Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Crash in darkroom #16841

Open
kofa73 opened this issue May 20, 2024 · 11 comments
Open

Crash in darkroom #16841

kofa73 opened this issue May 20, 2024 · 11 comments

Comments

@kofa73
Copy link
Contributor

kofa73 commented May 20, 2024

Describe the bug

I was working in the darkroom (either adjusting the perspective correction module or the contrast in filmic rgb), when darktable crashed. The backtrace file was empty; all I have is the console log from darktable -d pipe (that was the option I used the log a previous crash).
Today I edited many other files the same way, from the same camera, without a crash, so I think it won't be deterministic.

Steps to reproduce

Get the raw and the sidecar from https://tech.kovacs-telekes.org/files/2024-05-20-darktable-crash/

Expected behavior

No response

Logfile | Screenshot | Screencast

console.log

Where did you obtain darktable from?

self-compiled

darktable version

4.7.0+1257~g156a42d33a

What OS are you using?

Linux

What is the version of your OS?

Ubuntu 23.10

Describe your system?

Ryzen 5 5600X

Are you using OpenCL GPU in darktable?

Yes

If yes, what is the GPU card and driver?

NVIDIA GeForce GTX 1060 6GB; 535.171.04

@wpferguson
Copy link
Member

Couldn't reproduce with darktable 4.7.0+1271~gf98eda41f0 on 22.04 (gcc 12.3, nvidia 3070)

@jenshannoschwalm
Copy link
Collaborator

In the provided log we miss the "got details mask" ... could be the buffer data is wrong in this case or scaling/distorting of the mask failed. BTW no problem here so far.

@kofa73
Copy link
Contributor Author

kofa73 commented May 21, 2024

Thanks for looking into this.

  • This is the 2nd time my backtrace is empty. What could I do to prevent that?
  • What -d flags would provide the most useful output?

@jenshannoschwalm
Copy link
Collaborator

Likely -d pipe -d verbose -d opencl

@ralfbrown
Copy link
Collaborator

Given the choice between perspective and filmic, the fact that the pipe was running in fast mode indicates you were in perspective correction.

@jenshannoschwalm
Copy link
Collaborator

Yes, the perspective correction module was open/active. Still the processing was in diffuse and as some lines above would require the detail mask. This point was not reached so one idea would be the distorting of the detail mask somehow failed, the single modules distorting it including their roi would be shown via the -d verbose option

@kofa73
Copy link
Contributor Author

kofa73 commented May 22, 2024

Could it be related that if rotate and perspective is active, things like filmic's automatic exposure offset selector brings up a rectangle that is offset from the image edges? E.g. (this is another image, but I saw this for all that I recently edited):
image

If I disable crop, it looks like this:
image

@jenshannoschwalm
Copy link
Collaborator

Sorry, don't understand Your question

@kofa73
Copy link
Contributor Author

kofa73 commented May 23, 2024

Since it seems the filmic auto-picker area is off (it does not seem to handle the distortion introduced by rotate and perspective), maybe there's some other code that also does not handle it, leading to the crash.

@jenshannoschwalm
Copy link
Collaborator

Ah, got it. Likely the data calculated in filmic are not "best" but that should not be a problem.

Took you image and could not reproduce at all, would love a -d pipe -d verbose log :-)

@kofa73
Copy link
Contributor Author

kofa73 commented May 23, 2024

I'll share the logs when I encounter the next crash.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

4 participants