Upgraded Darktable 4.2.1 win64, now DT just quits

I just ran the darktable-4.2.1-win64.exe installer, without uninstalling prior (not sure of old version, but at best a year old) and now Darktable just quits when being ran independently, or when trying to open a .CR2 via right-click “Open With” → Darktable

When running the app independently, it flashes what looks to be an intro screen quickly before quitting: https://i.imgur.com/lDlI91h.png

i then uninstalled 4.2.1 and re-installed it, and it’s still doing the same thing.

… darktable ran fine prior to this upgrade, i use it pretty much constantly. now what?

You can launch from the terminal and put the output here.

C:\path\to\darktable.exe -d all

I just ran the uninstaller again, which was successful, but then i noticed the C:\Program Files\darktable folder was still there, with around 80mb of folders/files still in it, i deleted that folder (empty recycle bin) then re-ran the installer, it appears to be working now

I checked event log it wasn’t an app crash when it was quitting, something internal going on it would seem - everything OK now thank you

It has to upgrade the database and config files when using a new version for the first time.

i did see a message regarding that in the beginning and waited a while… but then nothing… i’m running a 3700x(not OC)/32GB@3600/EVO M.2/GTX 1080 - can’t imagine it’d take … several minutes without anything coming up? - anyways things seem OK now

Hard to say without more detail :upside_down_face:

Prior to this upgrade, i don’t think i’d ever had Darktable crash (close unexpectedly) a single time, and i’ve been using it frequently for at least 2 years now.

Since this upgrade, i’ve had it crash (close unexpectedly) several times when processing images. Nothing logged in Windows Event Viewer.

Because Darktable doesn’t save changes as you go along, all changes are lost when this occurs, potentially 20-30min into processing.

I’ve noticed if i simply visit Lighttable and re-load the image i’m currently working on it saves the steps I’d previously done - So i’m going to recommend either Darktable save steps as you’re processing, or give the user a save button? - (maybe that already exists and i haven’t noticed?)

This was occurring without a lot of predictability until recently. Generally each time this occurs it’s not specific to a module i’m using, but i’ve been able to “cause” this to occur repeatedly on a particular image within some steps of using Graduated Density, although again, i don’t think it has anything to do with that Module as i believe I was using other modules when this crash occurred when processing other images.

After googling i’d seen suggestions to update the display driver, which honestly i was reluctant to do as i can run any graphics or video editing program, old or new FPS game, flawlessly, there is nothing going on with this machine at all. Despite that i upgraded my (NVIDIA) drivers to 531.18, no difference, i then did a clean install of 531.18, no difference. I re-booted between each driver install.

If you are to run cmd/command from anywhere the syntax for an assumed default-installation location would be:

“C:\Program Files\darktable\bin\darktable.exe” -d all

After that and some googling, it appears the log file is @ C:\Users\\AppData\Local\Microsoft\Windows\INetCache\darktable\darktable-log.txt

I have a clean log isolated to the minimum steps involved to cause this to re-occur, after viewing the file it looks like i’d need to redact it a bit but the last few lines are consistent regardless, ending in the “Access violation.”

81.394313 [opencl memory] device 0: 14838720 bytes (14.2 MB) in use
81.394416 [dev_pixelpipe] took 0.005 secs (0.000 CPU) [full] processed base curve' on GPU, blended on GPU [dev_pixelpipe] module input color profile’ min: (1.066164; -47.160866; -33.037830) max: (110.451675; 39.998413; 488.294708) [preview]
81.396183 [process CPU] preview nlmeans. IN ( 0/ 0) 873x 573 scale=1.00. OUT ( 0/ 0) 873x 573 scale=1.00, final 873x573, backbuf 873x573
Magick: caught exception 0xC0000005 “Access violation”…Magick: caught exception 0xC0000005 “Access violation”…

If you move to the next image or back to lighttable (from darkroom), it will save the edits. Until you discover the crash issues, I suggest moving to the next image after 10min of editing so you dont loose your work.

When you upgrade darktable, it keeps the configuration files available. I’ve read suggestions to start a new config to avoid crashes and Ive seen reports of success with that method. I’m interested in discovering what the real issue is. The config directory is here: C:\Users\<username>\AppData\Local\darktable

If you turn off open CL, does it still crash?

