On 24 Apr 2014, at 13:46, Camille Teruel <camille.teruel@gmail.com> wrote:


On 24 avr. 2014, at 09:59, Marcus Denker <marcus.denker@inria.fr> wrote:


On 23 Apr 2014, at 19:30, Marcus Denker <marcus.denker@inria.fr> wrote:


On 23 Apr 2014, at 19:21, Camille Teruel <camille.teruel@gmail.com> wrote:


On 23 avr. 2014, at 19:00, Marcus Denker <marcus.denker@inria.fr> wrote:


On 23 Apr 2014, at 18:34, Marcus Denker <marcus.denker@inria.fr> wrote:


On 23 Apr 2014, at 18:18, Sven Van Caekenberghe <sven@stfx.eu> wrote:

And we're all GREEN, great work !
Let's keep it that way until release.
Everything that breaks a test should now be reverted.

Yes!

there are just two issues that are for pharo3 with fixes:

https://pharo.fogbugz.com/f/cases/13028/Adding-ClassVariables-corrupts-class-hierarchy


Ok, so loading that one brakes *everything*.

(the build did not get done, so no problem, other than a red build)

I will revert tomorrow...

Arg, I'm sick of that issue. 
So do we try to integrate it in two passes?


I have no idea what goes wrong� it seems that it works fine when I merge the slice.
But then the update process (which just saved the packages�) does something wrong.

Mayeb we could try to have a version without the method that sends #explicitRequirement


I integrated the issue as a change set (so there is some load order bug related to Traits with MC�).

This seems to have worked: all image tests are green.
BUT: the Metacello Platformtests fail with a �should not happen�  error.

Did you run the postscript? 
Smalltalk allClassesDo: [ :c | c superclass addSubclass: c ]

No, but if that would be inconsistent, there is a test that fails�

But I can do it the next update.

Marcus