issue 15733 - OK to change the test?

> On Mar 3, 2016, at 8:16 AM, PAUL DEBRUICKER <[hidden email]> wrote:
>
> See http://forum.world.st/The-Inbox-System-mt-697-mcz-td4802003.html for the orginial discussion that this fix is attempting to fix.  Basically there is an extra and unnecessary #home send in MessageTally>>#tally:by:  MessageTally>>#tally:in:by: and QSystemTally>>#tally:by:  
>
> Prior to the fix
>
>
> AndreasSystemProfiler spyOn: [
>    [ #((1 2 3)) do: [ :each |
>    each findLast: [ :ea |
>    ea squared = ea ] ] ] bench ].
>
>
>
>
> gives
>
> **Tree**
> 100.0 (5,016)  GLMMorphicPharoScriptRenderer [GLMMorphicPharoCodeRenderer]  actOnHighlightAndEvaluate:
>  100.0 (5,016)  RubSmalltalkEditor  highlightEvaluateAndDo:
>    100.0 (5,016)  RubSmalltalkEditor  evaluate:andDo:
>      100.0 (5,016)  OpalCompiler  evaluate
>        100.0 (5,016)  UndefinedObject  DoIt
>          44.7 (2,242)  Array [SequenceableCollection]  findLast:
>          44.7 (2,242)  Array [SequenceableCollection]  do:
>          1.2000000000000002 (60)  SmallInteger [Number]  squared
>
>
> and after the fix it gives
>
> **Tree**
> 100.0 (5,024)  BlockClosure  newProcess
>  100.0 (5,024)  MorphicUIManager  spawnNewProcess
>    100.0 (5,024)  WorldMorph  doOneCycle
>      100.0 (5,024)  WorldState  doOneCycleFor:
>        100.0 (5,024)  WorldState  doOneCycleNowFor:
>          100.0 (5,024)  WorldMorph  runStepMethods
>            100.0 (5,024)  WorldState  runStepMethodsIn:
>              100.0 (5,024)  GLMMorphicPharoScriptRenderer [GLMMorphicPharoCodeRenderer]  actOnHighlightAndEvaluate:
>                100.0 (5,024)  RubEditingArea [RubAbstractTextArea]  handleEdit:
>                  100.0 (5,024)  GLMMorphicPharoScriptRenderer [GLMMorphicPharoCodeRenderer]  actOnHighlightAndEvaluate:
>                    100.0 (5,024)  RubSmalltalkEditor  highlightEvaluateAndDo:
>                      100.0 (5,024)  RubSmalltalkEditor  evaluate:andDo:
>                        100.0 (5,024)  OpalCompiler  evaluate
>                          100.0 (5,024)  UndefinedObject  DoIt
>                            100.0 (5,024)  AndreasSystemProfiler class  spyOn:
>                              100.0 (5,024)  BlockClosure  ensure:
>                                100.0 (5,024)  AndreasSystemProfiler class  spyOn:
>                                  100.0 (5,024)  AndreasSystemProfiler  spyOn:
>                                    100.0 (5,024)  BlockClosure  ensure:
>                                      100.0 (5,024)  UndefinedObject  DoIt
>                                        100.0 (5,024)  BlockClosure  bench
>                                          100.0 (5,024)  BlockClosure  benchFor:
>                                            99.30000000000001 (4,989)  UndefinedObject  DoIt
>                                              98.4 (4,944)  Array [SequenceableCollection]  do:
>                                                55.300000000000004 (2,778)  UndefinedObject  DoIt
>                                                  53.2 (2,673)  Array [SequenceableCollection]  findLast:
>                                                    7.2 (362)  UndefinedObject  DoIt
>                                                      1.2000000000000002 (60)  SmallInteger [Number]  squared
>
>
>
> Basically a more accurate count that measures where the time is actually spent in blocks.  But that means I'll have to update the MessageTallyTest>>#testTallySends to reflect the new report.  Does that sound OK?