RUMM-2177: Use TLV format for data storage

[0xnm]

What does this PR do?

This change aligns the binary format used to write data to the disk with format implemented in DataDog/dd-sdk-ios#841.

For every event received we are going to write meta + event in the TLV (Type-Length-Value) format, where the format itself for each block (meta or event) has the following structure:

    +-  2 bytes -+-   4 bytes  -+- n bytes  -|
   |  block type | data size (n) |    data   |
    +------------+---------------+-----------+

It is safe to merge it without cleaning existing data, because this PR targets SDK v2 branch, where the change later will be done to use SDK instance ID in the feature folder naming, so data in the SDK v2 read/write paths won't overlap with data from existing SDK (but during the current development existing data should be erased from emulator).

Tests are verified against the latest testable commit.

Review checklist (to be filled by reviewers)

• Feature or bugfix MUST have appropriate tests (unit, integration, e2e)
• Make sure you discussed the feature or bugfix with the maintaining team in an Issue
• Make sure each commit and the PR mention the Issue number (cf the CONTRIBUTING doc)

[mariusc83]

LGTM

[maxep]

Nice work here!!

I think we could benefit from adding more flexibility on reading phase. We discussed that we should expect a pair of valid meta+event data but that would be when metadata is present. As we don't have yet needs for metadata, I dont think we should add it now. On iOS, we currently support stream with only events.

In short, when reading a block we would have these conditions:

1. If we encounter an event, keep it
2. If we encounter metadata, read next block and expect an event. Then we can decode both.

[maxep]

👍