[squeak-dev] The Trunk: Tests-fbs.205.mcz

Bert Freudenberg bert at freudenbergs.de
Tue May 7 19:27:40 UTC 2013


On 2013-05-07, at 16:00, Frank Shearar <frank.shearar at gmail.com> wrote:

> On 7 May 2013 14:35, Levente Uzonyi <leves at elte.hu> wrote:
>> On Mon, 6 May 2013, commits at source.squeak.org wrote:
>> 
>>> Frank Shearar uploaded a new version of Tests to project The Trunk:
>>> http://source.squeak.org/trunk/Tests-fbs.205.mcz
>>> 
>>> ==================== Summary ====================
>>> 
>>> Name: Tests-fbs.205
>>> Author: fbs
>>> Time: 4 May 2013, 9:29:08.634 pm
>>> UUID: add31da5-c70b-4e15-a7a6-f3c95d0d651b
>>> Ancestors: Tests-fbs.204
>>> 
>>> Add a #cleanUp so that ReleaseBuilder doesn't have to re-initialise
>>> MCMockPackageInfo.
>> 
>> 
>> I know I'm a bit late with this, but this is not exactly what the previous
>> behavior was. Implementing
>> 
>> cleanUp: aggressive
>> 
>>        aggressive ifTrue: [ self initialize ]
>> 
>> is more like what it was before. I'm not sure if it's a problem to call
>> #initialize during #cleanUp (or if it's necessary at all), but the behavior
>> changed. The same is true for the GetText package.
> 
> "Before" was quite a while before, because ReleaseBuilder (formerly
> known as ReleaseBuilderFor4dot5) has run "MCMockPackageInfo
> initialize" since ReleaseBuilder-fbs.90, in December 2012.
> 
> To be clear, you're saying that GetText's and Tests' cleanUp should be
> removed and cleanUp: aggressive implemented instead? I'm happy to do
> that, as long as we're clear that in a release process an aggressive
> cleanup's what we want.
> 
> frank

I meant to write something about this, but didn't have time to look at the code involved ... but I think Levente is right :)

- Bert -




More information about the Squeak-dev mailing list