Introduce canonical pointer PDA #374
Conversation
| pub(crate) fn get_wrapped_mint_address_with_seed( | ||
| unwrapped_mint: &Pubkey, | ||
| wrapped_token_program_id: &Pubkey, | ||
| ) -> (Pubkey, u8) { | ||
| get_wrapped_mint_address_with_seed_for_program(unwrapped_mint, wrapped_token_program_id, &id()) | ||
| } | ||
|
|
||
| pub(crate) fn get_wrapped_mint_address_with_seed_for_program( | ||
| unwrapped_mint: &Pubkey, | ||
| wrapped_token_program_id: &Pubkey, | ||
| program_id: &Pubkey, | ||
| ) -> (Pubkey, u8) { | ||
| Pubkey::find_program_address( | ||
| &get_wrapped_mint_seeds(unwrapped_mint, wrapped_token_program_id), | ||
| &id(), | ||
| program_id, | ||
| ) | ||
| } |
There was a problem hiding this comment.
Adding a _for_program() variants to allow for these forks to utilize our package for their address lookups
There was a problem hiding this comment.
Tbh I'd just have them all just take program_id and not support duplicate versions of all of the helpers.
If this is easier for SemVer, though, nbd.
There was a problem hiding this comment.
id() is so deeply embedded into the program+clients that requiring the program_id is a far reaching change. I think I'll keep this forward compatible for now and revisit for a later breaking change if we find lots of folks forking and wanting the simpler name.
buffalojoec
left a comment
buffalojoec
left a comment
There was a problem hiding this comment.
Lgtm! A few nits, take them or leave them!
| pub(crate) fn get_wrapped_mint_address_with_seed( | ||
| unwrapped_mint: &Pubkey, | ||
| wrapped_token_program_id: &Pubkey, | ||
| ) -> (Pubkey, u8) { | ||
| get_wrapped_mint_address_with_seed_for_program(unwrapped_mint, wrapped_token_program_id, &id()) | ||
| } | ||
|
|
||
| pub(crate) fn get_wrapped_mint_address_with_seed_for_program( | ||
| unwrapped_mint: &Pubkey, | ||
| wrapped_token_program_id: &Pubkey, | ||
| program_id: &Pubkey, | ||
| ) -> (Pubkey, u8) { | ||
| Pubkey::find_program_address( | ||
| &get_wrapped_mint_seeds(unwrapped_mint, wrapped_token_program_id), | ||
| &id(), | ||
| program_id, | ||
| ) | ||
| } |
There was a problem hiding this comment.
Tbh I'd just have them all just take program_id and not support duplicate versions of all of the helpers.
If this is easier for SemVer, though, nbd.
program/src/state.rs
Outdated
| /// The mint authority of an unwrapped mint's on-chain signal another token-wrap | ||
| /// deployment is the "canonical" one. |
There was a problem hiding this comment.
nit: this comment is a little bit confusing. Do you think we can clarify it a bit?
| return Err(ProgramError::MissingRequiredSignature); | ||
| } | ||
|
|
||
| let mint_data = unwrapped_mint_info.try_borrow_data()?; |
There was a problem hiding this comment.
Should we be checking the mint's owner here? Maybe it's inconsequential, since the mint address can't be duplicated.
There was a problem hiding this comment.
Sure thing, will add a check+test
Integrating feedback from Foundation
Problem
It's likely that a few high profile mint authorities will not consider the original deployment of token-wrap the one they want folks to use for their mint. They may prefer an alternative configuration (especially for token-2022s) and will fork+deploy their own. These authorities need a way to signal on-chain that a different token wrap deployment is the "canonical" one for their mint.
Solution
This PR adds a
CanonicalPointerPDA which allows mint authorities to set a field with a pubkey to the alternative deployment. One instruction is added to create/update this PDA.Edge cases
If there is no authority, a pointer cannot be created. If one is created and later the mint authority is set to None, the canonical pointer state is frozen as well.