thanks Stef, using it now. On Mon, May 19, 2014 at 3:35 AM, stepharo <stepharo@free.fr> wrote:

Hi Arturo

May be you spot a bug, sadly. I remember that we got a problem in the past and I thought it was fixed. You can use Epicea the new and working change recording system developed by Martin Dias.

Stef

On 18/5/14 14:58, Arturo Zambrano wrote:

...

Hi All,

Trying to recover lost changes from another (broken) image I noticed that only method definition and do-it are saved.

I'm in a middle of a refactoring where I changed hierarchies, renamed classes and inst vars.

When trying to recover the changes, class name changes are not there, so most of the method changes are not applicable.

I don't know if I am missing something important regarding how to recover changes. I just dragged the old changes file onto my new image and then browse the changes (where I can only see methods and do-its)

As a work around I manually did all the class definition changes (superclass, class name, inst var names), and re run all the changes from the "Recent Changes" window. As a result I recovered most of the work of the last night :) (but I lost some time trying to figure out the problem)

Is it OK that my changes file has not class definition changes? Am I missing sth? This is the expected behavior when recovering the changes from another image?

thanks

best regards

arturo

On Mon, May 19, 2014 at 1:49 PM, Sean P. DeNigris <sean@clipperadams.com>wrote:

Also, I've wondered why it's not an error to attempt to apply a change that creates a method on a non-existent class...

+1 when installing code with Gofer/Monticello to non- existent class an exception is raised. seems to be reasonable to expect the same behavior when importing changes...

----- Cheers, Sean -- View this message in context: http://forum.world.st/Class-definition-changes-in-change-log-tp4759413p47595... Sent from the Pharo Smalltalk Users mailing list archive at Nabble.com.