Re: UTF-8 editing

On 2 Oct, spam2011m@... typed:

> In message <0861d8f654.boase@...>
>  on 21 Aug 2015 Bernard Boase  wrote:

>> Have I missed a vital extra setting, or is UTF-8 support in MPro's
>> Editor precluded by that lack of UTF-8 support in the Wimp?

> I don't have any experience with a recent enough version of EmailEdit,
> but it's not a problem fundamentally to do with the Wimp.  There is
> nothing stopping a custom editor like EmailEdit displaying Unicode
> correctly.  You might have trouble typing characters if your alphabet
> is still set to Latin 1, but you'd probably have trouble typing the
> Unicode characters anyway.

> You didn't exactly say how it couldn't cope.  Does it display
> gibberish in the editor window if you quote the message when replying?
> Or is it displaying OK but cannot be edited?

My EmailEdit is version 2.02 (14-Feb-2015).

The incoming email had the Greek word 'laoutzikos' shown correctly in 
Greek characters. If I reply quoting the message, the word is 
displayed in EmailEdit as a sequence of 10 character pairs with the 
first of each pair being a capital I circumflex (&CE) or I diaresis 
(&CF) as you'd probably expect. Thus the initial lamda is &CEBB.

Key F8 to put the email into Edit and the lamda is shown as =CE=BB, 
presumably MPro's internal way of representing UTF-8.

In the same email there was a pair of sexed double quotes which 
displayed correctly but, in EmailEdit's reply, became the searate 
characters &E2 &80 &9C and &E2 &80 &9D, plausibly correct UTF-8 when 
strung together.

I don't wish to add any non-ASCII characters (yet!) but would like the 
quoted message to remain in UTF-8.

Rob Sprowson, speaking at ROUGOL last month, indicated that support 
for UTF-8 already exists in RISC OS 5 but there is work to do in the 
printing system and some applications such as Edit and the text areas 
in Draw. He added that setting *Alphabet UTF8 already enables anything 
using the Wimp for text input to work correctly.

So I tried that, and then Reply to sender rendered the UTF-8 
characters correctly, with an exception: tau (&CF84) produced a black 
lozenge question mark (font DejaVu's 'unknown'). Is this because &84 
(among others) is undefined in Acorns's Latin1 ISO 8859/1?

And, as expected, I can't actually edit in Unicode until a version of 
Chars or XChars appears from which to pick Unicode characters.

There is a corresponding issue with UTF-8 in headers, but that is 
addressed in a different thread here about the Subject line.

-- 
Bernard

______________________________________________________________________
This message was sent via the messenger-l mailing list
To unsubscribe, mail messenger-l+unsubscribe@...