<html>
<head>
<meta content="text/html; charset=windows-1252"
http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
Hi<br>
<br>
The green lamp is going again!<br>
<br>
Thanks a lot to everybody committing documentation updates, fixing
tests and providing inputs for the failing QgsAppStartup test.<br>
<br>
<div class="moz-cite-prefix">On 06/02/2015 02:49 PM, Tim Sutton
wrote:<br>
</div>
<blockquote
cite="mid:346C2260-35B3-4285-A2B9-38DEE729EECB@kartoza.com"
type="cite">
<meta http-equiv="Content-Type" content="text/html;
charset=windows-1252">
Hi
<div class=""><br class="">
<div>
<blockquote type="cite" class="">
<div class="">On 02 Jun 2015, at 18:15, Nathan Woodrow <<a
moz-do-not-send="true" href="mailto:madmanwoo@gmail.com"
class="">madmanwoo@gmail.com</a>> wrote:</div>
<br class="Apple-interchange-newline">
<div class="">
<div dir="ltr" class="">I'm all for strict unit test
policy however they need to be able to run on all
platforms consistently in order to enforce that
correctly. Taking the failing PyQgsAppStartup test on
Travis, i just ran it on my Linux setup and it passes
fine, yet fails on Travis. I'm skeptical of that test
anyway but it does make it hard to fix when it can't be
reproduced..<br class="">
<div class=""><br class="">
</div>
<div class="">Compare to Traivs I get these two tests
failing:</div>
<div class=""><br class="">
</div>
<div class="">
<div class="">The following tests FAILED:</div>
<div class=""><span class="Apple-tab-span"
style="white-space:pre"> </span>100 -
PyQgsLocalServer (Failed)</div>
<div class=""><span class="Apple-tab-span"
style="white-space:pre"> </span>134 -
PyQgsPalLabelingServer (Failed)</div>
</div>
</div>
</div>
</blockquote>
<div><br class="">
</div>
<div>Disabling a test is also a legitimate course of action -
if the test does not run as expected and there is no
reasonable way to see why in python you can tag it as
‘expected failure’ using the expectedFailure decorator.
Obviously that should not be the default action when a test
fails and should not be used to disable tests in masse, but
disabling the test puts travis back into a testable state
for other dev while you figure out the reason for the
regression.</div>
</div>
</div>
</blockquote>
<br>
I just disabled part of this test to have the CI running again. It's
a rather hard-to-debug test since it runs in a subprocess. But this
test prevented me once from breaking the python initialization code,
so it would be good to be able to re-enable it.<br>
<br>
Regards,<br>
Matthias<br>
<br>
</body>
</html>