As @anon41087856 notes he would need more information. On windows there should be a backtrace file in users\username\appdata\local\temp darktable_bt_2HNCY0.txt (11.2 KB)
As an example this is is a recent one I had from intermittent crashes happening on my system with a build from yesterday.
If you pass this along it will be helpful…
I tried to run darktable with the -d all option and this is the log that i get darktable-log.txt (1.0 MB)
I can’t find any backtrace file… i have searched in the suggested folder and in other folders but I cannot find it … if you give me some indication I can search again.
Some useful info:
Darktable 3.4.1
OS: Windows 10
OpenCL: enabled
Intel UHD Graphics 630 + Nvidia GeForce GTX 1050 Ti Max-Q
Try here
C:\Users\[username]\AppData\Local\Microsoft\Windows\INetCache\darktable\darktable-log.txt
You are not alone. Same for the picker in color calibration.
I keep several installations of darktable on my external drive, 3.4.0 does certainly not have this issue.
Mine was here: (Windows 10)
C:\Users\USER-ID\AppData\Local\Temp\darktable_bt*******.txt
I repeated several times all the following steps this morning:
@anon41087856 the dialog indicating the path of the backtrace file appeared only once ( i don’t know why it appeared only once)… this is the darktable_bt_file darktable_bt_QTJ0X0.txt (12.5 KB)
hope this will help.
Ale
I’m more one running Darktable 3.4.1 in Windows 10 and having this auto-picker problem.
The backtrace shows an error with a memory buffer freeing. Without further information, that leaves us with only 3467 possible causes. \o/
Is anybody able to launch darktable Windows within gdb ?
On a more global note, it’s useless to post bugs here. Most devs don’t hangout here, it’s not adapted to issues tracking, and tagging me everywhere filmic is mentionned is only adding more spam in my inbox (especially since filmic has nothing to do with it).
Please use Github.
especially since filmic has nothing to do with it
It’s seems to be the picker function. It effects also other modules.
If you have no copy of 3.4.0 - it’s hier, at the bottom: Release darktable 3.4.0 released · darktable-org/darktable · GitHub
Is this bug regression only in the Windows version of 3.4.1? I am on Windows. I suppose this is something that will not be fixed until the next release scheduled for Christmas 2021?
I am pretty sure there will be some release earlier.
IMHO they should delete the Windows 3.4.1 ASAP, since it’s basically unusable. 3.4.0 should stay the current Windows version for now. A dt newcomer on Windows may come to the conclusion that dt is bug ridden and not return for years.
I will stick with 3.4.0 for the time being.
The problem is the chronic lack of devs/testers in Windows, so unfortunately the releases for that OS are more prone to undetected bugs.
Most of the issues with 3.4.0 and 3.4.1 have been already fixed in the 3.5 development relase, so you could use a nightly build or learn to build DT yourself, it’s not difficult. Remember to alway make a backup of database before upgrading.
And if someone wants to help with testing of new releases, that’s more than welcome.
When installing 3.5 it asks to agree to convert the DB. Does it use a different DB format? I didn’t see any improvements in 3.4.1 that effect me, I can use 3.4.0 for now (and still be happy).
The Windows specific crash with color picker has been reproduced and the issue found, thanks to Murtuza Chhil for the quick testing.
A windows only 3.4.1.1 relaae has been made, grab it on github.
There is a list of bug fixes. How do I find more detail about what has been fixed? In particular, I am interested in finding out more detail about what this means:
Fix Lr metadata import, this is done only if no other XMP present.
Thank you.
You may want to browse the commits:
Find the commit for this and if there is an issue linked looked at the discussion on this issue.