p-token: Add batch instruction
#35
Conversation
febo
force-pushed
the
febo/fix-jon-comments
branch
from
e058e93 to
c12b54f
Compare
| let (discriminator, instruction_data) = instruction_data | ||
| .split_first() | ||
| .ok_or(ProgramError::InvalidInstructionData)?; | ||
| let instruction = TokenInstruction::try_from(*discriminator)?; |
There was a problem hiding this comment.
Was the usage of TokenInstruction supposed to be reverted?
There was a problem hiding this comment.
Yes, using the TokenInstruction enum increases 20 CUs at least once I added the Batch variant. I could not find a way to make it efficient.
| /// - `0`: `InitializeMint` | ||
| /// - `1`: `InitializeAccount` | ||
| /// - `3`: `Transfer` | ||
| /// - `7`: `MintTo` | ||
| /// - `9`: `CloseAccount` | ||
| /// - `16`: `InitializeAccount2` |
There was a problem hiding this comment.
Just to make sure, were these updated for better performance as a result of this PR?
There was a problem hiding this comment.
Yes, the idea was to keep all the Initialize*, MintTo, CloseAccount and Transfer on the "faster" path. We can tweak the list if you thing there are other instructions that are more critical.
| accounts: &[AccountInfo], | ||
| instruction_data: &[u8], | ||
| instruction: TokenInstruction, |
There was a problem hiding this comment.
Same here: was this supposed to be reverted?
There was a problem hiding this comment.
Yes, it is more efficient to use the u8 value directly.
This PR adds the
batchinstruction forp-token– this was originally proposed here.The purpose is to allow the execution of multiple token instructions in a single program instruction – e.g., it is possible to execute multiple instructions in a single CPI call, which saves at least
1000CUs (CPI base fee) for each additional instruction.Note
SInce tests now run without
test-sbf, this PR also updates thep-tokentests to remove thetest-sbffeature.