6 testsuite failures building for s390

[Jonas Smedegaard]

Hi,

Building Valentina for s390 fails 6 tests in the testsuite:

FAIL!  : TST_FindPoint::TestCurveIntersectAxis(Angle 0 degree) Compared values are not the same
   Actual   (result)     : QPointF(515.72,238.941)
   Expected (resultPoint): QPointF(515.72,238.941)
   Loc: [tst_findpoint.cpp(552)]
FAIL!  : TST_FindPoint::TestCurveIntersectAxis(Angle 180 degree) Compared values are not the same
   Actual   (result)     : QPointF(77.8886,238.941)
   Expected (resultPoint): QPointF(77.8886,238.941)
   Loc: [tst_findpoint.cpp(552)]
FAIL!  : TST_FindPoint::TestCurveIntersectAxis(Ignore base point) Compared values are not the same
   Actual   (result)     : QPointF(409.367,79.9382)
   Expected (resultPoint): QPointF(409.367,79.9382)
   Loc: [tst_findpoint.cpp(552)]

FAIL!  : TST_VAbstractPiece::EquidistantRemoveLoop(Issue 298. Case1) Compared values are not the same
   Actual   (ekv.size())    : 30
   Expected (ekvOrig.size()): 31
   Loc: [abstracttest.cpp(63)]

FAIL!  : TST_VAbstractPiece::BrokenDetailEquidistant(GAVAUDAN Laure.) Compared values are not the same
   Actual   (ekv.size())    : 19
   Expected (ekvOrig.size()): 21
   Loc: [abstracttest.cpp(63)]

FAIL!  : TST_VAbstractPiece::BrokenDetailEquidistant(Issue #687.) Compared values are not the same
   Actual   (ekv.size())    : 22
   Expected (ekvOrig.size()): 23
   Loc: [abstracttest.cpp(63)]

Full build log: https://buildd.debian.org/status/fetch.php?pkg=valentina&arch=s390x&ver=0.6.1%7Edfsg-3&stamp=1546372398&raw=0

[Roman Telezhynskyi]

TST_VAbstractPiece::EquidistantRemoveLoop TST_VAbstractPiece::BrokenDetailEquidistant(GAVAUDAN Laure.) TST_VAbstractPiece::BrokenDetailEquidistant(Issue #687.)

Add to ignore if you want. You can wrap with ifdef. Different platforms produce slightly different result.

TST_FindPoint

As to this. To fix this need to improve tests. Check is too precise.

[Jonas Smedegaard]

Sorry, can you please guide me with one of them, how to correlate the error message to the source piece needing to be ifdef'ed, then I can do similar for the rest.

Quoting Roman Telezhynskyi (2019-01-01 21:26:59)

[Roman Telezhynskyi]

Improving tests. Comparisons for TST_FindPoint is too precise. ref #918.

→ <<cset 0a47a319a945>>

[Roman Telezhynskyi]

Skip certain tests if architecture is s390 or s390x. ref #918.

→ <<cset a82704b1fc5d>>

[Roman Telezhynskyi]

Improving tests. Comparisons for TST_FindPoint is too precise. ref #918. (grafted from 0a47a319a9450b3d30f0d7f8d00675628744d58a)

→ <<cset 8aa3e2e68f4f>>

[Roman Telezhynskyi]

Skip certain tests if architecture is s390 or s390x. ref #918. (grafted from a82704b1fc5dba81a965616c0d3323bcc601472c)

→ <<cset 9b057f31bc02>>

[Jonas Smedegaard]

Thanks, I can confirm that changeset 0a47a319a945 indeed fixes variability across several archs (not only s390).

Quoting Roman Telezhynskyi (2019-01-02 15:13:14)

[Roman Telezhynskyi]

This repository has been archived. If you want to report a bug or your feature request, you need to do this in the new repository https://gitlab.com/smart-pattern/valentina. This issue ticket will be closed.