Choosing between per‑file copy and single‑archive backup for atomic LÖVE saves

Question

Backup goalEnsure a reliable snapshot of the LÖVE save directory despite non‑deterministic enumeration order, possible file changes during enumeration, and a limited number of simultaneous file handles.Constraints and uncertaintyPer‑file copying preserves individual file attributes and allows incremental updates but opens many handles and does not guarantee an atomic snapshot if the source changes while iterating. Packing all files into a single archive yields an atomic write and reduces handle usage, yet requires implementing or bundling an archive format and complicates restoration of individual files. The documentation notes that love.filesystem.enumerate may omit files created or deleted during the call, leaving developers to decide whether to accept possible inconsistency or to retry enumeration until a stable list is obtained.Which approach better satisfies the need for atomicity while respecting handle limits, and how should the enumeration instability be mitigated?

Tasadduq Burney · Answer

Recommended approachUse a single‑archive backup written to a temporary file and then renamed atomically. This guarantees a consistent snapshot even if the save directory changes during the backup, and it only needs one file handle at a time.Why per‑file copy is less suitableEach file opened consumes a handle; many saves can exceed the platform’s simultaneous‑handle limit.The copy is not atomic – a file may be modified after it has been read, leading to an inconsistent backup.Mitigating enumeration instabilityCall love.filesystem.enumerate on the save directory and store the list.After a short delay (e.g., 50 ms), enumerate again.If the two lists match (same files, same order), treat the directory as stable.If they differ, repeat the process up to a configurable retry limit (e.g., 5 attempts) with a brief sleep between attempts.If stability is still not reached, fall back to per‑file copying while checking that the number of concurrently open handles stays below the platform limit.Backup steps (stable case)Obtain a stable file list using the retry loop above.Create a temporary archive file (e.g., backup_tmp.zip) in the same directory as the final backup.For each file in the stable list, open it read‑only, add its contents to the archive, and close the handle immediately.Close the archive handle.Rename the temporary archive to the final backup name (e.g., backup.zip) – the rename is atomic on the underlying filesystem.This procedure uses at most two file handles simultaneously (one for the source file, one for the archive) and produces an atomic snapshot.

Choosing between per‑file copy and single‑archive backup for atomic LÖVE saves

Backup goal

Constraints and uncertainty

1 answer

Recommended approach

Why per‑file copy is less suitable

Mitigating enumeration instability

Backup steps (stable case)

Your perspective belongs here.

0 question comments