On Thursday 26 August 2004 13:18, Neil Schemenauer wrote:
> I haven't converted the _c_htmltext module to use unicode strings
> but I think pretty much everything else works:
>
Hi,

I've done a simple test:

        uni_val= u"\u00d2scar"  # Ok
        self.add_widget("string", "name", title=uni_val, value=uni_val,
                        size=20, required=1)  # Ok

        uni_val  # Ok
        '
  Uni es %s
' % uni_val  # XXX

log:

  File "/usr/local/lib/python2.3/site-packages/quixote/demo/forms.ptl", line
97, in action
    '
  Uni es %s
' % uni_val
  File "/usr/local/lib/python2.3/site-packages/quixote/_py_htmltext.py", line
85, in __mod__
    return htmltext(self.s % args)
UnicodeEncodeError: 'ascii' codec can't encode character u'\xd2' in position
0: ordinal not in range(128)

The same while tryes to show traceback:

  File "/usr/local/lib/python2.3/site-packages/quixote/publish.py", line 534,
in publish
    request.response.set_body(output)
  File "/usr/local/lib/python2.3/site-packages/quixote/http_response.py", line
219, in set_body
    self.body = unicode(body).encode(self.charset)
UnicodeDecodeError: 'ascii' codec can't decode byte 0xc3 in position 1142:
ordinal not in range(128)

Doing something wrong?

Also, I wonder where to call set_charset.

At first, thought on start_request, but it's not enough for exceptions.

Wouldn't a config var help publisher do it itself?
Ex:
    if request.response.charset is None:
        request.response.set_charset( self.config.charset )

Application users could still select a diferent encoding per session.
( If exceptions are showed, then, config ought to be UTF-8)

Thanks,

-Oscar Rambla