Typed data aggregate (SNIP-12)

[penovicp]

Motivation and Resolution

Aggregate for the reported typed data issues. Each commit addresses a corresponding issue (or PR).

Extracted to provide a unified overview so that potential impact can be evaluated and coordinated.

Issues:

• Enum encoding is different from the SNIP-12 spec #1278
• Incorrect SNIP-12 enum type signature encoding #1286
• SNIP-12 Incorrect encoding for empty enum variant #1341
• SNIP-12 merkletree doesn't include dependent types #1342
• SNIP-12 Incorrect type for enum #1343
• SNIP-12 incorrect encoding for type selector #1348
• Encoding for "shortstring" typed data (SNIP-12) #1039

[amanusk]

Summary of what we currently plan, please ack

@xJonathanLEI @sgc-code @franciszekjob @penovicp @tabaktoni @thiagodeev

#1278 - fix js implementation
#1286 - fix js implementation
#1341 - fix js implementation
#1343 - fix js implementation

#1039 - Don't change implementation, clarify the SNIP

#1342 - Don't change implementation, clarify the snip, have a new revision if the change is necessary
#1348 - Don't change implementation, clarify the snip, have a new revision if the change is necessary

@sgc-code, @penovicp @xJonathanLEI @tabaktoni @franciszekjob

[thiagodeev]

Thanks @amanusk.
Please let us know when these changes are released so I can update Starknet.go code

[sgc-code]

wont this break SNIP-9? https://github.com/starknet-io/SNIPs/blob/main/SNIPS/snip-9.md#version-2

[sgc-code]

#1039 - Don't change implementation, clarify the SNIP
#1342 - Don't change implementation, clarify the snip, have a new revision if the change is necessary
#1348 - Don't change implementation, clarify the snip, have a new revision if the change is necessary

This were actually addressed in this PR

[sgc-code]

I think it's missing the fixes for these two issues: #1287 and #1362

[penovicp]

#1039 - Don't change implementation, clarify the SNIP
#1342 - Don't change implementation, clarify the snip, have a new revision if the change is necessary
#1348 - Don't change implementation, clarify the snip, have a new revision if the change is necessary

This were actually addressed in this PR

Like the previous aggregate, the idea here is to have them all available in one place, from there to evaluate potential impact and pick which will be applied and in what way.

To address the other message here:


• #1287 - It was picked out in a similar approach as previously mentioned and has been included since v6.24.1.
• #1362 - I addressed in a comment there. The SNIP doesn't foresee the usage of the requested type/message format, and a more verbose one should be used.