Christian Boos wrote:
> Right, look at time.daylight usage in the code. Tim's buildbot must live 
> in a non-DST timezone.

Correct, I could finally reproduce the issue by switching my timezone to
PST. Actually, any timezone different from the one used in that
particular test (which happens to be mine) reproduces the issue.

> I suppose the test has to take this into account, for choosing an 
> appropriate expected value.

No, actually some of the changes in [8831] were wrong. The issue is
fixed with [8857]. I guess we now have enough reasons to make a second
release candidate...

-- Remy

Attachment: signature.asc
Description: OpenPGP digital signature

Reply via email to