Page 1 of 1

t/filter...FAILED tests 2, 7, 27, 36, 38 Expected???

Posted: 2007-09-04T22:21:06-07:00
by fisheggs
I just downloaded and built ImageMagick-6.3.5-8.tar.gz dated Sep 04 00:00 on Ubuntu Feisty x86_64 GNU/Linux and (just for the hell of it), ran make check.

The output included

Code: Select all

========================================================
All 696 tests behaved as expected (22 expected failures)
========================================================
.
.
.
t/composite.......ok
t/filter..........FAILED tests 2, 7, 27, 36, 38
        Failed 5/58 tests, 91.38% okay
t/getattribute....ok
.
.
.
Failed Test Stat Wstat Total Fail  Failed  List of Failed
-------------------------------------------------------------------------------
t/filter.t                58    5   8.62%  2 7 27 36 38
Failed 1/23 test scripts, 95.65% okay. 5/336 subtests failed, 98.51% okay.
Are the failures of t/filter expected?

Re: t/filter...FAILED tests 2, 7, 27, 36, 38 Expected???

Posted: 2007-09-05T07:14:14-07:00
by magick
It could be a sensitivity issue. Post the output of this command so we can verify:
  • perl t/filter.t

Re: t/filter...FAILED tests 2, 7, 27, 36, 38 Expected???

Posted: 2007-09-05T07:39:52-07:00
by fisheggs

Code: Select all

$ perl t/filter.t 
1..58
AdaptiveThreshold('5x5+5%') ...
ok 1
Annotate(text=>'Magick',geometry=>'+0+20',font=>'Generic.ttf',fill=>'gold',gravity=>'North',pointsize=>14) ...
  mean-error=0.0112754685847058, maximum-error=0.905882352941176
not ok 2
Blur('0.0x1.0') ...
ok 3
Border(geometry=>'6x6',color=>'gold') ...
ok 4
Channel(channel=>'red') ...
ok 5
Chop(geometry=>'80x80+5+10') ...
ok 6
Charcoal('0x1') ...
  mean-error=3.58199545802976e-09, maximum-error=0.00392156862745098
not ok 7
Set(fuzz=>3276) ...
ColorFloodfill(geometry=>"+25+45") ...
ok 8
Set(fuzz=>3276) ...
Colorize(fill=>"red", opacity=>"50%") ...
ok 9
Contrast() ...
ok 10
Convolve([1, 1, 1, 1, 4, 1, 1, 1, 1]) ...
ok 11
Crop(geometry=>'80x80+5+10') ...
ok 12
Set(page=>'0x0+0+0') ...
ok 13
Despeckle() ...
ok 14
Draw(fill=>'none',stroke=>'gold',primitive=>'circle',points=>'60,90 60,120',strokewidth=>2) ...
ok 15
Edge() ...
ok 16
Emboss('0x1') ...
ok 17
Equalize() ...
ok 18
Implode(0.5) ...
ok 19
Flip() ...
ok 20
Flop() ...
ok 21
Frame('15x15+3+3') ...
ok 22
Gamma(1.6) ...
ok 23
GaussianBlur('0.0x1.5') ...
ok 24
Implode(0.5) ...
ok 25
Level(level=>'20%') ...
ok 26
Magnify() ...
  mean-error=2.9849962150248e-10, maximum-error=0.00392156862745098
not ok 27
Set(fuzz=>3276) ...
MatteFloodfill(geometry=>"+25+45") ...
ok 28
MedianFilter() ...
ok 29
Minify() ...
ok 30
Modulate(brightness=>110,saturation=>110,hue=>110) ...
ok 31
Quantize(colorspace=>'gray',colors=>2,dither=>'false') ...
ok 32
MotionBlur('0x13+10-10') ...
ok 33
Negate() ...
ok 34
Normalize() ...
ok 35
OilPaint() ...
  mean-error=6.16461418326922e-06, maximum-error=0.16078431372549
not ok 36
Set(fuzz=>3276) ...
Opaque(color=>"#e23834", fill=>"green") ...
ok 37
Quantize() ...
  mean-error=0.00027541007878002, maximum-error=0.211764705882353
not ok 38
RadialBlur(10) ...
ok 39
Raise('10x10') ...
ok 40
ReduceNoise() ...
ok 41
Resize('60%') ...
ok 42
Roll(geometry=>'+20+10') ...
ok 43
Rotate(45) ...
ok 44
Sample('60%') ...
ok 45
Scale('60%') ...
ok 46
Segment() ...
ok 47
Shade(geometry=>'30x30',gray=>'true') ...
ok 48
Sharpen('0.0x1.0') ...
ok 49
Shave('10x10') ...
ok 50
Shear('-20x20') ...
ok 51
SigmoidalContrast("3x50%") ...
ok 52
Solarize() ...
ok 53
Swirl(90) ...
ok 54
Threshold(90%) ...
ok 55
Trim() ...
ok 56
UnsharpMask('0.0x1.0') ...
ok 57
Wave('25x150') ...
ok 58

Re: t/filter...FAILED tests 2, 7, 27, 36, 38 Expected???

Posted: 2007-09-05T11:18:31-07:00
by magick
All is fine. Some algorithms permit slight variations in the final result and we account for that with a fuzzy quality check. For your case we adjusted the sensitivity in the regression tests so the next release will pass on your system.

Re: t/filter...FAILED tests 2, 7, 27, 36, 38 Expected???

Posted: 2007-09-05T17:38:37-07:00
by fisheggs
Oh, OK, thanks.