dt 4.6.1 just built fails with segmentation fault.

Software darktable
1

I have now built 4.6.1 only using the source files from darktable.org. Build completed without any errors (but some notifications of missing items). Starting darktable from the command line ( /opt/darktable/bin/darktable) results in dt closing with a segmentation fault. A backtrace is available.

Given my history of the last few days of failing to build because of user errors and annoying too many people with requests for help, I now am concerned as to what to do with this back-trace data. do I raise another, potentially spurious, bug report or build again ?

2

You can use the same thread for all these issues, you Noe have four or five separate ones in the last two days, its getting hard to follow.

1 Like
3

It was something I thought about seriously. I concluded that this one addressed a different topic, concerned with what appears to be a genuine defect rather than the result of a user error, and so warranted a new post.

But I take your point, bearing in mind your level of understanding of dt in the round. Apologies that my judgement was not on target.

4

No need to apologize, Iā€™m just suggesting what I think will get you the best quality of help.

5

is there something I can do for myself by examining the output from darktable in the terminal during both the build and the run processes? Might I be able to get some indication of a user error in setting up or executing the build?So far the only errors which appear to me to be significant are:

  • missing jsonschema
  • couldnā€™t find configuration file for package OpenEXR compatible with version 3.0
  • missing optional package IMath version at least 3.1
  • missing package JXL, version at least 0.7.0
  • missing Cups package (synaptic begs to differ, telling Cups is installed)
6

Optional dependencies, and Cups misses development files (same as the other times you were told this)

Otherwise, post complete information. Small excerpts just showing part of the output can miss information or simply be misleading (e.g. that line with ā€œusr/local/include/lensfunā€¦ā€ would have helped if it was shown the first time you reported an error).

And if you have a backtrace, that may tell you where the error occurred. But, given the results from the other thread, I have to wonder if you did unpack the downloaded source code in its own directory, or if you unpacked it over the Git version. If the latter, you are bound to get problemsā€¦ (iow, donā€™t do that)

7

Thanks so much for your continued advice and support. If you have the time to read my apologetic response to member kmilos on my ā€˜fatal error in buildā€™ post, you will see that the whole problem of build failure and run failure has now been solved. And that you too had already seen and advised me about the problem - of old liblensfun files in /usr/local/lib.

Sorry, but I just lacked enough understanding to appreciate what you were telling me, so I didnā€™t check (via synaptic) where the files, which the correct version of liblensfun installs, go - I now realise that none go into /usr/local/lib.

Similarly I didnā€™t understand what you were telling me about Cups. But I can assure you that I did unpack the downloaded source code into its own directory - not named ;darktableā€™, which made it impossible to mix with the Git version (which I had removed by deleting the 'darktable; directory anyway).

So, there we have it: Iā€™m up and running on 4.6.1, days later than I might have been if I had fully understood the implications of your advice. Thanks for your patience.