On Sun, Apr 8, 2012 at 2:26 PM, Philippe Marschall <kustos@gmx.net> wrote:

Hi

I extracted a superclass and now I get a MNU during image startup because I have an instance of the now abstract base class in the start up list.

The image comes up for a fraction of a second and the crashes (PharoDebug.log). Any idea of how I can recover from this?

One way is in gdb put a breakpoint in the MNU processing (createActualMessage). Then when you hit the MNU edit the message object to something harmless such as Message selector: #class arguments: #(), and proceed. This isn't easy. You have to learn how to find symbols in the VM (you'll need to find eg. #class) and how to store pointers in an object. So only consider doing this if the image is worth a few hours learning arcane VM innards, and possibly building a debug VM that'll be less fragile in gdb.

Cheers Philippe

-- best, Eliot

For 1.3 see http://code.google.com/p/pharo/issues/detail?can=2&start=0&num=100&q=&colspe... For 1.4 I don't know. Max PS: Still working on a working fix… sorry for the delay but I'm trying to be thorough On 08.04.2012, at 23:26, Philippe Marschall wrote:

Hi

I extracted a superclass and now I get a MNU during image startup because I have an instance of the now abstract base class in the start up list.

The image comes up for a fraction of a second and the crashes (PharoDebug.log). Any idea of how I can recover from this?

Cheers Philippe