Bug 1218102

Summary: [Build 44.1] ImageMagick offset_pattern_fail.gif image is different
Product: [openSUSE] PUBLIC SUSE Linux Enterprise Server 15 SP6 Reporter: Jozef Pupava <jpupava>
Component: OtherAssignee: Petr Gajdos <pgajdos>
Status: RESOLVED INVALID QA Contact:
Severity: Normal    
Priority: P3 - Medium CC: eugenio.paolantonio, jpupava
Version: unspecified   
Target Milestone: ---   
Hardware: Other   
OS: Other   
URL: https://openqa.suse.de/tests/13077109/modules/ImageMagick/steps/222
Whiteboard:
Found By: openQA Services Priority:
Business Priority: Blocker: Yes
Marketing QA Status: --- IT Deployment: ---
Attachments: generated offset_pattern_fail.gif
reference offset_pattern_fail.gif

Description Jozef Pupava 2023-12-15 10:50:35 UTC 
Created attachment 871387 [details]
generated offset_pattern_fail.gif

## Observation

On 15-SP6 with ImageMagick-7.1.1.21-150600.21.x86_64 are differences between reference images and generated image, there are differences like image is ~1 pixel moved, but this one is biggest difference I found.
This is script executed in the test and the failed part of code
https://github.com/os-autoinst/os-autoinst-distri-opensuse/blob/master/data/imagemagick/bg_script.sh#L835

https://openqa.suse.de/tests/13077109/modules/ImageMagick/steps/222

## Reproducible

Fails since (at least) Build [34.1](https://openqa.suse.de/tests/12774078)


## Expected result

Last good: (unknown) (or more recent)


## Further details

Always latest result in this scenario: [latest](https://openqa.suse.de/tests/latest?arch=x86_64&distri=sle&flavor=Online&machine=64bit-2gbram&test=we-module&version=15-SP6)
Comment 1 Jozef Pupava 2023-12-15 10:51:06 UTC 
Created attachment 871388 [details]
reference offset_pattern_fail.gif
Comment 2 Petr Gajdos 2023-12-21 17:34:08 UTC 
 magick -tile-offset +20+20 -tile pattern:checkerboard \
          -size 80x80  xc: -draw 'color 30,20 reset'  offset_pattern_fail.gif

This comes from
https://imagemagick.org/Usage/canvas/

The change happened between following versions:
7.0.1.43
7.0.1.44

https://github.com/ImageMagick/ImageMagick/compare/7.1.0-43...7.1.0-44
Comment 4 Petr Gajdos 2024-01-03 11:28:27 UTC 
Hi Jozef,

this is only problem with the version update, or there are other, similar ones? 

Reading relevant part of
https://imagemagick.org/Usage/canvas/
it seems that you are testing a corner case and the doc discourages to use convert this way. The faulty command is followed by recommended one right after:

magick -size 80x80  xc: \
          -tile-offset +20+20 +size -tile pattern:checkerboard \
          -draw 'color 30,20 reset'  offset_pattern_good.gif

but I assume it is part of your testing already.

Despite my feeling I have filled:
https://github.com/ImageMagick/ImageMagick/discussions/6996
feel free to participate there.
Comment 5 Petr Gajdos 2024-01-04 08:24:56 UTC 
> https://github.com/ImageMagick/ImageMagick/discussions/6996
And there is an answer.

In short, the command behaves differently, but the result is wrong both before and after.

Purpose of
https://imagemagick.org/Usage/canvas/#tile-offset
is actually to show how to create
https://imagemagick.org/Usage/canvas/offset_pattern_good.gif

If there is no other similar regression, we can ignore it I would say.
Comment 6 Jozef Pupava 2024-01-04 10:02:48 UTC 
Hi Petr,

thank you for the gh discussion, I will review our test and results and let you know.
Comment 7 Jozef Pupava 2024-01-04 14:13:36 UTC 
There are no other issues, this test is running also the command generating  offset_pattern_fail.gif but it does not match the expected wrong image, it has to be the correct wrong image. :)
Comment 8 Petr Gajdos 2024-01-05 08:47:52 UTC 
:)
Thanks Jozef