Keeper tests are flaky in CI

[Marcus Tomlinson]

Some tests are failing in CI:

cppcheck:

[/«PKGBUILDDIR»/src/helper/helper.cpp:221]: (warning) Member variable 'Helper::state_changed' is ...
[/«PKGBUILDDIR»/src/service/keeper-user.cpp:28]: (warning) Member variable

helpers-test:

Starting the process...
Error waiting for tar process: "Process operation timed out"
QProcess: Destroyed while process ("tar") is still running.
g_dbus_connection_real_closed: Remote peer vanished with error: Underlying GIOStream returned 0 bytes on an async read (g-io-error-quark, 0). Exiting.
/«PKGBUILDDIR»/tests/integration/helpers/helpers-test.cc:791: Failure
Value of: checkLastStorageFrameworkFile(dir.path().toLatin1().data())
  Actual: false
Expected: true

helper-speed-test:

/«PKGBUILDDIR»/tests/unit/helper/speed-test.cpp:40: Failure
Value of: int(10*helper.percent_done())
  Actual: 6
Expected: int((10*(n_bytes-n_left)/n_bytes))
Which is: 7
/«PKGBUILDDIR»/tests/unit/helper/speed-test.cpp:40: Failure
Value of: int(10*helper.percent_done())
  Actual: 8
Expected: int((10*(n_bytes-n_left)/n_bytes))
Which is: 9

[unity-api-1-bot (unity-api-1-bot)]

Fix committed into lp:keeper/devel at revision 92, scheduled for release in keeper, milestone Unknown

[Charles Kerr (charlesk)]

There are a handful of intermittent failures here that should probably be treated as separate bugs.

Keeping them in a single ticket will just get confusing when the bots merge a fix to one of these bugs into devel and then mark this ticket as 'fixed'.

Separating into new bugs:
* https://bugs.launchpad.net/keeper/+bug/1614652 speed-test
* https://bugs.launchpad.net/keeper/+bug/1614653 TestHelpers fails with 'no task data for ""'
* https://bugs.launchpad.net/keeper/+bug/1614657 TestHelpers fails on percent-done
* https://bugs.launchpad.net/keeper/+bug/1614658 TarCreatorFixture fails on error_string.contains