Changes to Trunk (
http://source.squeak.org/trunk.html) in the last 24 hours:
http://lists.squeakfoundation.org/pipermail/packages/2013-November/006595.htmlName: Tools-nice.502
Ancestors: Tools-cmm.501
Directly ask ToolSet to browseClass: rather than self systemNavigation.
And implement proper specialization in Behavior.
This has two positive effects:
- If I ask (Object class browse) then it browses Object class, not Metaclass,
- If I ask (Trait someInstance browse), then it browses an instance of Trait, not Trait.
Answer the result of this send (a Browser), it can serve eventually.
=============================================
http://lists.squeakfoundation.org/pipermail/packages/2013-November/006596.htmlName: KernelTests-fbs.256
Ancestors: KernelTests-cmm.255
More #shouldnt:raise: Error removals. ClassBuilderTest's makeFooSubclassOf: methods now return a thing rather than setting an instvar, because really. It's just easier to read, and side effects where unnecessary are bad.
=============================================
http://lists.squeakfoundation.org/pipermail/packages/2013-November/006597.htmlName: CollectionsTests-fbs.208
Ancestors: CollectionsTests-fbs.207
"self shouldnt: [foo] raise: Error" -> "foo" because the original form does nothing but obscure real bugs. Use #shouldnt:raise: with a_subclass_ of Error.
Rename some tests where appropriate.
Use #assert:equals: rather than assert: foo = bar, to give better error messages.
=============================================
http://lists.squeakfoundation.org/pipermail/packages/2013-November/006598.htmlName: Tests-fbs.262
Ancestors: Tests-fbs.261
More #shouldnt:raise: Error removals.
A lot of the time one line says "shouldnt: [foo := bar baz] raise: Error" which makes as much sense as "foo := bar baz" only it's less readable.
=============================================