"Phillip Smith" <phillip.smith@weatherbeeta.com.au> writes:
> How has the software been misinformed and how can we avoid it then? :)
Well, if that were entirely clear then we could fix it.
> The start of the pgdump sets the client encoding, so shouldn't that tell PG
> that things could get strange?
The problems seem to arise when the data doesn't actually have the
claimed encoding. We've been gradually tightening the encoding checks
over the years, so data that older versions let pass without comment
might now be recognized as malformed. In the OP's case there may be
something else going on (such as an actual encoding-conversion bug).
I've encouraged him to try to trace things forward from his original
input data and see if we can find out exactly where it went off the
rails.
regards, tom lane