Python: Regexp: Handle repetions {n} (with no ,)

[yoff]

To solve FP report #2403 with internal issue https://github.com/github/codeql-python-team/issues/85.
The code currently only handles the syntax {n,m}. This adds support for {n}.

[yoff]

As Taus mentioned, we might consider checking if there is a dual problem with caret and if solving that should then be part of this PR.

[yoff]

I am not sure why 4-5 is in there, but both that and removing 2-3 is consistent with how {n,m} is handled..

[tausbn]

I was perusing the Python re module documentation just now (as you do), and I noticed the \A and \Z escapes, which match the start and end of the string respectively. Thus, something like

>>> re.match("(\Aab$|\Aba$)$\Z", "ab")
<_sre.SRE_Match object; span=(0, 2), match='ab'>

works as expected. I don't know that we're explicitly handling these cases here, so perhaps a few more tests would be in order? 🙂

[yoff]

Looks like lastPart has an explicit case for \$ and would need one for \Z.

[yoff]

Made the test @tausbn sugested pass, the code feels a bit ad-hock, though.

[RasmusWL]

Besides autoformatting and a maybe adding a test-case, it all looks good to me 👍

[yoff]

I added https://github.com/github/codeql-python-team/issues/114 to track the fact that, as @tausbn observed, we probably have the dual problem with carets.

[RasmusWL]

Still LGTM 🚀