Bruce Badger <bruce_badger@BadgerSE.com> writes:
> The description of AsciiRow says that for each field it contains "the
> value of the field itself in ASCII characters". The description is here:
> http://postgresql.org/users-lounge/docs/7.3/postgres/protocol-message-formats.html
> Looking at the bytes comming over the Frontend/Backend protocol,
> however, it seems that the field is really always a String encoded
> according to the settings in the database - is that right?
I believe this documentation's reference to ASCII shouldn't be taken
very literally; it really just means "text" (as opposed to some internal
binary representation) and is not meant to imply anything about one
character set encoding versus another. You're correct that the data
would actually appear in the currently-selected client encoding,
assuming that the backend is configured to support encoding conversions.
regards, tom lane