On 05/18/2011 03:46 PM, Thierry Hauchard wrote:
> Hi,
>
> We experienced similar problem : it was just the encoding of pg_hba.conf
> that was accidentally changed when modifying it (UTF8/ansi)
The encoding should not matter unless there are characters above byte
128, because UTF-8 and the ANSI encoding are the same from 0-127.
I guess it's possible that Pg chokes if it finds a byte-order mark (BOM)
at the start of pg_hba.conf; I haven't tested this. A BOM is optional
for UTF-8 text but some text editors do add it, and I've run into
software that chokes on a BOM before. Many editors do assume unknown
text is UTF-8 if they can't identify it as some other known encoding,
and some would add a BOM on saving in this case.
If there is a byte order mark (BOM) you'll be able to tell because the
first three bytes of pg_hba.conf or postgresql.conf , when examined in a
hex editor, will be 0xEF 0xBB 0xBF .
--
Craig Ringer