Unexpected messages on terminal when styles preview is calculated

[pehar1]

Is there an existing issue for this?

• I checked and did not find my issue in the already reported ones

Describe the bug

Beeing in darkroom, selecting the the button for "quick access for applying styles", then hovering over specific styles generates (error ?) messages on terminal.

Steps to reproduce

• Set Preferences -> Processing -> Auto-apply pixel workflow defaults = None (did not ckeck other settings)

• Download Sony Alpha 77 ( filmic ).dtstyle.txt and Sony ILCE 7M4 (filmic).dtstyle.txt

• Remove extension .txt

• Import the styles (lighttable -> styles -> import)

• Use an unedited image in darkroom

• Select button "quick access for applying styles"

• Hover mouse over Sony Alpha 77 ( filmic )
  

• See messages (repeated four times)

   182,3152 cannot get iop-order for rasterfile instance 0                                                     
   182,3153 cannot get iop-order for agx instance 0       

• Hover mouse over Sony ILCE 7M4 (filmic)

• See messages (repeated five times)

   62,9786 cannot get iop-order for agx instance 0

The messages appear when the preview is shown. Both styles do not contain agx and / or rasterfile. Happens with fresh install and also with current version for daily work of dt 5.5.0+303~gb075af99fd.

Expected behavior

Do not throw (error) messages for modules not used

Logfile | Screenshot | Screencast

No response

Commit

No response

Where did you obtain darktable from?

self compiled

darktable version

5.5.0+303~gb075af99fd

What OS are you using?

Linux

What is the version of your OS?

Ubuntu 24.04

Describe your system

OS : Linux 6.8.0-100-generic / Ubuntu 24.04.4 LTS
Memory : MemTotal: 65759196 kB
Graphics Card : Product Name : NVIDIA GeForce RTX 2060
Graphics Card : Driver Version : 580.126.09
OpenCL installed : Device OpenCL C Version OpenCL C 1.2
OpenCL activated : yes
Xorg : Version: 1:7.7+23ubuntu3
Desktop : LXQt
GTK+ : 3.24.41
Glib : ldd (Ubuntu GLIBC 2.39-0ubuntu8.7) 2.39
gcc (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0

Are you using OpenCL GPU in darktable?

Yes

If yes, what is the GPU card and driver?

No response

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

No response

[TurboGit]

The image is unedited, but was it imported some time ago? Before AgX & rasterfile modules where present in Darktbale? This could be the issue here.

[TurboGit]

So for rasterfile before May 10 2025.

[pehar1]

The image has been imported mid december 2025 with current master at this time. IIRC I also tried a history reset which did not change the behavior. Other styles (without filmic and sigmoid) do not generate the messages. I can do more tests tomorrow morning, tell me what I should test. I guess you can't reproduce with my styles?

[pehar1]

Additional information :

• Happens with completely fresh install (empty .config) and fresh imported image (no sidecar)
• Seems to be independent of the image used
• Happens with 5.5.0+303~gb075af99fd, 5.5.0+295~gc3debd7c8e, 5.5.0+261~g8b6ade32f6
• Does not happen with 5.5.0+153~g1435feaec3
• Terminal output is always the same :

with Sony Alpha 77 ( filmic )

   157.2323 cannot get iop-order for rasterfile instance 0
   157.2323 cannot get iop-order for agx instance 0
   157.2325 cannot get iop-order for rasterfile instance 0
   157.2325 cannot get iop-order for agx instance 0
   157.2328 cannot get iop-order for rasterfile instance 0
   157.2328 cannot get iop-order for agx instance 0
   157.2331 cannot get iop-order for rasterfile instance 0
   157.2331 cannot get iop-order for agx instance 0

with Sony ILCE 7M4 (filmic)

   161.3929 cannot get iop-order for agx instance 0
   161.3933 cannot get iop-order for agx instance 0
   161.3938 cannot get iop-order for agx instance 0
   161.3942 cannot get iop-order for agx instance 0
   161.3946 cannot get iop-order for agx instance 0

For my understanding the messages come from

int dt_ioppr_get_iop_order(GList *iop_order_list,
                           const char *op_name,
                           const int multi_priority)

in .../src/common/iop_order.c but my understanding of C is far too poor to understand the cause.

[pehar1]

More information :
The messages are not thrown if "keep module order" is unchecked in the styles.

[TurboGit]

@pehar1 : Just to be clear, you've:

• Created a style with a fresh install of master version
• Edited it (without module order)
• You use this style to preview on some image and you get error on console

Is that correct? Or was the style created before AgX was merged?

[pehar1]

Created a style with a fresh install of master version

• No, the styles are old, imported into the fresh install of dt. I attached them for testing puposes in the original post. In this state the messages are thrown.
• Edit the styles taking "keep module order" out
• No messages thrown
• Edit the styles again reactivating "keep module order"
• No messages thrown

You use this style to preview on some image and you get error on console

See screenshots in the original post, the messages are thrown when hovering over the item with the style when the small preview for the popup is calculated

Or was the style created before AgX was merged?

I searched for a reference to the creation date in data.db, but couldn't find anything. As far as I can remember, the Sony Alpha 77 ( filmic ) has been around for years, but has been modified repeatedly. As far as I can remember, the Sony ILCE 7M4 (filmic) was created from the alpha-77 style by making changes.

If the last changes have been made before agx and / or rasterfile ? I don't know.

[TurboGit]

@pehar1 : Ok so I understand what is happening. The old style where created with "module order" set but does not have AgX described (created before AgX was merged). So if you don't update module order you don't have AgX in the style and as the module order described in the style is used (but too old) there is an error reported.

I suppose to fix that we need to ensure that "module order" is updated at some point... But this seems delicate issue.

[pehar1]

Thank you for clarification. So the module order "mismatch" has been introduced with rasterfile and agx. This is visible since some commit between 5.5.0+153~g1435feaec3 and 5.5.0+261~g8b6ade32f6 making the logging more "sensitive".

seems delicate issue

As workaround I updated manually