Re: UTF8 encoding problem

On 18/giu/08, at 15:00, Michael Fuhr wrote:

> On Wed, Jun 18, 2008 at 08:25:07AM +0200, Giorgio Valoti wrote:
>> On 18/giu/08, at 03:04, Michael Fuhr wrote:
>>> Is the data UTF-8?  If the error is 'invalid byte sequence for
>>> encoding "UTF8": 0xa3' then you probably need to set client_encoding
>>> to latin1, latin9, or win1252.
>>
>> Why?
>
> UTF-8 has rules about what byte values can occur in sequence;
> violations of those rules mean that the data isn't valid UTF-8.
> This particular error says that the database received a byte with
> the value 0xa3 (163) in a sequence of bytes that wasn't valid UTF-8.
>
> The UTF-8 byte sequence for the pound sign (£) is 0xc2 0xa3.  If
> Garry got this error (I don't know if he did; I was asking) then
> the byte 0xa3 must have appeared in some other sequence that wasn't
> valid UTF-8.  The usual reason for that is that the data is in some
> encoding other than UTF-8.
>
> Common encodings for Western European languages are Latin-1
> (ISO-8859-1), Latin-9 (ISO-8859-15), and Windows-1252.  All three
> of these encodings use a lone 0xa3 to represent the pound sign.  If
> the data has a pound sign as 0xa3 and the database complains that
> it isn't part of a valid UTF-8 sequence then the data is likely to
> be in one of these other encodings.

Much clearer now, thank you Michael.

--
Giorgio Valoti