Fix ``Text File Busy`` errors at the source

import fcntl

from galaxy.util import unicodify

def fork_safe_write(path: str, contents: str):

    # The following write method looks a little funky and is inspired by https://twitter.com/_monoid/status/1317895053122150400.

    # This should guarantee that we wait until all forks that inherit open FDs have proceeded to execve,

    # because only then the shared lock can be acquired.

    # This **should** entirely avoid the "Text File Busy" error that `_handle_script_integrity` attempts to deal with.

    # The likelihood of "Text File Busy" happening increases as the load increases, and more work has to be done

    # when forking to copy memory pages, making the fork slower and therefore more likely to happen while the script

    # file is open for writing.

    with open(path, "w", encoding="utf-8") as f:

        f.write(unicodify(contents))

        fcntl.flock(f.fileno(), fcntl.LOCK_EX)

    with open(path) as f:

        fcntl.flock(f.fileno(), fcntl.LOCK_SH)

    unicodify,

)

from galaxy.util.resources import resource_string

from ..fork_safe_write import fork_safe_write

log = logging.getLogger(__name__)

DEFAULT_SHELL = "/bin/bash"

    dir = os.path.dirname(path)

    if not os.path.exists(dir):

        os.makedirs(dir)

    with open(path, "w", encoding="utf-8") as f:

        f.write(unicodify(contents))

    fork_safe_write(path, contents)

    os.chmod(path, mode)

    if job_io.check_job_script_integrity:

        assert job_io.check_job_script_integrity_count is not None