How do I determine whether an input or output is segwit? - revisited

Question

This is a discussion that has already been had here, but bear with me.

We're gonna be talking about tx faae7e55db14a32e083cbf6a91db8a5ab6a3b05e050d9cefdec80b47f966848e

• all I got is the encoded transaction
• it has been marked as segwit (marker + flag)
• tx has 3 inputs, only the last one is segwit
• the segwit input has a script sig: 1600140c6259927541c4f8e88fc1398691e2661d15591a
• got to decode the witnesses section correctly, implied there is only one

In the previous discussion, they mentioned that a segwit input is easily spotted as it has an empty script sig.

The segwit input we have here has a script sig... I know it's segwit cause I looked in a blockchain explorer. How can my software programatically tell it's segwit just by looking at the transaction's data?

Answer 1

The way the BIP141 segwit rules consider transactions is simply that every transaction input (even non-segwit ones) have a witness stack. Non-segwit inputs are required to have an empty witness stack however.

If every transaction input has an empty witness stack, the old pre-segwit transaction serialization is used for the transaction. If at least one inout has a non-empty witness stack, the BIP144 transaction serialization format (including marker & flag) is used. The non-segwit inputs then simply have an empty witness.

Now, there is a consensus rule that says that unless a transaction input is a segwit spend, its witness stack must be empty. An input is a segwit spend if the UTXO it spends is a witness program (OP_n + push of 2-40 bytes), or if it is a P2SH UTXO, whose redeemScript is such a witness program. If an input is a segwit spend, the scriptSig must be exactly empty or exactly the P2SH redeemScript and nothing else. If an input is not a segwit spend, the witness stack must be empty. But all of this is rules that govern whether a transaction is valid w.r.t. its witnesses and scriptSigs; it is unrelated to whether the transaction has witnesses.