Fix replies hover highlight

Clicking on the share icon didn't result in the share link being shown.
For now, the simplest way to fix this is to ensure that we get the
correct container element by using event.currentTarget (which will
always be the <a> tag) rather than event.target (which might be the <i>
tag).

Previously, the event from the initial click on the "share" event was
swallowed by the click handler on the parent thread:

    https://github.com/hypothesis/h/blob/fef9f43/h/static/scripts/directives/thread.coffee#L198-L221

Removing this has meant that the share dialog never shows.

Incidentally, this also fixes #1802, which was caused by the same code
in the thread directive, which swallowed all click events from its
children and hence prevented the closure of the share dialog.

The reply count being drawn in the thread directive causes display
issues if we want it to be in its current position alongside the
annotation controls.

This change moves the responsibility for (optionally) drawing a reply
count into the annotation directive. It results in simpler, clearer CSS
and in particular makes it possible to trigger a click event on the
reply count element.

It doesn't make any sense to uncollapse threads which have no replies,
so bake this into the toggleCollapsed method of the ThreadController.

Refine extension differentiators

Thread collapse clarity (and code clarity improvements)

This moves almost all of the logic determining what bits of the thread
to render when out from the template and into the controller where it
belongs.

Make threads start collapsed. This is possible because it now doesn't
affect the visibility of top-level cards.

The meaning of "collapsed" in the context of a thread was immensely
confusing:

- on a top-level annotation card it referred to whether replies were or
  were not shown
- on a reply, it referred to whether or not the entire annotation was
  hidden

This change removes all reference to the latter meaning. Annotations
(and replies) are still hidden appropriately when searches are
performed, but now collapsing and uncollapsing refers only to the
visibility of the *replies* of the addressed thread.

With the viewer template managing the root container, this is no longer
necessary.

This stops thread cards with no replies changing in height when clicked.

Also removes one unnecessary ng-hide directive.

Accoring to the thread directive options this component is optional. It
is, however, used extensively by the directive template. This isn't a
nice fix, but it does make it possible for the thread tests to pass.

The thread directive is doing rather a lot:

- rendering a list of cards
- rendering a card
- rendering a reply within a card

This change relieves the thread directive of the first of these
responsibilities.

The require'd 'thread' controller is not used.

Move frontend tests alongside tested modules

Use argparse to implement the main hypothesis CLI

Set the base path for the karma test runner to h/static/scripts.

The only slight subtlety here is that client templates are now keyed by
their basename: loading them from above the base path results in the
paths being absolute when they get to ng-html2js-preprocessor, so
stripPrefix doesn't work.

This commit moves our frontend unit tests into 'test' folders alongside
the modules they are intended to test.

Makes frontend tests consistent with Python tests as of 161baf9.

Change 'yesterday' timestamp to 'a day ago'

Fix og:title output; Handle list or string value

Browserify ALL THE THINGS

Use key derivation to provide secret keys

Add worker runner (hypothesis-worker)

Travis is being especially slow at the moment.

This brings bugfixes for infinite loop on FF.

Fixes 1983.

Fix notifications

Use karma-browserify to run the tests, rather than relying on prebuilt
assets. This commit alters our test harness so that it no longer relies
(at all) on the webassets pipeline. This means:

- test isolation is improved, as each test file runs in its own module
- bringing objects into test scope is now simply a matter of requiring
  the correct module, rather than bolting more and more objects onto the
  Annotator namespace
- autoreloading of tests when either test files or their dependencies
  are modified is now possible (and enabled by default)

This includes files that need to be shimmed under the "browserify-shim"
key. Annotator plugins that have no dependencies do not need to be
shimmed at the moment as they just attach themselves to the global
Annotator object.

jQuery is assigned as a global, this means that it is assumed to be
included separately and bound to the global scope.

Both jQuery and Annotator have aliases under the "browsers" key to
make them easier to reference.

Use a singular time expression one year

Close #1973

Removing some obsolete conditional code,
which was necessary for some obsolete version
of PDF.js.

Fortunately, by now, all supported versions
of PDF.js uses the same data format, so this
workaround can go.

Earlier, on some versions of PDF.js, we used to
use the PDF.js's FindController as a data source
for the text extraction.

However, at some point, we stopped using the
routines shipped with it, since it didn't always
provide use adequate spacing between the various
pieces of texts. So we ended up just contenating
the various pieces of text ourselves.

Then, for new versions of PDF.js, we introduced
other means of accessing the same information,
completely bypassing the PDFFindController.

This change simply unifies the access; now we can
do the same an all PDF.js versions.

Allow using the API as a virtual root

Add highlight test and change scrolling