i18n: updated translation for German

Currently translated at 46.2% (516 of 1116 strings)

notifications: use stupid mail static-except-[] subjects to please gmail and its broken threading

This gives reasonable threading, both with gmail and proper mail clients, at
the cost of making the email subjects slightly obscure.

Gmail is clueless - also about mail threading, ignoring our References headers
and the wisdom of jwz. Gmail will start a new thread for each subject. The only
way to make all mails related to the same PR belong to the same thread is thus
to let all mails for a PR have the same subject. Gmail will however ignore
content in square brackets at the beginning of the subject so we can put the
interesting parts there.

db: more consistently set primary keys to be unique, not nullable and without default

Setting primary_key will (in general) automatically set autoincrement and
setting a default is just odd.

comments: show status changes as a part of the comment instead of in the header

This makes it obvious what the user said and there is no reason to explain what
he is saying or where he is saying it.

Comments will always have a text or a status change - there is no need for "No
comments".

comment: replace floating status with span

Simpler ...

Similar clean-ups could be done in a lot of places - this is just something I
"needed" for the next change.

pullrequests: use correct References header when adding reviewers to (new?) PRs

Use the same message id when inviting to a PR as we do when commenting on a PR.

reposcan: avoid infinite recursion - don't follow symlinks when scanning

The whole reposcan mechanism is a mess and should be rewritten ...

changeset: hack to "reload" page to make #target links work first time

Like 9dfd9d242728 did for pull requests.

files: fix hiding of old content while reloading

Fixing typo from 43cafe348c9a.

i18n: added translation for Belarusian

Currently translated at 80.8% (902 of 1115 strings)

diff view: improve appearance of line numbers

* display the ellipsis just once and centered
instead of showing in twice in both - and + columns
* expand column width as needed to accomodate longer numbers
(previously, the column width was fixed)
* enable box-sizing: border-box for the number links to
make it easier to set paddings

comments: remove line breaks after user mentions in comment previews

Previously, the preview might look like this:

@username

, have you seen it?

Now it should look like this:

@username, have you seen it?

rst: in @mention parser, escape spaces so they don't go to HTML

This eliminates extra spaces around @mentions. Every time mention was followed
by a comma, for example:

@username, have you seen it?

it turned into:

@username , have you seen it?

So an extra space was inserted. It was inserted because otherwise rst parser
might not recognise the markup (i.e. @user1,@user2 is replaced by
**user1**,**user2** — that would be interpreted as <b>user1**,**user2</b>).

See http://docutils.sf.net/docs/ref/rst/restructuredtext.html#character-level-inline-markup

hg: show code function in diffs

Just like already done for git diffs, show the function containing the
change in the hunk header of a diff, if available:

@@ -19,7 +19,7 @@ int func(void)

instead of

@@ -19,7 +19,7 @@

pullrequests: add PullRequest.nice_id method

Currently, a pull request id is referenced as #5, and the '#' symbol is
fixed and repeated in several places. This commit adds a class method
make_nice_id, that returns a string reference to a pull request,
currently in the form '#5'.

This function could be overridden by an organization if they need references
in another form, for example PR-5.

The method nice_id uses this for the pullrequest's own id.

diff view: do not hide overflow on very long lines that cannot be wrapped

When a change contains very long lines that cannot be wrapped because they
contain very long 'words' without spaces, the overflow is currently not
visible at all. This occurrence of such very long non-wrappable words is
pretty uncommon, but can for example happen in comma-separated value files.

Not touching the existing behavior of trying to wrap long lines, this patch
adds horizontal scrollbars for these exceptional cases of unavoidable
overflow.
On normal lines and long lines that can be wrapped, the diff block remains
the same as before.

ChangesetCommentsModel: refactor duplicate code

Comments are now always ordered by creation date ... but that should not make
any difference.

ChangesetCommentsModel: cleanup of doc strings

Docstrings that are out of date is a bug. Docstrings that doesn't add any value
is thus a liability we would like to get rid of.

[Changeset extracted and modified by Mads Kiilerich.]

db: remove redundant Column namings missed in dbbddd1e45cf

sed -i "s,^ \(.*\) = Column(\"\1\"\, , \1 = Column(,g" kallithea/model/db.py

rcextensions: cleanup of code and documentation

More cleanup is needed - this is just to clarify what it currently does.

permissions: by default, don't show entities with no access

There might be a lot of entities the user/group don't have access and it is
much more important to be able to see at a glance what the user/group has to
access to than what it doesn't have access to.

templates: drop autocomplete="off" - it is ignored by browsers

This is annoying for admin when setting passwor for other users - we will have
to do that in some other way.

templates: only bind pager event handlers once - avoid "leaks" and repeated page loads

Summary page ended up hammering the server and when pages were loaded one time
extra for each paging click.

Just moving javascript code from the data template (where it would be re-run on
each paging) and to the main page template.

changelog_summary: show changeset status label rather than raw status

Instead of showing the changeset status as e.g. under_review, use 'Under
Review', like all other similar code.

changelog: use status label directly from c.statuses.get() output

In the changelog template, c.statuses.get() returns a list:
[ str(status), status_lbl, pr_id, pr_repo ]

So to obtain the status label, one can directly use the second element,
rather than using a helper function on the first.

comments: avoid storing 'No comments' text when changing status

When a general comment (with or without status change) is added to a
changeset or pull request, and no text was added, Kallithea automatically
used 'No comments' as text. The stub text is added to the database as if it has
been entered by the user and it can thus not easily be identified as an
automatic comment.

