Pass in own sorting / tie breaking function

[aphelionz]

From: ipfs-inactive/dynamic-data-and-capabilities#50

Remove lamport clocks as the implicit tie breaker and instead make it a configuration. The reason is that lamport clocks occupy a lot of space when there are a lot of replicas, which is unnecessary in some use-cases, such as the one in Discussify

[aphelionz]

cc @satazor

[aphelionz]

@satazor Which issues will you be tackling first? I'm going to be working on these as well, starting with this one.

[satazor]

In parallel to all these issues, I’m leading an identity effort on the IPFS DDC WG as well as some other stuff such as recording a video about the status of discussify and preparing a workshop that I will be giving next week. This to say that I will be doing a lot of stuff “concurrently”.

I need to first analyze the codebase and ideally I would pick an easier one first. Which one do you think it’s the best?

[aphelionz]

Great question.

What about one of these two:

• IPLD Support #200
• Memory Footprint #203

[satazor]

#200 seems easy, but it will be a breaking change. Is it a problem to start with this one?

[aphelionz]

Not a problem at all. Personally I'm excited to see what comes of it w/r/t CBOR.

[aphelionz]

@satazor I have this working in theory. Do you have an example of a function you'd want to pass in instead of the default LastWriteWins to write some unit tests with?

[haadcode]

@aphelionz I believe @satazor wanted something like described in ipfs-inactive/dynamic-data-and-capabilities#50 (comment). As to a specific function, dunno exactly what to compare :)

[aphelionz]

I just made it the last argument of the Log constructor. I'll just make a PR now and y'all can have at it :)

[aphelionz]

@satazor @haadcode #209