sre negated lookaheads/behinds

[Andrew Kuchling]

"Fredrik Lundh" <effbot at telia.com> writes:
> you're right, of course; negative lookbehind is clearly broken
> in 2.0.  my fault (shouldn't have let you add it to the docs ;-)

>From my quick glance at the code, though, it doesn't look massively
"oh-God-this-can-never-possibly-work-let's-rip-it-out" broken, just
"still-a-few-glitches-left" broken, which isn't too bad.  (Or do you
think it's massively broken?  Is my proposed fix correct?)

Will you add a test case to re_tests.py once the bug is fixed, or
shall I do that?

--amk