[Pharo-project] [Fix of the day contest] 4/7/2012 with extras

Stéphane Ducasse stephane.ducasse at inria.fr
Thu Jul 5 01:39:32 EDT 2012


we will have that with manifest.
Manifest are simple class where we are storing SmallLint false positive and they will be package meta data container.

Stef

On Jul 4, 2012, at 9:49 PM, Chris Muller wrote:

> No one uses it but maybe a good idea to consider -- it seems like a
> good place for things like license, copyright, about text..
> 
> 
> On Wed, Jul 4, 2012 at 2:34 PM, Marcus Denker <marcus.denker at inria.fr> wrote:
>> 
>> On Jul 4, 2012, at 9:25 PM, Guillermo Polito wrote:
>> 
>>> Hi!
>>> 
>>> I've done a little fix today that allows monticello tests to run -green btw- :). http://code.google.com/p/pharo/issues/detail?id=6273
>>> So, I expect this will put us 11 tests closer to a green build.
>>> 
>>> 
>>> This was a problem related with the monticello-mocks and monticellomocks packages.
>>> There are problems still remaining, like, as giving an example:
>>> 
>>> - MCMock is categorized Monticello-Mocks, but is in the Monticello RPackage...
>>> 
>>> That happens with all the classes in the Monticello-Mocks category.  The root of the problem is that we have some strange objects installed as packages (PackageInfo registrations).  Have a look at the result of:
>>> 
>> 
>> The problem is that the MC tests use a featuer of MC that nobody uses: subclasses of PackageInfo to be able to make packages out of code that
>> is actually not following the convention.
>> 
>> See MCMockPackageInfo
>> 
>> it's pacakgeName is "MonticelloMocks", but the idea was originally to never save that package, I think.
>> 
>> 
>> --
>> Marcus Denker -- http://marcusdenker.de
>> 
>> 
> 





More information about the Pharo-dev mailing list