On 26.07.19 14:59, Chapman Flack wrote:
> pgJDBC's toJavaSecs() and toPgSecs() methods contain some dubious
> adjustments, apparently there in an attempt to make the java.sql
> mapping less weird. But it really succeeds only in smearing the weird
> around so it's not where it belongs, and some even gets on the java.time
> mapping, which ought to be free of weird.
So, I had a look at this. I can only reproduce it with the binary
protocol, not with the text protocol. So as a short term work around
switching to the text protocol should work.
java.time inherits the weird because it also goes through toJavaSecs. I
am working on a patch but it needs more tests and polish. I hope I'll be
able to submit a bug and PR shortly.
Cheers
Philippe
