After reading bug 527682 I also remembered that some test failures in the past were related with this random DISPLAY value, also not unsetting it makes some tests to ever draw windows while running making them really annoying.
I am 100% for unsetting this, but I am unsure if maybe this would lead to many more gnome packages needing virtualx.eclass to be run and, then, I am unsure if maybe we should have a default src_test phase running the tests always inside virtualx :|
Thanks for your thoughts
The virtualx eclass was recently changed, maybe we should add that there ?
I am unsure because lots of this packages only show this problems when a DISPLAY is found but, when it's not set (like, for example, when building from console) they won't show any problem and, then, they won't need to use virtualx.eclass
Hum, I see what you mean. In any case, if the package does indeed need X, we will need to fix its src_test. We will need to check this to some extent (like emerge -e @world) before pushing this to the tree.
(In reply to Gilles Dartiguelongue from comment #3)
> Hum, I see what you mean. In any case, if the package does indeed need X, we
> will need to fix its src_test. We will need to check this to some extent
> (like emerge -e @world) before pushing this to the tree.
We can maybe introduce this only for EAPI6 to be safer ;)
[master dc527ce] gnome2.eclass/gnome2-utils.eclass: implement eapi6 support and take advantage of that for finally fixing many bugs/enhancements that were waiting for this (#571378)
2 files changed, 113 insertions(+), 47 deletions(-)