The Trunk: Kernel-topa.834.mcz

Tobias Pape uploaded a new version of Kernel to project The Trunk:
http://source.squeak.org/trunk/Kernel-topa.834.mcz

==================== Summary ====================

Name: Kernel-topa.834
Author: topa
Time: 25 January 2014, 11:55:18.392 pm
UUID: 0184b773-5c2a-40a9-9922-3b1ba2443767
Ancestors: Kernel-topa.833

Move trait-backstops back to Kernel, so trait unloading still works.

My previous change is now unecessary.

=============== Diff against Kernel-topa.833 ===============

Item was added:
+ ----- Method: Behavior>>allTraits (in category 'Backstop-Traits') -----
+ allTraits
+ "Backstop. When traits are unloaded there are no traits hiding here."
+ ^#()!

Item was added:
+ ----- Method: Behavior>>hasTraitComposition (in category 'Backstop-Traits') -----
+ hasTraitComposition
+ "Backstop. When traits are unloaded we never have a trait composition"
+ ^false!

Item was added:
+ ----- Method: Behavior>>traitComposition (in category 'Backstop-Traits') -----
+ traitComposition
+ "Backstop. When traits are unloaded we never have a trait composition"
+ ^#()!

Item was added:
+ ----- Method: Behavior>>traitCompositionString (in category 'Backstop-Traits') -----
+ traitCompositionString
+ "Backstop. Monticello needs a traitCompositionString even with traits unloaded"
+ ^'{}'!

Item was changed:
  ----- Method: ClassDescription>>obsolete (in category 'initialize-release') -----
  obsolete
  "Make the receiver obsolete."
+ self hasTraitComposition ifTrue: [
+ self traitComposition do: [:each |
+ each removeTraitUser: self]].
- (self respondsTo: #hasTraitComposition) ifTrue: [
- self hasTraitComposition ifTrue: [
- self traitComposition do: [:each |
- each removeTraitUser: self]]].
  superclass removeSubclass: self.
  self organization: nil.
  super obsolete.!

On 25 January 2014 22:55,  <[hidden email]> wrote:
This means that in an image with unloaded Traits we'd still have those
backstops in Kernel, right? At some stage I'd like to see that not
happen. We can take another run at the problem in 4.6.

frank

yes.
note that reliance on these backstop methods is dispersed all over the system,
starting with the browsers, ranging over the compiler and not ending in  monticello.
this will be an intense task.

--
Tobias Pape
sent from a mobile device