This leads to confusion, and redundant work on lotus side as it needs to manage that file between PC1/2.
Ideally Precommit1 would either not create the sealed file at all - that would happen in early PC2, or the sealed file would actually contain the correct, sealed data (in which case PC2 could possibly not require unsealed data access?)
Today precommit1 just copies unsealed bytes to the 'sealed' file, and does nothing else with them -> https://github.com/filecoin-project/rust-fil-proofs/blob/128f7209ec583e023f04630102ef1dd17fbe2370/filecoin-proofs/src/api/seal.rs#L91-L98
This leads to confusion, and redundant work on lotus side as it needs to manage that file between PC1/2.
Ideally Precommit1 would either not create the sealed file at all - that would happen in early PC2, or the sealed file would actually contain the correct, sealed data (in which case PC2 could possibly not require unsealed data access?)