Fix "Sign in as user" acceptance test with PHP 7.0 and MySQL

[danxuliu]

It seems that the tests run slower on PHP 7.0 and MySQL in Travis CI (probably due to the use of XDebug); this caused the Sign in as user acceptance test to fail when some actions were executed on an element that had not appeared yet. To fix that now it is checked that the offending elements have appeared in the page before trying to use them.

[codecov[bot]]

Codecov Report

:exclamation: No coverage uploaded for pull request base (master@dc5a1dd). Click here to learn what that means. The diff coverage is n/a.

@@            Coverage Diff            @@
##             master     #379   +/-   ##
=========================================
  Coverage          ?   82.55%           
  Complexity        ?      360           
=========================================
  Files             ?       38           
  Lines             ?     1313           
  Branches          ?        0           
=========================================
  Hits              ?     1084           
  Misses            ?      229           
  Partials          ?        0

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update dc5a1dd...0248932. Read the comment docs.

[oparoz]

Congrats, all tests have passed :) and a big THANK YOU. This probably needs some cleanup and squashing. I'll merge once that's done.

[danxuliu]

@oparoz

Congrats, all tests have passed :)

I have seen that an API test (ConnectWithTokenCest) sometimes still fail on PHP 7.0 and MySQL with a timeout :-( It may be needed to make the default timeout higher... but before that let's see if it is something common or it was just a one-time problem ;-)

and a big THANK YOU.

A big you are welcome :-)

This probably needs some cleanup and squashing.

Done!

[oparoz]

I have seen that an API test (ConnectWithTokenCest) sometimes still fail on PHP 7.0 and MySQL with a timeout :-( It may be needed to make the default timeout higher... but before that let's see if it is something common or it was just a one-time problem ;-)

Seems to happen regularly. It's the one test using Imagemagick... but it shouldn't time out.

[danxuliu]

Seems to happen regularly. It's the one test using Imagemagick... but it shouldn't time out.

Maybe when Travis servers are loaded, combined with the instrumentation needed for the coverage, the bridge to call the underlaying ImageMagick native libraries is too slow.

In any case, I have increased the timeout in #387 and with the larger timeout it finally seems to work fine :-D (I have restarted the build several times).