Bugzilla – Bug 1214831
[Build 20230830] graphicsMagick fails two tests in openQA
Last modified: 2023-09-04 08:23:27 UTC
## Observation graphicsMagick has been updated to version 1.3.41 (SR included two regression patches in plus) openQA fails all instances of gm tests. The relevant failures reported are: 85 - KO - gm montage degradation.png logo-primary.png ./85-1.png;montage_compare PAE ./85-1.png montage1.png 0 86 - KO - gm montage -texture noise.gif degradation.png logo-primary.png ./86-1.png;montage_compare PAE ./86-1.png montage2.png 0 OK: 104 , ERRORS: 2 openQA test in scenario opensuse-Tumbleweed-DVD-x86_64-system_performance@64bit fails in [graphicsMagick](https://openqa.opensuse.org/tests/3538170/modules/graphicsMagick/steps/18) ## Test suite description Maintainer: QE Core An explicit system performance test. https://progress.opensuse.org/issues/36100 ## Reproducible Fails since (at least) Build [20230830](https://openqa.opensuse.org/tests/3538170) (current job) ## Expected result Last good: [20230828](https://openqa.opensuse.org/tests/3534103) (or more recent) ## Further details Always latest result in this scenario: [latest](https://openqa.opensuse.org/tests/latest?arch=x86_64&distri=opensuse&flavor=DVD&machine=64bit&test=system_performance&version=Tumbleweed)
> gm compare -metric PAE ./0-1.png montage1.png Image Difference (PeakAbsoluteError): Normalized Absolute ============ ========== Red: 0.0000152590 1.0 Green: 0.0000152590 1.0 Blue: 0.0000152590 1.0 Total: 0.0000152590 1.0 > gm compare -metric PAE ./1-1.png montage2.png Image Difference (PeakAbsoluteError): Normalized Absolute ============ ========== Red: 0.0000152590 1.0 Green: 0.0000152590 1.0 Blue: 0.0000152590 1.0 Total: 0.0000152590 1.0
So if I understand correctly, basically $ gm montage degradation.png logo-primary.png ./85-1.png leads to different result than before as shown with $ gm compare -metric PAE ./85-1.png montage1.png (where montage1.png is the previous result) However, the resulting images look the same for naked eye. I have tried also: 15sp3/GraphicsMagick (1.3.35): $ gm montage logo: logo: out.png; md5sum out.png 7a5099d37a6785c57c017224bf80100b out.png $ TW/GraphicsMagick (1.3.40): $ gm montage logo: logo: out.png; md5sum out.png 7a5099d37a6785c57c017224bf80100b out.png $ TW/GraphicsMagick (1.3.41): $ gm montage logo: logo: out.png; md5sum out.png 08c64b70d283466effd924df83bca6a3 out.png $ $ gm compare -metric PAE out-before.png out.png Image Difference (PeakAbsoluteError): Normalized Absolute ============ ========== Red: 0.0000152590 1.0 Green: 0.0000152590 1.0 Blue: 0.0000152590 1.0 Total: 0.0000152590 1.0 $
15sp3/GraphicsMagick (1.3.35): $ gm montage logo: logo: out.png; md5sum out.png 7a5099d37a6785c57c017224bf80100b out.png $ TW/GraphicsMagick (1.3.40): $ gm montage logo: logo: out.png; md5sum out.png 7a5099d37a6785c57c017224bf80100b out.png $ TW/GraphicsMagick (1.3.41): $ gm montage logo: logo: out.png; md5sum out.png 08c64b70d283466effd924df83bca6a3 out.png so 'something' changed... in previous versions, the picture was even bid-identical.. As we likely won't come to a conclusion today, I reverted GM to the previous version in Factory so that we can unblock TW over the weekend. Once we know what is going on here, we can update again to 1.3.41
Created an upstream bug: https://sourceforge.net/p/graphicsmagick/bugs/723/ I have noticed meanwhile that 1.3.41 release got discarded, so yes, my plan was to revert to 1.3.40, even in develproject. Have a nice weekend
https://sourceforge.net/p/graphicsmagick/news/2023/08/because-1341-is-discarded-i-has-been-published-2-builds-for-win32-architecture/
Ah, I have accepted your request because I thought you have solved it. Will revert in devel project again.
https://sourceforge.net/p/graphicsmagick/bugs/723/ Factory: reverted to 1.3.40. In future upstream version this should be fixed.