On 4 October 2012 17:03, Anton Gulenko
<anton.gulenko at student.hpi.uni-potsdam.de> wrote:
> Dear Frank,
>
> The MethodEngine project is not an implementation of method wrappers.
> The focus is on automatic (but also controlled) generation of program mutations.
> In fact, the MethodEngine can USE an implementation of method wrappers
> to inject a mutation into the program. It could also use other
> mechanisms like directly compiling the modified code or using AOP.

Hi Anton,

That's kind've what I meant: part of the MutationEngine implements
method wrappers, and I've seen implementations elsewhere, and wondered
if making a separate, shared, implementation might not be worthwhile.

frank

> Best regards,
> Anton
>
> 2012/10/4 Frank Shearar <frank.shearar at gmail.com>:
>> On 4 October 2012 09:59, Anton Gulenko
>> <anton.gulenko at student.hpi.uni-potsdam.de> wrote:
>>> Dear Community,
>>>
>>> We would like to announce a new Squeak Community Project by the Software
>>> Architecture Group, Hasso-Plattner-Institute. Our new MutationEngine
>>> is a library
>>> to inject random modifications (mutations) in order to insert defects
>>> into Smalltalk
>>> programs and to make them crash. With that, you can automatically check and
>>> evaluate the effectivity of your debugging and testing approaches. The
>>> MutationEngine also provides a safe clean up at the end so that your Squeak
>>> system is not affected after a critical mutation.
>>
>> This looks very interesting. It's also about the third implementation
>> of method wrappers that I've seen (the others being the in-image
>> example, and ObjectsAsMethodsWrapper), and makes me wonder if we
>> shouldn't have a single best-of-everything library for what's clearly
>> a broadly useful tool.
>>
>> frank
>>
>>> A description, setup instructions and a tutorial can be found in the following
>>> Wiki article:
>>> https://www.hpi.uni-potsdam.de/hirschfeld/trac/SqueakCommunityProjects/wiki/mutations
>>>
>>> Best regards,
>>> Anton
>>>
>>
>