Using more than 7 bit ASCII on windows.

[Paul Moore]

On Mon, 30 Oct 2000 15:08:21 +0100, Paul Moore
<paul.moore at uk.origin-it.com> wrote:
>That sounds very likely - £ in latin-1 is different from £ in the
>DOS-codepage - a fact I had forgotten, which has probably caused all
>of this confusion.

OK, This proves something...

In a console window
-------------------

>>> print unicode("£","latin1").encode("cp437")
£
>>> unicode("£","cp437")
u'\243'
>>> unicode("£","latin1")
u'\234'
>>> print "%o" % ord("£")
234

In PythonWin
------------

>>> print unicode("£","latin1").encode("cp437")
Ã?Å?
>>> unicode("£","cp437")
u'\u252C\372'
>>> unicode("£","latin1")
u'\302\243'
>>> print "%o" % ord("£")
Traceback (innermost last):
  File "<interactive input>", line 1, in ?
TypeError: expected a character, length-2 string found
>>> print map(ord, unicode("£","latin1").encode("cp437"))
[194, 156]

... where the output on the first line displays as
capital-A-with-a-circumflex followed by lowercase-oe. Actually, what I
*typed* was "£" - the cut-and-paste has added the capital-A-hat in the
quotes! (PS, given that it may not make it through Usenet, what is
displayed in my newsreader is capital-A-with-two-dots, comma,
capital-A-with-a-circle, double-straightline-opening-quotes.

And worse, if I type ord("£") (those exact keypresses) I get the
length-2 string error. Which patently isn't what I typed, but is
obviously what Python sees...

Now, the console is comprehensible, but PythonWin is - to put it
politely - odd. Actually, the length-2 string stuff implies that
something is converting to UTF-8 somewhere, somehow...

Is PythonWin trying to use UTF-8 internally? If so, it looks like
you're doing one to many encodes, or some other nasty but easy-to-make
error...

Paul.

PS When I tried to send this message, I got a comment from the
newsreader that it couldn't be sent in Latin-1, because of
non-displayable characters. I've sent latin-1 anyway, as it's more
likely to be at least visible on other screens. But it just goes to
show that what is going on here is pretty nasty to get right...