Optimize Diff creation

[badetitou]

By commits

1. I created tests for GLPHEDiffRange>>#newFrom. And optimize this method a bit. I basically killed one string duplication
2. same for GLPHEChange. My bench show two times faster

[badetitou]

The last commit try to intriduce an optimization of the addAll:unless: method.
Here is the script I used to benchmark the update

users := OrderedCollection new.
1 to: 300 do: [ :idx |
	| user |
	user := GLHUser new.
	user id: idx. 
	users add: user ].

[	user2b := GLHUser new.
	user2b id: 298.
	user3b := GLHUser new.
	user3b id: 300.
	user4 := GLHUser new.
	user4 id: 302.

	result := users
		          addAll: {
				          user2b.
				          user3b.
				          user4 }
		          unless: [ :a :b | a id = b id ]] bench
"Before"
"'14 437.625 per second'"
"'16 029.394 per second'"
"After"
"'48 639.688 per second'"
"'44 872.426 per second'"

[alkalinan]

I am nicolas and I approve these commits !

[guillep]

Hi Nicolas! :)

[badetitou]

Hi Nicolas :)

[alkalinan]

I approve this commit !

[NicolasAnquetil]

I am nicolas and I approve this message ;-)