Open Zebradil opened 5 months ago
Currently, race conditions (at least the ones I identified) are avoided by using mutex locks. This feels like a workaround. Consider rewriting the flow of the sync stage to another approach (channels, workers, task deduplication, etc.)
Currently, race conditions (at least the ones I identified) are avoided by using mutex locks. This feels like a workaround. Consider rewriting the flow of the sync stage to another approach (channels, workers, task deduplication, etc.)