[pygtk] utf8 validating string

[Yann Leboulanger]

John Ehresman wrote:
> Yann Leboulanger wrote:
>> I'd like not to have it. But I getthis string by gpg-decodding a message
>> send by Miranda IM. I think it's a bug in their GnuPG implementation,
>> but anyway I'd like my client to detect those bad string and a) print
>> message correctly if I can or b) don't traceback and print a warning
>> message. But for that I need a function that tells me that
>> g_utf8_validate will fail ...
> 
> You probably should explicitly decide how to handle \0.  If it's always
> at the end, it's probably just a simple bug and can be chopped off but
> it may be something more if valid text follows the \0.
> 
> But in general, I think this'll work:
> 
> def valid_glib_utf8(s):
>   try:
>     unicode(s, 'utf-8')
>   except Exception:
>     return False
>   else:
>     return '\x0' not in s
> 
> In case you need it s.replace('\x0', '') will remove the \0's.
> 
> Cheers,
> 
> John
> 

That doesn't work:
>>> import gtk
>>> tv = gtk.TextView()
>>> b = tv.get_buffer()
>>> t = "test\x00"
>>> u = unicode(t, 'utf-8')
>>> b.set_text(t)
__main__:1: GtkWarning: gtk_text_buffer_emit_insert: assertion
`g_utf8_validate (text, len, NULL)' failed

it's the same if I try with the unicode:
>>> import gtk
>>> tv = gtk.TextView()
>>> b = tv.get_buffer()
>>> t = "test\x00"
>>> u = unicode(t, 'utf-8')
>>> b.set_text(u)
__main__:1: GtkWarning: gtk_text_buffer_emit_insert: assertion
`g_utf8_validate (text, len, NULL)' failed

-- 
Yann