Re: pg_do_encoding_conversion glitch

ITAGAKI Takahiro <itagaki(dot)takahiro(at)oss(dot)ntt(dot)co(dot)jp> writes:
> I have a question about the result contract of pg_do_encoding_conversion().

It's poorly designed :-(

As near as I can tell, all uses of the function either pass a
null-terminated string or special-case the result == src case in such a
way that it doesn't matter. However it's certainly not obvious that you
have to do that.

The calls in contrib/sslinfo might be broken --- not sure how much
that module has been tested.

Do you have a proposal for a different API, or do you just want to
improve the comment for the function? Bear in mind that a lot of the
callers do know the string length, and so we shouldn't impose an
unnecessary strlen() operation on those cases.

regards, tom lane