Re: Patch for jdbc ResultSet.getTimestamp()

> I certainly will not be applying any jdbc stuff to 7.1.X.  I was told
> not to apply stuff during beta.  I am sure not going to apply anything
> to a minor release.

OK, but if someone takes responsibility for the accuracy and
effectiveness of a patch, then it should be considered for application.
We had trouble during beta applying larger patches or patches which were
not correct; some of which could be vetted by simple inspection. This
one would actually be in the "simple inspection" category, but if
someone (a third person) wants to test it first that would be better and
that step might be considered essential.

Without any patches, we would not have any minor releases so I'm not
sure that this is violating any standing rules for branch management ;)

Re: the code itself...

I see that this section of code assumes that time zones are on an even
hour boundary (that is, that there are only two digits in the ISO time
zone). That is not true for Canada/Newfoundland and Asia/Calcutta time
zones (perhaps others too), which are on half-hour boundaries. Since the
code is looking for a sign character exactly three positions away from
the end of the timestamp string, it is not likely to catch those cases
istm.

Anyone want to take a stab at handling both 3 ("-08") and 6 ("-02:30")
character time zone fields too? Or am I misreading the code??

                    - Thomas