Juliaup launcher Crash Report

[marcom]

I previously added a comment to #732 (comment), but I opened a new issue as the error message is different.

Running juliaup or julialauncher results in a crash. I think it is due to file locking not working on the filesystem (NFSv3 at the university, NLM seems to have stopped working). This is on Linux. I think something on our network config was changed by the admins, resulting in NLM not working anymore, which results in file locking failing on the NFS mounted dirs (where I have installed julia/juliaup).

A possible workaround would be to allow an environment var to change the lock file location as described in #844 .

"name" = "Juliaup launcher"
"operating_system" = "Linux [64-bit]"
"crate_version" = "1.14.5"
"explanation" = """
Panic occurred in file 'src/config_file.rs' at line 166
"""
"cause" = "called `Result::unwrap()` on an `Err` value: Os { code: 37, kind: Uncategorized, message: \"No locks available\" }"
"method" = "Panic"
"backtrace" = """

   0: 0x5565c6ac5d60 - juliaup::config_file::load_config_db::h7e9d39e2d7499445
   1: 0x5565c6a33372 - julialauncher::run_app::hb135a07abdb38f63
   2: 0x5565c6a3ae9b - julialauncher::main::h865068469b405a8f
   3: 0x5565c6a2a8d3 - std::sys_common::backtrace::__rust_begin_short_backtrace::hbcdfd77d36b49889
   4: 0x5565c6a2b261 - std::rt::lang_start::h8731c304bc39dbb6
   5: 0x7fcd3a0281b0 - __libc_start_call_main
   6: 0x7fcd3a028279 - __libc_start_main@GLIBC_2.2.5
   7: 0x5565c6a2a699 - _start
   8:        0x0 - <unresolved>"""

I used this little python program to confirm that file locking doesn't work on the NFS mounted home dir:

#!/usr/bin/env python3

import argparse
import fcntl
import sys
import time

def test_lock(file_path):
    with open(file_path, 'w') as f:
        print(f"Attempting to lock the file: {file_path}")
        try:
            fcntl.flock(f, fcntl.LOCK_EX | fcntl.LOCK_NB)
            print("Lock acquired. Holding for 5 seconds...")
            time.sleep(5)
            print("Releasing lock.")
            fcntl.flock(f, fcntl.LOCK_UN)
        except IOError as e:
            print(f"Error: Unable to acquire the lock on {file_path}. Error: {e}")
            return 1
    return 0

def main(argv):
    parser = argparse.ArgumentParser(description='Test file locking')
    parser.add_argument('file_path', type=str, help='Path to the file for locking test.')
    args = parser.parse_args(argv[1:])  # skip the script name in argv[0]

    return test_lock(args.file_path)

if __name__ == '__main__':
    sys.exit(main(sys.argv))

It works on /tmp:

$ ~/bin/locktest.py /tmp/foo
Attempting to lock the file: /tmp/foo
Lock acquired. Holding for 5 seconds...
Releasing lock.

But fails on the home dir (mounted via NFS):

$ ~/bin/locktest.py ~/foo
Attempting to lock the file: ~/foo
Error: Unable to acquire the lock on ~/foo. Error: [Errno 37] No locks available