2014-10-29 9:30 GMT+01:00 Max Leske <maxleske@gmail.com>:
On 29.10.2014, at 08:22, Marcus Denker <marcus.denker@gmail.com> wrote:
On Wednesday, Oct 29, 2014 at 8:18 am, Max Leske <maxleske@gmail.com>, wrote:
On 29.10.2014, at 01:17, stepharo wrote:
Good question :).
I do not know I would say that having the package Unpackaged would be better than nil @free.fr>
So maybe the bug is that it should not be allowed to pass nil as category?
@free.fr>
We need the concept of "this is code that is automatically generated, it is not part of any package and this is meant to be like that".
Nautilus (or RPackage?) does provide the 'Unpackaged' category. Wouldn't that be "good enough" for now?
RPackage :) Thierry
(there is now code in a package Generated-code-non-existing-package that is of that kind, but the problem is that this package is always dirty of code gets generated).
https://pharo.fogbugz.com/f/cases/14097/generated-code-non-existing-package-...
Would be nice if someone would implement a solution for this.
For libgit2 I simply sublassed NB and put the auto-generated methods into the protocols I choose. Of course that's not a general solution.
On 28/10/14 16:26, Max Leske wrote:
I just fixed a couple of Fuel tests that failed because it is possible to create a trait with the category set to nil. When creating a new trait with the same name this results in an exception being signaled. Should we consider this a bug? It's not cool to create a trait in category nil but since it's possible... Should I open an issue?
Cheers,
Max
On 28.10.2014, at 20:42, Max Leske wrote:
Oh yeah! Maybe now the Fuel tests will be back to green :)
On 28.10.2014, at 20:23, stepharo wrote:
Thanks esteban!
I'm unstuck :)
On 28/10/14 07:41, GitHub wrote:
40336
14337 Removing a trait usage raises an exception (2nd attempt)
@free.fr>@gmail.com>@free.fr>