[squeak-dev] #reompileAll causes segfaults on both Mac and Pi 32bit systems

tim Rowledge tim at rowledge.org
Wed Aug 7 19:17:18 UTC 2019


We talked briefly at the board meeting this morning about the experimental Sista conversion and so I quickly tried out the 'ReleaseBuilder recompileAll' snippet on a Pi too see how long it takes; thinking it would still take ages.

It actually proceeded quite quickly but unfortunately causes a segfault about one third of the way through. So I tried a clean image instead of my working one - same problem. On my mac, using a recent #18602/32bit image - same problem - and interestingly a very, very, similar crash dump trace. So probably not a Pi-only problem.

The trace was virtually identical - the stack exactly, the list of recent prims very similar but not identical.

A 5.2 #18229 64 bit image on Mac ran the snippet ok.
A 5.2 #18229 32bit image on the Pi got most of the way through but crashed with an again almost identical dump.

This is the trace part - 
Smalltalk stack dump:
0xbfeea95c M [] in PasteUpMorph>processEvent:using: 0x6953ac0: a(n) PasteUpMorph
0xbfeea97c M FullBlockClosure(BlockClosure)>ensure: 0x63ca090: a(n) FullBlockClosure
0xbfeea9a8 I PasteUpMorph>processEvent:using: 0x6953ac0: a(n) PasteUpMorph
0xbfeea9c8 M PasteUpMorph(Morph)>processEvent: 0x6953ac0: a(n) PasteUpMorph
0xbfeea9f0 M [] in HandMorph>sendEvent:focus:clear: 0x6b7a4e8: a(n) HandMorph
0xbfeeaa10 M BlockClosure>ensure: 0x63c9f70: a(n) BlockClosure
0xbfeeaa38 I MouseEvent(MorphicEvent)>becomeActiveDuring: 0x63c9bd8: a(n) MouseEvent
0xbfeeaa60 M [] in HandMorph>sendEvent:focus:clear: 0x6b7a4e8: a(n) HandMorph
0xbfeeaa80 M BlockClosure>ensure: 0x63c9e78: a(n) BlockClosure
0xbfeeaaa0 M HandMorph>becomeActiveDuring: 0x6b7a4e8: a(n) HandMorph
0xbfeeaac8 M [] in HandMorph>sendEvent:focus:clear: 0x6b7a4e8: a(n) HandMorph
0xbfeeaae8 M BlockClosure>ensure: 0x63c9d78: a(n) BlockClosure
0xbfeeab10 I PasteUpMorph>becomeActiveDuring: 0x6953ac0: a(n) PasteUpMorph
0xbfeeab34 M HandMorph>sendEvent:focus:clear: 0x6b7a4e8: a(n) HandMorph
0xbfeeab58 M HandMorph>sendMouseEvent: 0x6b7a4e8: a(n) HandMorph
0xbfeeab78 M HandMorph>handleEvent: 0x6b7a4e8: a(n) HandMorph
0xbfeeaba0 M MouseOverHandler>processMouseOver: 0x85346f0: a(n) MouseOverHandler
0xbfeeabcc M HandMorph>processEvents 0x6b7a4e8: a(n) HandMorph
0xbfeeabe8 M [] in WorldState>doOneCycleNowFor: 0x6aff9e0: a(n) WorldState
0xbfeeac0c M Array(SequenceableCollection)>do: 0x658e098: a(n) Array
0xbfeeac28 M WorldState>handsDo: 0x6aff9e0: a(n) WorldState
0xbfeeac48 M WorldState>doOneCycleNowFor: 0x6aff9e0: a(n) WorldState
0xbfeeac64 M WorldState>doOneCycleFor: 0x6aff9e0: a(n) WorldState
0xbfeeac88 I PasteUpMorph>doOneCycle 0x6953ac0: a(n) PasteUpMorph
0xbfeeaca0 M [] in MorphicProject>spawnNewProcess 0x6c4fbb0: a(n) MorphicProject
0xbfeeacc0 I [] in BlockClosure>newProcess 0x911c430: a(n) BlockClosure

Bit puzzled. A FullBlockClosure issue?

tim
--
tim Rowledge; tim at rowledge.org; http://www.rowledge.org/tim
Strange OpCodes: OI: Vey




More information about the Squeak-dev mailing list