This commit makes following changes:
- allow adding an empty comment to the database when there is a status
change. An empty comment without status change is ignored.
- do not add a stub text to the database, but generate it on demand
- the stub text is shown in italic font to differentiate it from user-entered
text

Currently there is a large amount of duplication between
controllers/changeset.py and controllers/pullrequests.py, which is to be
cleaned up in a later commit.

tests: update for git sometimes emitting 'Cloning into' on stderr

Tests failed with Git 2.1.0 .

tests: rename test_vcs_operations.py to avoid automatic pick-up

This is a special test that requires special setup and must be run manually.

Nose already skips the test because it is executable.

tests: restore pytest norecursedirs default patterns

This replaces d741b99effdf by adressing the root cause: 5c8c77b7ca56 removed
the default exclude of .* (and thus also .eggs/ used by latest setuptools) and
*.egg (used by older setuptools).

tests: Move pytest config (norecursedirs) from tox.ini to setup.cfg

Avoid (conceptual) dependency on the tox meta test runner configuration.

tox.ini: Add .eggs to norecursedirs

This prevents py.test from trying to run tests in the .eggs directory, which
setuptools can create.

tests: avoid executing tests in base classes

Fixes lots of failures with pytest by making sure base test classes don't
inherit from unittest.TestCase or by prefacing name with an underscore so tests
don't get executed by pytest.

tests: add pytest section and put tox and tests/scripts to norecursedirs

kallithea/tests/scripts are non-integrated messy tests
that cant be part of a testsuite as is

pullrequest overview: fix 'my pullrequests' overview

Commit bd4f453a0055 fixed the pager links on pullrequest overviews, but
broke the 'my pull requests' page due to insufficient testing and
incomplete Python understanding.

pullrequest overview: fix 'my pullrequests' overview

Commit bd4f453a0055 fixed the pager links on pullrequest overviews, but
broke the 'my pull requests' page due to insufficient testing and
incomplete Python understanding.

changeset status: fix common case where a reviewer did not yet review anything

When a reviewer did not yet make any review, its status for the changeset is
None, causing the code to crash. The unit tests also did not cover this case.

pullrequest overview: fix missing pager links

Commit 0f9a48e0adc3 refactored the pullrequest overview pages, but has a
bug causing no more than 10 pull requests to be shown, without pager links.
The code assumed that 'pager' was a variable, rather than a method, of the
pullrequests structure.

fonts: update symbol codes to be closer to real Unicode

Let's use character codes closer to what Unicode defines, so if the
custom font can't be used for some reason what user sees is still
more or less readable.

comments: display username and name instead of only username

The full name is more significant to many people than a username.

changeset status: set status to rejected if at least one reviewer rejected

Currently, a reject of a change by a reviewer does not affect the overall
status at all. A logical policy is to reject the change if at least one
reviewer rejects it.

Note: different repositories/organizations may require different policies
for the overall status. Currently, the policies for approve/reject are
fixed. A possible improvement is to provide several alternative
policies and allow selecting the policy per instance or per repo.

changeset_status: add unit tests for calculation of overall status

Add unit tests for the calculation of the review status of a changeset/pull
request. To allow this, some reorganization of the ChangesetStatusModel is
needed.

db: remove redundant Column namings

sed -i "s,^ \(.*\) = Column('\1'\, , \1 = Column(,g" kallithea/model/db.py

db: fix gist_owner database field name - it is user_id

Fix bug introduced in 3533ee5cbb62.

error: fix handling of errors without response

Broken by 6f9dae4d39d9.

tests: set app_globals and translater in conftest.py for py.test

Copy over code from the pylons nose plugin for setting app_globals and for
setting a translator object.

This allows more of the functional tests to pass; particularly the ones related
to i18n.

tests: Destroy user_user_group 'perms_group_1'

This doesn't seem to be a problem when running `nosetests`, but with `py.test` for some reason it becomes a problem:

$ py.test kallithea/tests/models/test_user_permissions_on_repo_groups.py kallithea/tests/models/test_user_group_permissions_on_repo_groups.py --maxfail=1 --tb=short
...
build/bdist.macosx-10.6-intel/egg/sqlalchemy/engine/default.py:331: in do_execute
cursor.execute(statement, parameters)
E IntegrityError: (IntegrityError) UNIQUE constraint failed: users_groups.users_group_name
u'INSERT INTO users_groups (users_group_name, user_group_description, users_group_active, users_group_inherit_default_permissions, user_id, created_on, group_data) VALUES (?, ?, ?, ?, ?, ?, ?)' ('perms_group_1', 'DESC', 1, 1, 2, '2015-04-08 17:38:03.457963', None)
...

tests: Eliminate use of nose.tools.with_setup - it is nose-specific

Instead, simply make `permissions_setup_func` set the baseline permissions.

tests: Make `parameterized` stuff work with both pytest and nose

- Use `nose.tools.nottest` if `nose` is available
- Use `pytest.mark.skipIf` if `pytest` is available
- Rename from `nose_parametrized` to `parameterized` since it's no longer
nose-specific

tox.ini: Set PYTHONHASHSEED = 0

because there is some non-deterministic stuff that can fail without this