Polochon-street
commented
3 years ago Hi,
Thanks for reporting :)
I don't think the cue files are a problem (I have some on the same kind in my own library, and they don't make the program segfault, just error like what you see here) I also have a 200GiB library, so, it's probably not the problem.
It analyzes as many songs in parallel as you have cores, so perhaps it takes too much RAM, or something?
Otherwise, if you have gdb, could you run it, then in the prompt, run --rescan ~/slsk/music/, and if it segfaults, show me the output of bt ?
You can compile / run https://github.com/Polochon-street/bliss-rs/blob/master/examples/analyse.rs by cloning the repo, and then run cargo run --release --example=analyze /path/to/music/file.flac, maybe that would help finding the file that segfaults (if it's a specific file)
BachoSeven
I did not, it seems to coredump around PinkFloyd, but not always at the same file.
How should I compile just that piece of code?
I think this is due to
blissifytrying to analyse all music files that a.cuefile lists, and in my case many of those are non-existent(wrong/different naming). Or, could it be a problem of sheer size of the collection(?) I do have many hundred gigs of music in that folder.Log: