Subj : codepage
To   : Rob Swindell
From : Michiel van der Vlist
Date : Sun Mar 05 2023 02:01 pm

Hello Rob,

On Saturday March 04 2023 13:11, you wrote to me:

 >> So let me get this straight:
 >>
 >> 1) If the message that is responded to, is encoded in CP437,
 >> Synchronet answers in CP437. Yes?

 RS> No. The message response itself determines the encoding and only CP437
 RS> terminals can faithfully author CP437 encoded messages. If a UTF-8
 RS> terminal user responds to a CP437 encoded message (with non-ASCII
 RS> chars), the original message text is converted to UTF-8 before it is
 RS> quoted and the response will be UTF-8. Unless there are no non-ASCII
 RS> chars in the response, in which case the response charset witll just
 RS> be ASCII.

I see... So it is the terminal - or whatever functions as its equivalent - and only the terminal that determines the encoding of the message at hand.

 >> So what happens if the response does not fit into CP437?

 RS> I think this question is making false assumptions.

It is making assumtions, but they are not false I would say. Read on.. I will come back to that further down.

 >> What happens if the original message is encoded in a one byte
 >> encoding other than CP437?

 RS> The only encodings Synchronet supports for message text are ASCII,
 RS> CP437, and UTF-8.

Hmmm... That leaves out a big part of Fidonet. These days the majority, maybe the vast majority is writen in a language that uses the Cyrillic alfabet and the encoding is CP866.

 >> 2) If the message that is responded to is encoded in UTF-8,
 >> Synchronet answers in UTF-8 if the terminal that is used supports
 >> UTF-8. Yes?

 RS> Yes.

OK, so far so good...

 >> So what happens in that case if the terminal does not support
 >> UTF-8?

 RS> The message text would be converted to CP437 before being quoted and
 RS> the response would be in CP437.

And now I come back to my previous question: what happens if it does not fit into CP437? That can easely happen. A Euro sign '€' can be composed in UTF-8 but it  does not fit into CP437.

 >> My software translates the CP437 encoded degree sign into UTF-8 as
 >> you can see.

 RS> Yup, most software does the same, when appropriate.

My Golded does, but the reverse is a bit problematic.


Cheers, Michiel

--- GoldED+/W32-MSVC 1.1.5-b20170303
 * Origin: Nieuw Schnøørd (2:280/5555)

.