Multiple test failures on sparc64

[wljjn]

trafficstars

Several tests failed or crashed on sparc64.

Running testCopyDeepScanLine

Testing raw data copy in DeepScanLineInput/OutputFile:

Testing files with 1 channels 100 times. generating compression 0 writing copying /bin/bash: line 5: 322288 Bus Error (core dumped) ${dir}$tst FAIL: IlmImfTest 1 of 1 test failed make[3]: *** [Makefile:552: check-TESTS] Error 1 make[2]: *** [Makefile:675: check-am] Error 2 make[1]: *** [Makefile:382: check-recursive] Error 1

Running testCopyMultiPartFile/components/openexr/test-failures.txt" 140L, 4940C Testing copying multi-part files Testing file with 2 part(s). pixelType = 2 partType = 2 line order =1 pixelType = 1 partType = 3 tile order =0 levelMode = 0 Writing files copying /bin/bash: line 5: 331140 Bus Error (core dumped) ${dir}$tst FAIL: IlmImfTest 1 of 1 test failed make[3]: *** [Makefile:552: check-TESTS] Error 1 make[2]: *** [Makefile:675: check-am] Error 2 make[1]: *** [Makefile:382: check-recursive] Error 1

Running testOptimizedInterleavePatterns Testing SSE optimisation with different interleave patterns (large images) ... 0, 0: RGBHalf read as RGBHalf... OK 0, 1: RGBHalf read as RGBAHalf... OK 0, 2: RGBHalf read as ABGRHalf... OK 0, 3: RGBHalf read as RGBFloat... OK 0, 4: RGBHalf read as BGRHalf... OK 0, 5: RGBHalf read as RGBLeftRGB... OK 0, 6: RGBHalf read as RGBRightRGB... OK 0, 7: RGBHalf read as RGBALeftRGBA... OK 0, 8: RGBHalf read as RGBARightRGBA... OK 0, 9: RGBHalf read as LeftRGB... OK 0,10: RGBHalf read as RightRGB... OK 0,11: RGBHalf read as LeftRGBA... OK 0,12: RGBHalf read as RightRGBA... OK 0,13: RGBHalf read as TripleView... OK 0,14: RGBHalf read as Trees... OK 0,15: RGBHalf read as TreesAndBirds... OK 0,16: RGBHalf read as RGBLeftRGBA... OK 0,17: RGBHalf read as RGBRightRGBA... OK 0,18: RGBHalf read as RGBALeftRGB... OK 0,19: RGBHalf read as RGBARightRGB... OK 0,20: RGBHalf read as TwinRGBLeftRGB... OK 0,21: RGBHalf read as TwinRGBRightRGB... OK 0,22: RGBHalf read as TwinRGBALeftRGBA... OK 0,23: RGBHalf read as TwinRGBARightRGBA... OK 0,24: RGBHalf read as TripleTripleView... OK 0,25: RGBHalf read as Alpha... OK 0,26: RGBHalf read as RedAlpha... OK 0,27: RGBHalf read as RG+BA... OK 0,28: RGBHalf read as RGBpassiveA... OK 0,29: RGBHalf read as RGBpassiveleftRGB... OK 0,30: RGBHalf read as RGBFloatA... /bin/bash: line 5: 201061 Bus Error (core dumped) ${dir}$tst FAIL: IlmImfTest 1 of 1 test failed make[3]: *** [Makefile:552: check-TESTS] Error 1 make[3]: Leaving directory '/builds/jinji/openexr-upgrade/components/openexr/build/sparcv7/IlmImfTest' make[2]: *** [Makefile:675: check-am] Error 2 make[2]: Leaving directory '/builds/jinji/openexr-upgrade/components/openexr/build/sparcv7/IlmImfTest' make[1]: *** [Makefile:382: check-recursive] Error 1 make[1]: Leaving directory '/builds/jinji/openexr-upgrade/components/openexr/build/sparcv7'

The following one caused the test to stuck here.

Running testMultiPartFileMixingBasic Testing the mixed (ScanLine, Tiled, DeepScanLine and DeepTiled) multi-part file Testing file with 1 part(s). Generating headers and data pixelType = 2 partType = 2 line order =1 Writing files Reading whole files Reading first part Reading partial files simulating incomplete file

Running testFutureProofing Testing reading future-files Testing file with 1 part(s). Generating headers and data pixelType = 2 partType = 2 line order =1 Writing files Reading first part Reading whole files part 0 ok simulating new part type ERROR -- caught exception: Out of Memory Assertion failed: false, file /builds/jinji/openexr-upgrade/components/openexr/openexr-2.2.0/IlmImfTest/testFutureProofing.cpp, line 1382 /bin/bash: line 5: 268985 Abort (core dumped) ${dir}$tst FAIL: IlmImfTest 1 of 1 test failed make[3]: *** [Makefile:552: check-TESTS] Error 1 make[3]: Leaving directory '/builds/jinji/openexr-upgrade/components/openexr/build/sparcv7/IlmImfTest' make[2]: *** [Makefile:675: check-am] Error 2 make[2]: Leaving directory '/builds/jinji/openexr-upgrade/components/openexr/build/sparcv7/IlmImfTest'

[meshula]

@wljjn A great many changes have recently been committed in order to address bit-size correctness across the board. If the sparc64 target is still one that you are interested in, it would be good to know if the work that has been done addresses the problems you were seeing.