css_parser regex broken for python < 2.7.5 about soupsieve

[facelessuser]

Worse case scenario, I think I can disable selectors that I cannot make work for Python < 2.7.5. I should be able to setup a portable Python 2.7.4.1 environment on my system to at least analyze the issue. If I can confirm, and more importantly resolve, then I'll add a fix. If I can't resolve the issue, I may just disable that selector for Python < 2.7.5.

[facelessuser]

Hmm, that is a pretty old version of Python 2.7. I don't explicitly test versions that old. I probably have no intentions of explicitly supporting anything below 2.7.

If the fix is easy enough, I don't mind adding it for < 2.7.5. I'd have to first identify what the issue is first. It looks like we would be providing a workaround for a bug in the old regex version in regards to * patterns I imagine, but I'm not sure yet.

[facelessuser]

I'll have to see if I can even get access to a python version that old...

[facelessuser]

FYI, this is not related to < 2.7.5, but instead seems to be < 2.7.4. The issue is actually this sub-pattern:

VALUE = r'''(?:"(?:\\.|[^\\"]*)*?"|'(?:\\.|[^\\']*)*?'|{ident}+)'''.format(ident=IDENTIFIER)

It should be as follows to work on older Python 2.7:

VALUE = r'''(?:"(?:\\.|[^\\"]+)*?"|'(?:\\.|[^\\']+)*?'|{ident}+)'''.format(ident=IDENTIFIER)

Subtle. It shouldn't normally break a regular expression engine, but using + in this case does make more sense.