[Development] CI failing regularly

Nicolás Alvarez nicolas.alvarez at gmail.com
Wed Aug 21 20:49:14 CEST 2013

2013/8/21 Sarajärvi Tony <Tony.Sarajarvi at digia.com>:
> Hi
> The problem was test cases that don't pass. We have a patch that will address these issues: https://codereview.qt-project.org/#change,63361
> Too bad it stumbled upon a machine that apparently was broken (don't understand why, since they should be clones). It's been restaged again.

Is that the correct patch link? It looks more like an illustration of
the problem! It was staged three times and every time it failed on a
different and unrelated test.

> In the long run if we want to have a mechanism that detects what has been changed and which tests should be run, then we need to start designing it. How would it work?

I don't think you can reliably figure out when a test is or isn't
relevant to a change. And the only good reason to do it is if the
tests take too long. If unrelated tests fail, they should be *fixed*,
not skipped when they aren't relevant to the change.


More information about the Development mailing list