Under the Darktable preferences/processing i unchecked [ ] “Activate OpenCL support” - closed/re-opened Darktable and went through a nearly identical process which did not change the outcome, Darktable still unexpectedly closes

The log file does look slightly different leading up to ending in the same “Access violation” message:

56.429792 [modify roi IN] full graduatednd. ( 0/ 0) 1189x 780 scale=0.27 → ( 0/ 0) 1189x 780 scale=0.27
56.429838 [process CPU] full graduatednd. IN ( 0/ 0) 1189x 780 scale=0.27. OUT ( 0/ 0) 1189x 780 scale=0.27, final 1189x780, backbuf 1189x780
56.431792 [dev_pixelpipe] took 0.002 secs (0.000 CPU) [full] processed graduated density 1' on CPU, blended on CPU [dev_pixelpipe] module base curve’ min: (0.000758; 0.000000; 0.000000) max: (1.349730; 1.000671; 0.833403) [preview]
56.432773 [process CPU] preview hazeremoval. IN ( 0/ 0) 873x 573 scale=1.00. OUT ( 0/ 0) 873x 573 scale=1.00, final 873x573, backbuf 873x573
[dev_pixelpipe] module graduated density 1' min: (0.000249; 0.000000; 0.000000) max: (2.225230; 1.762949; 1.352550) [full] 56.443390 [process CPU] full basecurve. IN ( 0/ 0) 1189x 780 scale=0.27. OUT ( 0/ 0) 1189x 780 scale=0.27, final 1189x780, backbuf 1189x780 56.445741 [dev_pixelpipe] took 0.002 secs (0.000 CPU) [full] processed base curve’ on CPU, blended on CPU
56.451112 [dev_pixelpipe] took 0.018 secs (0.047 CPU) [preview] processed haze removal' on CPU, blended on CPU [dev_pixelpipe] module haze removal’ min: (0.000152; 0.000000; 0.000000) max: (1.421041; 1.021479; 0.854791) [preview]
56.456399 [process CPU] preview colorin. IN ( 0/ 0) 873x 573 scale=1.00. OUT ( 0/ 0) 873x 573 scale=1.00, final 873x573, backbuf 873x573
[dev_pixelpipe] module base curve' min: (0.000269; 0.000000; 0.000000) max: (1.422194; 0.991846; 0.812522) [full] 56.456940 [process CPU] full hazeremoval. IN ( 0/ 0) 1189x 780 scale=0.27. OUT ( 0/ 0) 1189x 780 scale=0.27, final 1189x780, backbuf 1189x780 56.457621 [dev_pixelpipe] took 0.001 secs (0.000 CPU) [preview] processed input color profile’ on CPU, blended on CPU
[dev_pixelpipe] module `input color profile’ min: (0.940598; -44.352142; -30.340349) max: (110.260414; 39.207397; 445.159302) [preview]
56.464426 [process CPU] preview nlmeans. IN ( 0/ 0) 873x 573 scale=1.00. OUT ( 0/ 0) 873x 573 scale=1.00, final 873x573, backbuf 873x573
Magick: caught exception 0xC0000005 “Access violation”…

  • snips of opencl related info from the beginning section of log:

[dt_get_sysresource_level] switched to 1 as `default’
total mem: 32694MB
mipmap cache: 4086MB
available mem: 16347MB
singlebuff: 255MB
OpenCL tune mem: OFF
OpenCL pinned: OFF

[opencl_init] opencl related configuration options:
[opencl_init] opencl: OFF
[opencl_init] opencl_scheduling_profile: ‘default’
[opencl_init] opencl_library: ‘default path’
[opencl_init] opencl_device_priority: ‘/!0,///!0,*’
[opencl_init] opencl_mandatory_timeout: 200
[opencl_init] opencl library ‘OpenCL.dll’ found on your system and loaded
[opencl_init] found 1 platform
[opencl_init] found 1 device

