Prevent concurrent lockfile conflicts using flock and add backup

[0xZaddyy]

Summary

This PR introduces a safe and stable mechanism to manage Cargo.lock during testing across multiple scripts.

• File Locking with flock: Ensures only one script can manipulate Cargo.lock at a time using a .cargo.lock.flock file.
• Backup and Restore: test.sh creates a backup of the existing Cargo.lock before overwriting it and restores it afterward using an EXIT trap.

This prevents race conditions and corruption when running tests concurrently

closes #773

[coveralls]

Pull Request Test Coverage Report for Build 16203386079

Details

• 0 of 0 changed or added relevant lines in 0 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage remained the same at 85.558%

---

Totals
Change from base Build 16201907810:	0.0%
Covered Lines:	7571
Relevant Lines:	8849

---

💛 - Coveralls

[zealsham]

Do you really intend to check for the literal string "LOCKFILE" ? . if you meant the variable , then you forgot the "$" character

[0xZaddyy]

Thanks for pointing it out!

[nothingmuch]

not sure what this is trying to do, but exec is definitely the wrong thing here, it replaces the current shell script with the program 9 which should not normally exist

Suggested change

	exec 9>"$LOCKFILE_LOCK"
	touch "$LOCKFILE_LOCK"

[nothingmuch]

this would need to be something like:

(
    flock 9
    ...
) 9>"$LOCKFILE_LOCK"

() 9>"$LOCKFILE_LOCK" creates a subshell, opening $LOCKFILE_LOCK for writing as file descriptor 9 and then forking the current program

within the subshell, flock can then inherit file descriptor 9 from the subshell, creating the locking behavior

the ... represents the entirety of the test script, i.e. all the commands that need to be protected by the lock

[nothingmuch]

Suggested change

	trap 'if [ -f "$LOCKFILE_BAK" ]; then mv "$LOCKFILE_BAK" "$LOCKFILE"; fi; flock -u 9' EXIT
	trap '[ -f "$LOCKFILE_BAK" ] && mv "$LOCKFILE_BAK" "$LOCKFILE"' EXIT

since the file descriptor will be closed after the subshell exits, there's no need to unlock

make sure the trap command is within the ( ... ) subshell, or there will be a race condition and the lock will be dropped before restoring

[nothingmuch]

this seems a bit more intuitive to me but doesn't really matter:

Suggested change

	cp "$LOCKFILE" "$LOCKFILE_BAK"
	mv "$LOCKFILE" "$LOCKFILE_BAK"

[nothingmuch]

i as a little confused when no Cargo.lock.bak was present because it's cargo.lock.bak

other than that, tACK, tested concurrently, including interrupting

[nothingmuch]

Suggested change

	LOCKFILE_BAK="cargo.lock.bak"
	LOCKFILE_LOCK=".cargo.lock.flock"
	LOCKFILE_BAK="Cargo.lock.bak"
	LOCKFILE_LOCK=".Cargo.lock.flock"