> On 2020-04-30, at 12:59 AM, Marcel Taeumel <marcel.taeumel at hpi.de> wrote:
> 
> Hi all!
> 
> I will now test 202004292351 as a release candidate for a new VM release to be also used in Squeak 5.3 to fix ARM support.

It looks 'mostly OK' to me.
a) it doesn't segfault on startup; always a bonus.
b) tinyBenchmarks are typical, so likely no broken cog mechanism
c) running the full test suite  -
4726 run, 38 expected failures, 5 failures, 17 errors.

The 6 failures include
 - TextAnchorTest>>#testResizeAnchoredMorph (anchoredMorph is shadowed and the 
self assert: priorExtent y + 15 equals: textMorph height.
fails because priorExtent is 84 and textMorph height is 84 - so adding 15 isn't a happy maker

 - HashedCollectionTest>>#testSizeFor fails because the very first value tested is 0 and we end up noticing the 0.0 / 5 is not > than 0.75. 

 - IntervalTest>>#testIntervalOfFloatLast fails because of floating point insanity. Will we ever be able to trust them?

 - IntervalTest>>#testIntervalOfFloatReversed & #testSurprisingFuzzyInclusionBehavior same basic issue

The errors -

 - AllocationTest>>#testOutOfMemorySignal fails to set the free space limit of 1.5 GB on a 4GB Pi

 - 3 Decompiler tests fail because somebody thought using _ for := is still acceptable a million years after we changed that.

 - DecompilerTests>>#testDecompilerInClassesVAtoVM fails because the first assert has conniptions over the tempVector having voted for trump

 - 12 SqueakSSLTests fail because they know it would be annoying. Only the #testFaceBookAPI passes, which is also to be annoying.

Only the SSL failures seem like a potential problem here - does it indicate the plugin is borked?

tim
--
tim Rowledge; tim at rowledge.org; http://www.rowledge.org/tim
"How many Trinoc does it take to change a lightbulb?” 
"Why do you want to know about our maintenance schedules? Are you planning to attack us in the dark?"