is instead some problem in the query parser, its tests, date calculations, jvm handling of thai dates, something else, but not the locale-switching.
if the assertion is triggered, then we need to figure how why before changing the locale-swapping mechanism.
I ran this test 2500 times last night but i couldnt make it fail. i have a very small portable right now and I will need to wait till the new year to really look at this more...