We should add a new replay phase that serializes the recording to JSON, re-parses it into memory, then replays that recording as a second replay phase. Otherwise, it could be very hard to find bugs in the serialization code.
Note that it's probably a lot of work to compare the in-memory representations of the original and deserialized recording. The original will likely contain some platform-specific stuff that is not necessary for deterministic execution. We really just care that playing the recording results in the same execution.
<rdar://problem/16480349>
<rdar://problem/16480403>
Closing web replay-related bugs until we resume working on the feature again.