[Qgis-developer] 14 hours of broken travis build in master

[Nyall Dawson]

On 16 February 2016 at 04:37, Sandro Santilli <strk at keybit.net> wrote:
> I've some pulls to merge but current master does not pass travis
> since 14 hours: https://travis-ci.org/qgis/QGIS/builds
>
> The first failure was after Nyall commit:
> https://github.com/qgis/QGIS/commit/dfdcec89223d4167850c7ebbb0af1d65c21135f8
> and is currently still failing in that area ("area" computation):
>
> http://dash.orfeo-toolbox.org/testDetails.php?test=37439889&build=218090
>
> Is anyone on the task of fixing current master build already ?
>
> As a travis build takes about 1 hour, it may be advisable not to
> push a change later than 1.5 hours before the end of the workday,
> so to eventually revert a change that could be breaking.

I posted an earlier email to the list regarding this. For now I've
taken the brute force approach and just disabled that test on the
Travis OSX builds.

Note that it's not a regression, it's a newly added unit test covering
previously untested parts of code which have revealed this issue.

Nyall

>
> --strk;