pink colour generated in the "retouching" module

[cjose38]

Describe the bug

the pink colour appears when the area to be replaced is close to the edge of the frame. the colour is visible when the module is activated. after a few seconds, the pink colour disappears, leaving the area that matches the one used as a template. When the "retouch" module is closed, the pink area appears again. You can also see this pink in the miniature of the light table. However, this pink colour is not present in the exported file.

I changed the red colour of the grid to green and the area remained pink.

Steps to reproduce

copy an area near the edge using the "retouch" module and the cross tool.

Expected behavior

no pink colour.

Logfile | Screenshot | Screencast

No response

Commit

No response

Where did you obtain darktable from?

downloaded from www.darktable.org

darktable version

4.8.0

What OS are you using?

Windows

What is the version of your OS?

w10 pro

Describe your system?

Device name DESKTOP-ICR7E22 Processor 12th Gen Intel(R) Core(TM) i9-12900K 3.20 GHz Installed RAM 32.0 GB (31.8 GB usable) System type 64-bit operating system, x64 processor

Windows 10 Professional Edition Version 22H2

Are you using OpenCL GPU in darktable?

Yes

If yes, what is the GPU card and driver?

Gigabyte NVIDIA GeForce RTX 3060 EAGLE V2 - 12GB GDDR6 pilot 31.0.15.5176

Please provide additional context if applicable. You can attach files too, but might need to rename to .txt or .zip

No response

[jenshannoschwalm]

Could you share raw file and the xmp?

[cjose38]

thank you. here are the files :+1: Nouveau dossier.zip

[jenshannoschwalm]

Thanks for reporting this. can reproduce here. At first when testing your xmp i thought about some blending or raster mask problem or some sort of race condition but could rule that out, issue is there even with a few modules active.

Please use this xmp as it make issue easier to reproduce with less processing time. 20240629_JOSE0034.ORF.xmp.zip

The real problem is in the lens module (embedded TCA correction) only on the right border, both CPU & OpenCL code. In the screenshot you will see a) you pinkish spot but b) also the pinkish line at the very right of the canvas.

Haven't seen this before, likely a bug in our code, yet it could also be a problem with wrong embedded data. Is this the latest macro lens? Quite impressive btw.

[cjose38]

yes, it's the m.zuiko 90 mm at max magnification. i tested with your xmp and funny thing, to see the pink border with the module closed, i cropped at 100% and the pink disappears! With the same development and a different display the problem appears or not. Thank you for having spent time on the problem. For me, it's not critical because by modifying the points of the mask, you manage not to have this pink area.

[jenshannoschwalm]

I further investigated and in fact it's a bug in the embedded lens correction algo principally valid for all lenses and not only for tca so it's relevant for us all :-) understood and fix following asap.

[cjose38]

very good, thank's.

[jenshannoschwalm]

If you compile dt yourself you might check the above PR

[cjose38]

Sorry, I don't know how. I'll wait for the next version. thanks for the efficiency.