pytest migration: api: switch to standard assert statements

Use unittest2pytest to replace unittest-style assert statements (e.g.
assertEqual) with standard Python assert statements to benefit from pytest's
improved reporting on assert failures.

The conversion by unittest2pytest was correct.

pytest migration: models: switch to standard assert statements

Use unittest2pytest to replace unittest-style assert statements (e.g.
assertEqual) with standard Python assert statements to benefit from pytest's
improved reporting on assert failures.

The conversion by unittest2pytest was correct except for:
- 'import pytest' is not automatically added when needed
- line wrapping in string formatting caused a syntax error in the
transformed code. Reported upstream at
https://github.com/pytest-dev/unittest2pytest/issues/3 .
- in assertRaises with a lambda, the lambda needs to be removed

pytest migration: other: switch to standard assert statements

Use unittest2pytest to replace unittest-style assert statements (e.g.
assertEqual) with standard Python assert statements to benefit from pytest's
improved reporting on assert failures.

The conversion by unittest2pytest was correct except for:
- 'import pytest' is not automatically added when needed
- line wrapping in string formatting caused a syntax error in the
transformed code. Reported upstream at
https://github.com/pytest-dev/unittest2pytest/issues/3 .

pytest migration: add args/kwargs to assertRaises replacement

Even though we'll soon get rid of the assertRaises replacement, the tests in
kallithea/tests/other/test_validators.py expect to pass it args and kwargs
so we need to adapt it first.

tests: users: enable commented UsersController test

Previously, a test was written but could not be enabled until pytest
was fully available. Now that it's the case, enable the test.

tests: api: fix intertest dependency on repository locking

In test classes based on unittest, tests are executed in alphabetical order.
In test classes based on pytest, tests are executed in the order they are
specified. This difference revealed a problem in the API tests:

- test_api_lock_repo_lock_optional_locked locks the test repository
- test_api_get_locks_regular_user gets the current locks and expects it to
be empty

With unittest as base class, this worked fine because the 'get_locks' group
of tests are executed before the 'lock_repo' group (alphabetical order).
Using a real pytest-based test class, the order is swapped and the locked
repository from the first test invalidates the preconditions of the second
test.

Fix this specific problem by releasing the lock from
test_api_lock_repo_lock_optional_locked.

This commit does not fix other interdependencies between tests. For example,
test_api_lock_repo_lock_optional_locked expects the existing lock state to
be 'locked' but did not lock the repo itself; instead it expects a previous
test to have locked. In practice, this is
test_api_lock_repo_lock_aquire_optional_userid.
A full solution would make each test fully self contained so that tests can
be executed in random order. The pytest extension pytest-random can help
detecting these problems.

tests: add set_test_settings fixture

Uses https://pytest.org/latest/yieldfixture.html#yieldfixture to restore
settings after tests. Takes into account current specifics of list-valued
settings.

tests: add create_test_user fixture

Uses https://pytest.org/latest/yieldfixture.html#yieldfixture to provide
auto-deleting users to tests.

subprocess: output on stderr doesn't mean that a process failed

Git sometimes does that. Threading non-determinism did that we sometimes didn't
see it, and sometimes we failed.

templates: be more strict about mako errors

We prefer to find errors early.

pullrequests: reword the grumpy "Rejected" to the more neutral and factual "Not approved"

"Rejecting" could be considered rude and some reviewers were thus reluctant to
use it. Rewording it might give reviewers a voice to better express their
feelings.

docs: tweak email documentation - add gmail example

smtp_auth is too obscure to have in the .ini template.

pull requests: refactor update_reviewers

This avoids a redundant database lookup for every removed reviewer,
and shows usernames (not just IDs) in associated log messages.

pull requests: prevent adding DEFAULT user as reviewer

Add a helper method to resolve reviewers, with an added check to prevent
the adding of the DEFAULT user as reviewer.

The __add_reviewers method, although internal, had a troubling interface
where the method was responsible for resolving reviewers, but the caller
was responsible for resolving mention_recipients. The method is changed
to always take two sets of database User objects, leaving the resolution
to the caller in both cases.

utils: improve extract_mentioned_users usefulness

Previously, extract_mentioned_users performed a regex search, converted
to a set to remove duplicates, converted that set back into a list, and
sorted the list on username, ignoring case. Every single caller of the
function then promptly took the carefully pruned and sorted result and
put it back into a set, rendering the entire exercise pointless.

In addition, every caller also resolved the usernames to database User
objects.

In this changeset, extract_mentioned_users is changed to return a set
of database User objects directly. A new extract_mentioned_usernames
function, which does only username extraction, is kept for the sole
purpose of providing a more testable interface (no database needed).

Bonus feature: The new extract_mentioned_users will prune non-existent
users, as well as the DEFAULT user. This means it is no longer possible
to @mention (and send notifications to) the DEFAULT user.

auth: prevent misuse of PermFunction in bool context

Evaluating a PermFunction as a boolean, rather than calling it, is
almost certainly an error. If not, "pf is not None" can be used.

notifications: untangle notification access check

This removes a broken permission check when viewing notifications (the
HasRepoPermissionAny object was created, but never actually called with
a repo_name argument as required). It would be non-trivial to actually
implement the check, as notifications don't track their repository
relationship explicitly, and even then, it's unclear why it would
make sense to allow a repository admin to see notifications to
other users.

It was never a vulnerability, due to a subsequent (and much stricter)
ownership check, which remains but has been untangled for readability.
In short, this changeset is a pure refactoring, except that specifying
a non-existent notification ID will now produce error 404, not 403.

setup.py: enable pytest-sugar plugin

pytest-sugar enables a nice visual representation of test progress and
results.
Minimum version is set to 0.7.0 as it fixes a problem where skipped tests
would not be shown in the test summary.
The upper version 1.0.0 is arbitrarily chosen and could be increased as seen
fit.

helpers: don't truncate words when when truncating texts for urlification

This would especially cause confusion when URLs were truncated.