[Pkg-octave-devel] Bug#871214: Bug#871214: octave-image FTBFS on i386: test hangs

[Rafael Laboissière]

* Adrian Bunk <bunk at debian.org> [2017-08-07 01:57]:

> Source: octave-image
> Version: 2.6.1-1
> Severity: serious
> Tags: buster sid
>
> Some recent change in unstable and buster makes octave-image FTBFS on i386:
>
>  https://tests.reproducible-builds.org/debian/history/octave-image.html
>  https://tests.reproducible-builds.org/debian/rb-pkg/unstable/i386/octave-image.html
>
>  ...
>  Checking CC files ...
>  warning: function /build/1st/octave-image-2.6.1/inst/private/iscolormap.m shadows a core library function
>  warning: called from
>     /tmp/filePMKFm8 at line 1 column 1
>  [__spatial_filtering__]
>  PASSES 21 out of 21 tests
>  [graycomatrix]
>  PASSES 2 out of 2 tests
>  [watershed]
>  ***** test
>  im = [
>      2     3    30     2
>      3    30     3    30
>    255    31    30     4
>      2   255    31    30
>      1     2   255     5];
>
>  labeled8 = [
>      1     1     0     3
>      1     1     0     3
>      0     0     0     0
>      2     2     0     4
>      2     2     0     4];
>  assert (watershed (im), labeled8);
>  assert (watershed (im, 8), labeled8);
>  !!!!! test failed
>  ASSERT errors for:  assert (watershed (im),labeled8)
>
>   Location  |  Observed  |  Expected  |  Reason
>    (3,4)          3            0         Abs err 3 exceeds tol 0
>    (4,4)          0            4         Abs err 4 exceeds tol 0
>  [bwdist]
>  Sat Sep  8 05:36:25 UTC 2018 - pbuilder was killed by timeout after 18h.

I would guess that the culprit is the following unit test in bwdist:

 ## The quasi-euclidean method is apparently sensitive to a machine precision
 ## error that happens in x86 systems only. This test will cause an endless
 ## loop in case of a regression.
 %!test
 %! bw = [  0   1   1   0   0   0   1   0
 %!         0   0   0   0   0   0   0   0
 %!         1   1   0   0   0   0   0   0
 %!         0   0   0   0   0   0   1   0
 %!         0   0   0   0   1   0   0   1
 %!         0   0   0   0   0   0   0   0
 %!         1   0   0   0   0   0   0   0
 %!         0   0   1   0   0   1   1   0];
 %! out = single ([
 %! 1.00000   0.00000   0.00000   1.00000   2.00000   1.00000   0.00000   1.00000
 %! 1.00000   1.00000   1.00000   sqrt(2)   sqrt(2)+1 sqrt(2)   1.00000   sqrt(2)
 %! 0.00000   0.00000   1.00000   2.00000   2.00000   sqrt(2)   1.00000   sqrt(2)
 %! 1.00000   1.00000   sqrt(2)   sqrt(2)   1.00000   1.00000   0.00000   1.00000
 %! 2.00000   2.00000   2.00000   1.00000   0.00000   1.00000   1.00000   0.00000
 %! 1.00000   sqrt(2)   2.00000   sqrt(2)   1.00000   sqrt(2)   sqrt(2)   1.00000
 %! 0.00000   1.00000   1.00000   sqrt(2)   sqrt(2)   1.00000   1.00000   sqrt(2)
 %! 1.00000   1.00000   0.00000   1.00000   1.00000   0.00000   0.00000   1.00000
 %! ]);
 %! assert (bwdist (bw, "quasi-euclidean"), out);

Should we run the test above only on non-x86 architectures?

Rafael



>
> _______________________________________________ 
> Pkg-octave-devel mailing list 
> Pkg-octave-devel at lists.alioth.debian.org 
> http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-octave-devel
>