This enables CSP in dev and CI

Using the sed script from
https://gitlab.com/gitlab-org/gitlab-ce/issues/59758

Fix whitespace in specs because normalize_ws is slightly
different from Capybara 2 behavior

This is a temporary measure to prevent some system specs
causing Chrome to crash

Hoping that we don't hit into

    Selenium::WebDriver::Error::NoSuchDriverError:
      invalid session id

Again.

Signed-off-by: Rémy Coutable <remy@rymai.me>

According to the Chrome source code (https://chromium.googlesource.com/chromium/src/base/+/master/base_switches.cc#120):

The /dev/shm partition is too small in certain VM environments, causing Chrome
to fail or crash (see http://crbug.com/715363). Use this flag to work-around
this issue (a temporary directory will always be used to create anonymous
shared memory files).

Addresses gitlab-org/gitlab-ee#4252 but doesn't appear to cure it completely.

Signed-off-by: Rémy Coutable <remy@rymai.me>

Don't reset the session when the example failed, because we need capybara-screenshot to have access to it

Closes #30876

capybara/rspec already calls Capybara.reset_sessions! in an
`after` hook, but `block_and_wait_for_requests_complete` is called
before it so by calling it explicitely here, we prevent any new
requests from being fired.
See
https://github.com/teamcapybara/capybara/blob/ffb41cfad620de1961bb49b1562a9fa9b28c0903/lib/capybara/rspec.rb#L20-L25



Signed-off-by: Rémy Coutable <remy@rymai.me>

This change fix a memory leak due to a Webkit bug:
https://github.com/ariya/phantomjs/issues/12903



Also:

- Whitelist only localhost and 127.0.0.1 in Capybara + JS specs
- Blacklist all requests to media such as images, videos, PDFs, CSVs etc.
- Log all the requests made.

Signed-off-by: Rémy Coutable <remy@rymai.me>

Signed-off-by: Rémy Coutable <remy@rymai.me>

Signed-off-by: Rémy Coutable <remy@rymai.me>

Also disabled assets warming for Capybara/RSpec and remove a check to
skip assets warming when ./tmp/cache/assets/test exists because assets
are now served by webpack-dev-server.

Signed-off-by: Rémy Coutable <remy@rymai.me>

We were unintentionally hitting `gravatar.com` whenever a test that used
Poltergeist was run. This was certainly wasting their resources and
slowing down our tests even further, for no reason.

The problem occurred because asset compilation takes a long time, so
when the asset cache didn't exist and the first test ran, it would often
(randomly) time out during the generation before the actual test even
had a chance to run.

Now we check if the cache exists before the suite runs, and if not, we
manually fire a request to the root URL in order to generate it. This
should allow subsequent tests to use the cached assets.