"Andrew Dunstan" <andrew(at)dunslane(dot)net> writes:
> Tom Lane said:
>> Sure. If we'd been seeing these failures all along, I might think that
>> --no-locale was a good solution, but since they just popped up I think
>> suspicion has to fall on the latest patches.
> That's it alright. When I change the ifndef in varlena.c to #ifndef xxxWIN32
> the regression tests sail through.
OK, so what is your default locale on that machine? In particular, is
it UTF16-based?
Now that I look at it, the strncoll code is utterly obviously broken :-(
regards, tom lane