From the perspective of a TestCase...
* "running" a test wraps execution in error handling so it can log errors without TestRunner breaking open a debugger.��
* "debugging" executes without that error handling so an error breaks open a debugger.

The functionality required is that a debugger opens immediately on the first line of the test method.
cheers -ben

On Sun, Apr 12, 2015 at 3:35 AM, Max Leske <maxleske@gmail.com> wrote:

On 11 Apr 2015, at 21:31, Ben Coman <btc@openinworld.com> wrote:



On Sun, Apr 12, 2015 at 3:07 AM, Ben Coman <btc@openinworld.com> wrote:
Returning to build 40486 which exhibits the freeze, as a first step, why don't we replace all that complicated stack manipulation stuff with...

�� �� ��NautilusUI>>debugTest: aMethod
[ aMethod methodClass debug: aMethod selector ] fork.


Sorry. Pretend I didn't say that. That is the <run test> behaviour.

Really? Why does it send #debug: then??

cheers -ben��