Source code encoding is ignored in IDLE console
sessions currently.  To allow to use encodings other
than Latin-1, encoding marker must be added before
passing it to tokenizers. (see the attached patch)

current behavior:

>>> u'\ '
u'\xc7\xd1\xb1\xdb'

this must be like ...

>>> u'\ '
u'\ud55c\uae00'

Logged In: YES 
user_id=21627

I think the patch is wrong, as it messes with the source
code typed. At a minimum, line number will be off by one, no?

I'm currently working on a different solution, allowing
compile() to accept an encoding. However, it is too late to
add this to 2.4 now.

Logged In: YES 
user_id=55188

Surprisingly, it doesn't affect source code lines.  But, I'm
still wondering why it isn't counted as a distinct line. :)

>>> raise ValueError
Traceback (most recent call last):
  File "<pyshell#3>", line 1, in -toplevel-
    raise ValueError
ValueError

>>> if 1:
	pass
	raise ValueError
Traceback (most recent call last):
  File "<pyshell#9>", line 3, in -toplevel-
    raise ValueError
ValueError

Okay. I agree that it's too late to change behavior, but it
may need to write some errata item about this inconsistency
of IDLE vs terminal interactive mode on documentation.

Can this be closed? Is it still an issue in 3.0?

No response from OP, closing.