[dt_opencl_device_init]
DEVICE: 0: ‘NVIDIA GeForce GTX 1080’
CANONICAL NAME: nvidiageforcegtx1080
PLATFORM NAME & VENDOR: NVIDIA CUDA, NVIDIA Corporation
DRIVER VERSION: 531.18
DEVICE VERSION: OpenCL 3.0 CUDA, SM_20 SUPPORT
DEVICE_TYPE: GPU
GLOBAL MEM SIZE: 8192 MB
MAX MEM ALLOC: 2048 MB
MAX IMAGE SIZE: 16384 x 32768
MAX WORK GROUP SIZE: 1024
MAX WORK ITEM DIMENSIONS: 3
MAX WORK ITEM SIZES: [ 1024 1024 64 ]
ASYNC PIXELPIPE: NO
PINNED MEMORY TRANSFER: NO
MEMORY TUNING: NO
FORCED HEADROOM: 400
AVOID ATOMICS: NO
MICRO NAP: 250
ROUNDUP WIDTH: 16
ROUNDUP HEIGHT: 16
CHECK EVENT HANDLES: 128
PERFORMANCE: 3.407
TILING ADVANTAGE: 0.000
DEFAULT DEVICE: NO
KERNEL DIRECTORY: C:\Program Files\darktable\share\darktable\kernels
CL COMPILER OPTION: -cl-fast-relaxed-math
KERNEL LOADING TIME: 0.0244 sec
[opencl_init] OpenCL successfully initialized.
[opencl_init] here are the internal numbers and names of OpenCL devices available to darktable:
[opencl_init] 0 ‘NVIDIA GeForce GTX 1080’
[opencl_init] FINALLY: opencl is AVAILABLE on this system.
[opencl_init] initial status of opencl enabled flag is OFF.
[dt_opencl_update_priorities] these are your device priorities:
[dt_opencl_update_priorities] image preview export thumbs preview2
[dt_opencl_update_priorities] 0 -1 0 0 -1
[dt_opencl_update_priorities] show if opencl use is mandatory for a given pixelpipe:
[dt_opencl_update_priorities] image preview export thumbs preview2
[dt_opencl_update_priorities] 0 0 0 0 0
[opencl_synchronization_timeout] synchronization timeout set to 200
[dt_opencl_update_priorities] these are your device priorities:
[dt_opencl_update_priorities] image preview export thumbs preview2
[dt_opencl_update_priorities] 0 -1 0 0 -1
[dt_opencl_update_priorities] show if opencl use is mandatory for a given pixelpipe:
[dt_opencl_update_priorities] image preview export thumbs preview2
[dt_opencl_update_priorities] 0 0 0 0 0
[opencl_synchronization_timeout] synchronization timeout set to 200
0.330778 [wb_presets] loading wb_presets from `C:\Program Files\darktable\share\darktable/wb_presets.json’

  • it appears there’s no other reference to “opencl” in the log file after the above

  • since it seems unrelated i’ve turned the opencl option back on

i renamed that particular darktable folder to “darktable.old” and moved it out of the AppData folder, initial tests looking positive, folder re-created itself and this may have resolved the issue.

Update: i really tried to “break” it here with multiple instances of the same (Graduated Density) steps/changes, that’d consistently caused Darktable to close, issue seems resolved now.

Assuming that was the resolution, I’ll armchair another 2c here, and i could be completely wrong on this, but i’m not too surprised to see this play out under the AppData folder, with an “Access violation” or denied file/folder error.

Within the last year or so i’ve noticed people on (standard user) restricted profiles lose their ability to modify or remove user-profile installed programs they’d previously installed themselves where no administrative rights were originally required, the most common i’ve seen is Cisco Webex, but I’d also expect Zoom or maybe even Teams, really any downloadable program that didn’t require admin rights to install, and puts itself (or data) under AppData in some way.

Then at some point (and i’m not tracking it) users lost their ability to remove the programs they’d installed on their profiles, and suddenly to remove them admin rights is required. This is abnormal behavior, normally any user profile installed program can be then uninstalled by the user. (same rights in, same rights out) - but it seems Microsoft changed something specifically in this area. I’ve witnessed the above situation (with apps) play out a few times lately.

I guess the only thing i could suggest is re/verifying file rights access upon re/install in order to detect/anticipate the problem, I noticed another user in another thread having strange file “rights” issues recently and maybe something is in common there.

You should be able to move the library.db and data.db back to the new folder to restore your library and presets.

Can you share the darktablerc old and new file?

My .old folder is pretty big, ~300MB

  • Fair amount of files in .old are set to read-only. I see files in the new folder set to read-only too, i’m sure some of that is normal, but someone might want to check into that.

screen shot of old folder: https://i.imgur.com/7cWZ84g.png

screen shot of new folder: https://i.imgur.com/OfCnHtO.png

I’m OK with losing presets etc just started with a new camera, i’ll survive

This is normal. The library.db and all the iterations is what links your library to the file location. By default, darktable will create a backup (snapshot) every so often (weekly?) And whenever there is a significant update in the database structure. You can delete most/all of the old stuff.

Similar with the data.db.

The file I’m interested to compare is the first one, darktablerc.

is there a particular area or lines in that file you’re looking for?

looking within the contents of that file, i can’t help noticing:

plugins/imageio/storage/gphoto/id=642055548087-n01fgvugnbns7a9jq8jfucjsn5l1t6so.apps.googleusercontent.com

plugins/imageio/storage/gphoto/secret=o29QcbsDWS5cauRqdmGdF3sP

anyone share any insight onto what that’s pointing to? - kinda curious

https://www.google.com/search?&q=642055548087

Have you configured a google photos plugin? Looks like it might be an encrypted API key.

Not at all, but you can see it showing up in other peoples darktablerc files if you take a look at that google search. I had the same similar/thought though, where is this going, or who’s account/key is that? did someone compile or provide that from DT?

https://github.com/darktable-org/darktable/blob/master/data/darktableconfig.xml.in

I do not see anything like this in my darktable.rc