09 - scriptLen and scriptPubKey

Part of the transaction pre-image is the scriptPubKey held in the UTXO being spent in the input. This is broken down as two fields as follows:

1. scriptLen - the locking script length, a VarInt (1, 3, 5 or 9 bytes depending on script length)

2. scriptPubKey - the locking script for the UTXO being spent (Length as defined by previous parameter)

To extract the locking script, we must first extract the length. There are 4 possible sizes for the VarInt, depending on the length of the script. The size of this field can be inferred from the value in its first byte:

• If the value is equal to or less than 0xFC, the varInt is a 1 byte integer value containing the integer value of scriptLen.

• If the first byte is 0xFD, the varInt is 3 bytes long, with the last 2 bytes containing the integer value of scriptLen.

• If the first byte is 0xFE, the varInt is 5 bytes long, with the last 4 bytes containing the integer value of scriptLen.

• If the first byte is 0xFF, the varInt is 9 bytes long, with the last 8 bytes containing the integer value of scriptLen.

Example 1: 2 byte length field

Typically, we can know roughly how big a script might be. A 2 byte length is valid from 253B up to 64kB so we can assume for our purposes this is what we are expecting.

Stack	Script	Description
<r_tx_preimg>	...	Version, hash_prevouts, hash_nSequence and hash_outpoints have been removed
<r_tx_preimg>	0P_3	VarInt is 3 bytes long
<r_tx_preimg> 0x03	OP_SPLIT	Split length
<varint> <rr_tx_preimg>	OP_SWAP	Move to top of stack
<rr_tx_preimg> <varint>	OP_1	First byte is VarInt length field
<rr_tx_preimg> <varint> 0x01	OP_SPLIT	Calculate txid using OP_HASH160
<rr_tx_preimg> <varint_id> <length_be>	OP_NIP	Nip varint ID
<rr_tx_preimg> <length_be>	OP_SPLIT	Split script from pre-image
<rrr_tx_preimg> <lock_script>	...	Rest of script

Now that the script is on the stack, it is possible for it to utilise data stored in itself to enforce the conditions of the next output state. We will get into this shortly.

Example 2: Variable VarInt handling

This example splits a script of unknown length from the stack. It first splits off the 1-byte VarInt type identifier, checks whether the varInt is 1 byte, 3 bytes or 5 bytes long, and then where needed splits the length value from r_tx_preimg before separating the script from the pre-image.

Stack	Script	Description
<r_tx_preimg>	...	Version, hash_prevouts, hash_nSequence and hash_outpoints have been removed
<r_tx_preimg>	0P_1	VarInt type ID is 1 byte
<r_tx_preimg> 0x01	OP_SPLIT	Split type
<type> <rr_tx_preimg>	OP_SWAP	Move to top of stack
<rr_tx_preimg> <type>	OP_2
<rr_tx_preimg> <type> 0x02	OP_NUM2BIN
<rr_tx_preimg> <type>	OP_DUP	Duplicate type
<rr_tx_preimg> <type> <type>	0xFC	0xFC or less is 1 byte varInt
<rr_tx_preimg> <type> <type+00> 0xFC00	OP_GREATERTHANOREQUALTO	Is type >= 0xFE? - Use GREATERTHANOREQUAL because 0xFE is a NEGATIVE integer
<rr_tx_preimg> <type> <result>	OP_NOTIF	If NOT, enter loop. Otherwise <type> is length
<rr_tx_preimg> <type>	OP_DUP	Duplicate type
<rr_tx_preimg> <type> <type>	0xFD	Is length 2 bytes?
<rr_tx_preimg> <type> <type> 0xFD	OP_EQUAL	Test
<rr_tx_preimg> <type> <result>	OP_IF	If 2 bytes then...
<rr_tx_preimg> <type>	OP_DROP	Drop type
<rr_tx_preimg>	OP_2	2 byte length
<rr_tx_preimg> 0x02	OP_SPLIT	Split
<length> <rrr_tx_preimg>	OP_SWAP	Move length_bigendian to top of stack
<rr_tx_preimg> <type>	OP_ELSE	If not 2-byte
<rr_tx_preimg> <type>	0xFE	Is length 4 bytes?
<rr_tx_preimg> <type> 0xFE	OP_EQUAL	Check equality
<rr_tx_preimg> <result>	OP_IF	Enter if statement
<rr_tx_preimg>	OP_4	Length is 4 bytes
<rr_tx_preimg> 0x04	OP_SPLIT	Split length
<length> <rrr_tx_preimg>	OP_SWAP	Swap it to front
<rr_tx_preimg>	OP_ELSE	If not 4, must be 8
<rr_tx_preimg>	OP_8	Length is 8 bytes
<rr_tx_preimg> 0x08	OP_SPLIT	Split length
<length> <rrr_tx_preimg>	OP_SWAP	Swap to front
<rrr_tx_preimg> <length>	OP_ENDIF	Exit IF loop
<rrr_tx_preimg> <length>	OP_ENDIF	Exit IF loop
<rrr_tx_preimg> <length>	0x00	Add 00 to stack (cannot use OP_FALSE)
<rrr_tx_preimg> <length> 0x00	OP_CAT	Add zeroes to the length to ensure it will be interpreted as a positive integer
<rrr_tx_preimg> <length+00>	OP_BIN2NUM	Optimally encode the nubmer
<r_tx_preimg> <length>	OP_SPLIT	split the script from the pre-image remainder
<lock_script> <rrrr_tx_preimg>	OP_SWAP	Swap to the front
<rrrr_tx_preimg> <lock_script>	...	Rest of script

These script elements can easily be customised to your requirements as you define your OP_PUSH_TX script. Simple checks may require scripts smaller than 253 bytes allowing these checks to be optimised as needed. Understanding the processing of the scriptLen value is an important aspect of this process. Much care must be taken to handle integer values so that they are interpreted correctly.