[squeak-dev] Re: Improved MVCToolBuilderTests in source.squeak.org/tests

Andreas Raab andreas.raab at gmx.de
Sun Jul 12 22:15:48 UTC 2009


Bernhard Pieber wrote:
> I just looked into the problem a little bit further. It seems that the 
> new version is not loaded from the test repository because there is no 
> update.mcm file there. Now I don't understand anymore why my 
> ToolBuilder-MVC-bp.14 package version was loaded by the update mechanism.

That's because I had copied that earlier into trunk for a different 
reason. See previous message.

Cheers,
   - Andreas

> 
> Confused,
> Bernhard
> 
> Am 12.07.2009 um 19:18 schrieb Bernhard Pieber:
> 
>> Thanks, Ian!
>>
>> However, the update does not seem to work, though. I tried:
>> MCMcmUpdater updateFromRepositories: 
>> #('http://source.squeak.org/trunk' 'http://source.squeak.org/tests')
>>
>> The "load code updates" button did not work either, i.e. the version 
>> KernelTests-p4s.64.mcz was not updated.
>>
>> Cheers,
>> Bernhard
>>
>> Am 12.07.2009 um 00:26 schrieb Ian Trudel:
>>
>>> People hi,
>>>
>>> It's great to see that Bernhard has taken the initiative. In my mind,
>>> writing tests in spare time is exactly that. A little thing here, a
>>> little thing there, then next thing you know everything is properly
>>> tested.
>>>
>>> I have also decided to conduct a small experiment with one of my
>>> programmer. He has absolutely no prior experience with Squeak or
>>> Smalltalk. His background is Delphi (personal projects) and Ruby with
>>> NetBeans (professionally).
>>>
>>> The experiment is relatively straightforward. He has to write few
>>> tests every week but without much of instructions. We keep a small
>>> diary to be able to provide data on to the community later. I have
>>> however gave him a presentation how to sync with the Squeak community
>>> trunk and tests. Then, we have connected on his computer for the first
>>> tests he would write, and I have supervised. That's about as much
>>> instructions as he will get.
>>>
>>> The first presentation lasted 80 minutes and it also took 80 minutes
>>> to write and commit tests for SmallInteger#odd and SmallInteger#even.
>>> Both few days apart.
>>>
>>> One of the most obvious observation is the confusion when he has to
>>> perform specific actions, which is likely to be a button or in a menu.
>>> Even with the buttons in his face, he would search in EVERY menu
>>> Squeak could have. Just like our favourite headless chicken. What it
>>> indicates me is the lack of clarity in the menus, it's disorganized
>>> and counter intuitive. It came to one point that I HAD TO TELL HIM
>>> where to find what he needed. Confusion (as in completely lost) is
>>> what described our 80 minutes remote desktop session, when he was the
>>> one in command.
>>>
>>> He has commited code under the initials p4s. You can comment here and
>>> I'll pass on him. Hopefully, a small contribution every week but small
>>> is beautiful.
>>>
>>> Ian.
>>> -- 
>>> http://mecenia.blogspot.com/
>>>
>>
>>
> 
> 
> 




More information about the Squeak-dev mailing list