Yes, post materialization should be added in Debugger>>serializeStack
like it's been done in
HDTestReport>>serializeTestFailureContext:toFileNamed:

To avoid code replication, we may create a extension method to
MethodContext, #serializeOn: that receives a stream.

Martin

On Sat, Feb 16, 2013 at 3:31 PM, Esteban Lorenzano <[email protected]> wrote:
> yep... something like that :)
>
>
> On Sat, Feb 16, 2013 at 3:25 PM, Ben Coman <[email protected]> wrote:
>>
>> Camillo Bruni wrote:
>>>
>>> On 2013-02-16, at 14:14, Camillo Bruni <[email protected]> wrote:
>>>
>>>
>>>>
>>>> On 2013-02-16, at 14:11, Ben Coman <[email protected]> wrote:
>>>>
>>>>
>>>>>
>>>>> Marcus Denker wrote:
>>>>>
>>>>>>
>>>>>> Hi,
>>>>>>
>>>>>> To make it easy to explore why a test fails on the test server we
>>>>>> serialize the stack of falling tests.
>>>>>>
>>>>>> With two changes this has now gotten much easier:
>>>>>>
>>>>>> 1) Drag and Drop support for .fuel file
>>>>>> Just drag the file to the Pharo Desktop --> Debugger opens
>>>>>>
>>>>>>
>>>>>
>>>>> Could you advise how to do this from an operational image (that is, not
>>>>> a CI test image).  I would like to use this to demonstrate Pharo's
>>>>> outstanding development environment.     A killer feature would be the
>>>>> debugger having a <File out> button, that a user could use to mail to
>>>>> support.  Does any other programming system allow this ?
>>>>>
>>>>
>>>> you can just select "fileout / serialize stack" in the debugger
>>>> somewhere.
>>>> The you get a local .fuel file with the serialized stack :)
>>>>
>>>
>>>
>>> it's the last entry "File out Stack" in the right-click menu of the
>>> debugger
>>>
>>>
>>
>> Brilliant!  I thought the idea was obvious :)
>>
>> Now that sparks a following thought, what might be interesting for
>> customer/support interaction would a mechanism packaging together:
>> * the serialized stack
>> * a manifest of all loaded ConfigurationOfXXX versions
>> * the diff of any dirty ConfigurationOfXXX
>> subject to user review to avoid leaking proprietary information.
>>
>> cheers, Ben
>>
>

Reply via email to