Re: [PATCH]console:UTF-8 mode compatibility fixes

["Alexander E. Patrakov" <patrakov@ums.usu.ru>]

Adam Tla/lka wrote:
> On Sun, Feb 19, 2006 at 09:24:26AM +0500, Alexander E. Patrakov wrote:
>   
>> Adam Tlałka wrote:
>>
>>     
>>> Maybe I should remember all bytes of the UTF-sequence to use their 
>>> values as a last resort char in case of malformed sequence and 0xfffd
>>> not defined?
>>>       
>> Please don't do that. Display question marks instead in the case when 
>> 0xfffd is not defined.
>>     
>
> Look at the original code. If conv_uni_to_pc fails and there is no replacement
> char (after a clear_unimap for example) and we using US-ASCII we rather
> should see something then sequences of '?' chars.
> Maybe I could change this to:
>
> if (tc == -4) {
> 	if (c < 128)
> 		tc = c;
> 	else
> 		tc = '?';
> }
>
> What about that?
>   
I'd let someone else judge, but that is clearly a broken case that just 
has to be declared broken. <joke>could you please also adapt to a font 
that has all glyphs looking as smileys?</joke> But it's only three extra 
lines of code, so let's accept that "c<128" check.
> Remembering of original bytes is needed if we could then remember
> them in a way so paste from screen gives us the same sequence as it was
> in input.
This doesn't match the behaviour of X.
>  With current console design it is impossible is case
> of correct UTF-8 sequences containing undisplayable glyphs or malformed
> sequences.
I agree that, in some cases, it makes sense to copy and paste 
undisplayable glyphs. However, IMHO, this should not be allowed for 
malformed sequences.

-- 
Alexander E. Patrakov