This replaces the two custom Google authentication backends originally
written in 2012 with using the shared python-social-auth codebase that
we already use for the GitHub authentication backend. These are:
* GoogleMobileOauth2Backend, the ancient code path for mobile
authentication last used by the EOL original Zulip Android app.
* The `finish_google_oauth2` code path in zerver/views/auth.py, which
was the webapp (and modern mobile app) Google authentication code
path.
This change doesn't fix any known bugs; its main benefit is that we
get to remove hundreds of lines of security-sensitive semi-duplicated
code, replacing it with a widely trusted, high quality third-party
library.
If a url doesn't have a scheme, browsers would treat it as a relative
url and open something like: https://chat.zulip.org/google.com instead.
This PR fixes the issue on the backend; the frontend implementation
remains out of sync and the user sending the message wouldn't see
any linkification for urls without a scheme.
Fixes#12791.
* backports-abc: For old Python versions.
* backports.ssl-match-hostname: For old Python versions.
* docopt: Has never been used directly.
* gitdb: Has never been used directly.
* ndg-httpsclient: No longer used by requests ≥ 2.12.1.
* pycrypto: Has never been used directly.
* smmap: Has never been used directly.
* typing: For old Python versions.
* typing_extensions: For old Python versions.
PROVISION_VERSION is not bumped because these were already unused
since at least the last major bump.
Signed-off-by: Anders Kaseorg <anders@zulipchat.com>
Reverts c09962b and 697b4b2. The lines that break look like
stripe.Invoice.finalize_invoice(stripe_invoice).
Something like stripe_invoice.finalize_invoice() would work, but it's a big
change given how the tests in test_stripe.py currently work.
This is required for Python 3.7 support; line_profiler needs a release
with Cython rerun following the Python 3.7 release, and this achieves
that goal.
This commit leverages the ahocorasick algorithm to build a set of user_ids
that have their alert_words present in the message. It runs in linear time
of the order of length of the input message as opposed to number of
alert_words. This is after building a ahocorasick Automaton which runs
in O(number of alert_words in entire realm) which is usually cached.