[spam][crazy][wrong][ot] ramblings

Karl gmkarl at gmail.com
Wed Nov 24 16:59:36 PST 2021


omigoodness!  the thing i thought was 'in an hour' is not actually
happening this week.  instead i have a trip to prepare for.

for now, i'll work a little more.  i generated a private key for
testing!  i'll send it funs!

i use mainnet for testing [so that i burn funds in doing so.  helps my
psychological issues to waste money.  the ones where i think i'm being
beaten up by a group that hates me.  they do less imaginary beating up
if i waste my money.]


i sent $0.99 in BSV to my test address ! 1749.

here's the private key line for gits and shiggles:
# note: this private key is not private
privkey = bitcoinx.PrivateKey.from_hex('088412ca112561ff5db3db83e2756fe447d36ba3c556e158c8f016a2934f7279')

1750 now for testingish ummmmmmmmmmmmmmmmmmmmm
i'll want to enumerate that utxo from the address.  so i can prep for
buliding transactions. 1751

1756 the rpc server doesn't like my scripthash =/  i have its source
code! i'll see what's wrong with it!

RPCError: (1, '423410ca994a0641076449 is not a valid script hash')

1757 ooookay hashX and scripthash are actually different formats.

1803 hashx's are truncated to 11 bytes.  scripthashes are 32 hexadecimal bytes.

1808 woohoo!  i got a utxo! a scripthash is the hexadecimal
byte-reversed sha256 hash of a script output's raw bytes.  there's one
per address, kinda.

y'know ... that means a lot of other things.  anything that's a
repeated script output, you could ask an electrum server to list them.
.... maybe?  dunno, maybe not.  it's 1810.

here's my utxo:

[{'tx_hash': '86ca468461094f502eb21b3161366e626056f0d299d99080539cfcb4e02a8dc3',
  'tx_pos': 0,
  'height': 714954,
  'value': 647200}]

i guess i'll want my api function to convert that to a structure from
a library.  luckily it includes value, which is super useful for
transaction building.

bitcoinx doesn't have a good structure =/  i guess i'll just return the dict.

1815 i don't remember about hte 'script_sig' field in transaction
inputs.  i think that's where we complete he script of the preceding
output, with a cryptographic signature?  i wonder how ot generate it.
1816.  maybe the bitcoinx tests will show how to do this.

1845 i've signed a dummy input in a repl using the electrum (not
electrumx) libraries.

        pre_hash = sha256d(bfh(self.serialize_preimage(txin_index,
        privkey = ecc.ECPrivkey(privkey_bytes)
        sig = privkey.sign_transaction(pre_hash)
        sig = bh2u(sig) + sighash_type

looks like it serializes something, then double-sha256's the bytes,
then performs an EC signature on that hash ... then passes that to
bh2u and appends sighash_type (a basic byte)

signing_pubkey=pubkey, sig=sig)

so 3 funcs to sort out

        elif txin.script_type in ['p2pkh', 'p2wpkh', 'p2wpkh-p2sh']:
            pubkey = pubkeys[0]
            pkh = bh2u(hash_160(bfh(pubkey)))
            return bitcoin.pubkeyhash_to_p2pkh_script(pkh)

gets p2pkh script

def pubkeyhash_to_p2pkh_script(pubkey_hash160: str) -> str:
    return construct_script([

in electrumx, there's electrumx.lib.script.ScriptPubKey.P2PKH_script:
    def P2PKH_script(cls, hash160):
        return (bytes((OpCodes.OP_DUP, OpCodes.OP_HASH160))
                + Script.push_data(hash160)
                + bytes((OpCodes.OP_EQUALVERIFY, OpCodes.OP_CHECKSIG)))

this appears to be the electrum-non-x scriptsig for p2pkh:
        elif _type == 'p2pkh':
            return construct_script([sig_list[0], pubkeys[0]])

i'll try to do it by hand.  1901.  hmm.

1907 this makes a p2pkh script in bitcoinx:

1920 in bitcoinx, privkey.sign(data, bitcoinx.double_sha256)

1926 i've drafted code to sign an op_return transaction using only the
bitcoinx library

1940 1926 might have been a typo for 1936 or such.  i'm noticing that
bitcoinx is bsv-specific.  might be a better library out there.

1951 transaction was rejected by network rules !  which is much farther.
1953 i believe i had set the transaction version to 0 instead of 1.  uncertain.
1954 version fixed.  now: 'missing inputs'.

i vaguely recall that 'missing inputs' might mean that the node i'm
sending the transaction to, hasn't received the utxo i'm spending.

1958 didn't byte-reverse the txid in the input !

1958 next network rule: mempool min fee not met.  i wonder what my fee
was calculated to be.  is there a way to query for this min fee?

i have accomplished much.  sending this spamlog.

More information about the cypherpunks mailing list