[Python-Dev] [patch #100912] should we keep the \xnnnn escape in unicode strings?

[M.-A. Lemburg]

Fredrik Lundh wrote:
> 
> mal wrote:
> 
> > There were objections from Finn Bock and myself: \xXXXX is
> > defined to mean "read all hex chars until the next non-hex char
> > and then cast to the underlying type (char or wchar_t)" in C9X.
> 
> finn wrote:
> 
>     "the implementation of \x in JPython is so full of bugs
>     and inconsistencies that I'm +1 on your proposal."
> 
> since when is +1 an objection? ;-)

Ooops, I must have overlooked that one.
 
> > Not that this definition is optimal, but we should stick to what
> > the standard says and only depreciate usage of \xXXXX in favour
> > of \uXXXX. Code using escapes like "\xABCD" which then results
> > in "\xCD" is broken anyway -- having u"\xABCD" return "\uABCD"
> > wouldn't make much of a difference (+ the bug would become pretty
> > obvious if viewed in a Unicode aware viewer: Asian characters are
> > very easy to recognize in ASCII text ;-)
> 
> as Tim pointed out in the SRE thread, the only standard that
> means something here is Python 1.5.2.  And in Python 1.5.2,
> \u means read as many hex digits as you can, and cast the
> result to a character.
> 
> I'd say it's more important to be compatible with Python (1.5.2
> string literals, SRE regular expression literals), than to be com-
> patible with C9X...

Ok, ok, go ahead :-) (please also update the docs, if there
are any on this subject -- a grep didn't show anything up on \xXXX)

-- 
Marc-Andre Lemburg
______________________________________________________________________
Business:                                      http://www.lemburg.com/
Python Pages:                           http://www.lemburg.com/python/