Little-endian functions

[fewwwww]

For The Graph TypeScript Library (graph-ts), why all functions related are in little-endian, instead of Ethereum's default big-endian?

[fewwwww]

I was trying to ask the question in the discord. But the technical support engineer told me to open an issue.

[That3Percent]

A typical processor uses little-endian representation for integers because casts are a no-op.. Everything should be in little-endian because otherwise the system needs to convert to big-endian and back (which is extra work for no benefit). The real question is why is Ethereum using big-endian?

[fewwwww]

Thank you for your reply. So in the development of Subgraph using graph-ts, do we need to pay special attention to Ethereum's big-endian?

For example, when using fromSignedBytes (https://github.com/graphprotocol/graph-tooling/blob/dcc246333397952506a9db6a6c556fc57b8017c2/packages/ts/ common/numbers.ts#L87), do I still need to call .reverse(), then pass in the input?

[leoyvens]

Yes, that's correct as the function documentation states. If you have suggestions for helpers we could introduce or possibly new names for these APIs that would reduce confusion, those are welcome.