Reports again

[James Wilde]

On Aug 29, 2011, at 18:56 , Fred Bone wrote:

> On 29 August 2011 at 9:56, James Wilde said:
> 
>> values and not text strings.  The hang-up concerned the translation of the
>> Swedish letters å, ä and ö (and their higher face equivalents, which I
>> didn't test), and this is an issue I must take up with LibreOffice, not
>> GC.  It seems that LibO uses a slightly different coding for these
>> letters.  In GC and the html report file, å is represented as C3A5.  In
>> LibO it is represented as C383C2A5 and comes out as two characters.
>> Similarly with the other Swedish letters.
> 
> This is what happens when you send the data in UTF-8 and the receiving
> system thinks it's in the ISO8859-1 (single-byte-per-character) charset,
> and then proceeds to convert it into Unicode/UTF-8 all over again.

Thanks for that confirmation, Fred.  I assumed it must be something like
that.

> You
> could look at how you transferred the data; there might be an alternative
> which lets you convince LibreOffice that it's being given UTF-8 and not
> ISO8859-1.
> 
I'm somewhat at a loss to see how LibO could translate it as ISO8859-1
when the document it is opening clearly states that it is in UTF-8.  I'm
taking this up with the gang on the LibO list, but so far I've not had
much response.

I think I'll enter it as a bug.

//James