Harden Lambda scan temp-file and timeout handling

[docwho2]

Motivation

• Prevent file name collisions in /tmp when multiple scans use the same S3 object base name by using unique temp file names.
• Avoid passing a negative or too-small timeout to Process.waitFor which can cause unexpected behavior near Lambda timeouts.
• Improve correctness on interruption by restoring the thread interrupt status when InterruptedException is caught.

Description

• Replace deterministic /tmp path generation with a new helper createTempFilePath(String) that generates a UUID-based filename (preserving the original extension) and returns a Path under /tmp (file: lambda/src/main/java/cloud/cleo/clamav/lambda/ScanningLambda.java).
• Add getClamScanWaitMillis(int) to compute a safe wait time (remainingMillis - 10000) and bail out if there is insufficient execution time, destroying the clamscan process and tagging the object ERROR when configured to do so.
• Restore thread interrupt status inside the catch for InterruptedException to avoid swallowing the interrupt.
• Add Paths import and ensure the temp file is deleted in the finally block after scanning.

Testing

• Ran the project's test suite with mvn test -q and the tests completed successfully.

---

Codex Task