chains icon indicating copy to clipboard operation
chains copied to clipboard

Published 20 hours ago • verified publisher icon ethereum-lists

Field "chain" is used twice with different meaning/definition

Open kopeboy opened this issue 2 years ago • 6 comments

I would clarify what does "chain" mean, given that we have "parent.chain" that has a different definition. Since the latter is dependent on / specific to this repository and doesn't have a correspondence on a on-chain value (EIPs are off-chain right?), I would consolidate to the former and make a clear definition of what is it. There are irregularities currently on the "chain" field values, examples:

  • Capitalized/UPPERCASE/lowercase (ETH/ETH vs Evmos/evmos)
  • equal to the symbol of the native currency vs. to the name or shortname of the chain (ETH vs Solana vs evmos)
  • is it a unique representation of the parent chain or parent native currency (by slip44 maybe?)? Seems so with children of Ethereum but not otherwise, like Aurora-NEAR, Avalanche-AVAX, etc..

In general, I would add comments to https://github.com/ethereum-lists/chains/blob/master/model/src/main/kotlin/org/ethereum/lists/chains/model/Chain.kt to help listers provide consistent values.

kopeboy avatar Mar 07 '23 12:03 kopeboy

true - this is inconsistent I think we just remove the "chain" in the root and keep it in parent "chain" in the root is kinda redundant to "name" anyway

ligi avatar Mar 07 '23 12:03 ligi

migration path:

  • we remove it in all files from _data
  • the processor adds it again for the legacy API
  • in API v2 it is not added (#1889)

ligi avatar Mar 07 '23 12:03 ligi

Excuse me? How would I index to the parent chain if the value of "parent.chain" is NOT on the parent?! Either you add a field holding the "EIP155-X" like values at the higher level, OR you define the parent.chain by an existing identifier for the chain, like chainID or shortname..

kopeboy avatar Mar 07 '23 13:03 kopeboy

the filename is CAIP2 and the values of chain fields in parent is CAIP2

ligi avatar Mar 07 '23 13:03 ligi

They would need to be "eip155:1" instead of "eip155-1" for CAIP2. Anyway, why would you want someone to parse github folders for file names when you can have everything in a single json that is already the list? Can't we override either "chain" or "parent.chain" with the same chosen id to make the list file sufficient?

kopeboy avatar Mar 07 '23 14:03 kopeboy

from the feedback have you like that we need to do the old behind and welcome new

95HenryFong95 avatar Apr 27 '23 03:04 95HenryFong95