7175d73573
create-close-tx: create a mutual close tx from the close sigs.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
c6ed5926cb
close-channel: add --complete option.
Code is identical, so don't create separate close-channel-complete binary.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
38f7a23064
close-channel: create message to mutually close channel.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
16baa1dae9
create-commit-spend-tx: spend the commit tx (after timeout)
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
114b7afbe2
bitcoin_script: fix bug in complex anchor redeem script.
We need to pop off the stack after OP_CHECKSEQUENCEVERIFY.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
3a4d8cd282
bitcoin_script: use OP_NOP dummy for now, as bitcoind rejects OP_NOP3
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
3c7e11fd65
bitcoin_script: DER correctly when we have leading zeroes.
Happens one in 128 times, I guess...
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
a25ba452f3
permute_tx: fix horrible thinko.
Map needs to be inverted, otherwise non-trivial shuffles get confused.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
4dd6b8e385
lightning.proto: Rename 'to_me' to 'final', use for commit_tx.
This is where the commit tx outputs should pay to.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
baf5d11643
getinput.sh: Switch to -regtest
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
9ed6968fe5
add_varint: encode correctly (LE)
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
0203d51ec0
pull_varint: fix.
Adapted badly from pettycoin.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
fc8552318a
opt_bits: parsing routines for 'bits' == 100 satoshi.
Also fix open-channel usage message.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
126192a842
create-steal-tx: get my money back!
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
904ef32459
bitcoin_tx: handle \n at end of file for tx.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
ca18eeeaa7
open-channel: fix revocation hash.
Revocation hash is SHA256 of output of generator. Sometimes
that is the same as the next revocation hash (on even counts), but
not otherwise.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
6c62e20701
Note that we should check their locktime, minconfirms etc.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
6048a15db9
bitcoin_tx: bitcoin_txid_from_hex helper.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
e1532ce044
bitcoin_script: get rid of scriptpubkey_pay_to_pubkeyhash.
We no longer generate any pay-to-pubkeys.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
d30c470c7d
anchor: change is not an arbitrary output, but a pubkey we p2sh to.
Gets rid of the last pubkey user; they're generally deprecated.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
29d0cdc3cd
open-channel: use pubkey, not arbitrary script for final outout address.
We need to pay to this from two places: on their side, it's a simple
payment, on our side, it's a complex timeout-or-mutual-or-hval script,
which doesn't lend itself to arbitrary scripts.
Use P2SH, of course.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
0742e5356f
get-anchor-depth: simple helper to get min_confirms from OpenChannel packet
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
33bbd38691
Makefile: cleanup and generalize
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
a6b08dc393
protocol: add revocation hash to update_accept for other commit tx.
And clarify that we're going to be closing the anchor tx directly.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
f7f7f3f1bb
check-anchor-scriptsigs: don't require txs.
It's convenient to double-check that commit txs spend anchor, but it's
misleading since we wouldn't have access to the other side's tx, and
in practice we wouldn't even sign ours until we need it.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
8516798645
signature: fix signatures for p2sh inputs.
The subscript in this case is the redeemscript, not the input script.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
cbad9a1fa5
bitcoin_script: remember to push extra 0 for OP_CHECKMULTISIG.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
769d2fd722
bitcoin_script: fix out-by-one error in multisig.
Was using 3 instead of 2. OP_LITERAL() was a bad idea anyway,
as it would only handle numbers up to 16 (and we use 32).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
17c56a8cfc
Fix DER encoding.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
623c6562af
Makefile: add TAGS target.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
abac31072e
open-anchor-scriptsig: use pubkey, not addr in scriptsig.
Oops.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
50609c6771
open-channel: expect hex txids to be in LE (ie. reversed!) order.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
3ed87c5e6f
anchor: fix anchor production for RIPEMD hash.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
617d21eb43
check-anchor-scriptsigs: dump out anchor tx if scriptsigs make expected tx.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
a6d1c65683
Use new opt_usage_exit_fail().
opt_usage_and_exit() exits with status 0, and prints to stdout.
This exits status 1 and prints to stderr.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
67f98fe6b6
Rename "anchor sig" to "anchor scriptsigs" to be clearer.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
cf333e45e1
Fix thinko: open-commit-sig needs to sign *their* commit tx.
And check-commit-sig needs to check it against ours.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
fe4972fda1
Anchor: apply permutation to get txid.
Otherwise both sides create different anchors.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
c291d19af6
Add check-commit-sig.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
316f29cb69
bitcoin_script: wean entirely off protobuf types.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
f911b2b6df
Use pubkey structures in bitcoin_script, rather than protobufs.
This ensures we do checking beforehand, and keeps abstractions clear.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
bd38003db6
Add function to linearize tx into bytes, by generalizing hash code.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
393400fa39
Make an explicit bitcoin_txid() call.
Neater than open-coding it. We still need the raw version for
signatures though.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
07c7214edb
Explicit pubkey structure.
This checks that the protobuf is the right form, also handles uncompressed
keys (though you shouldn't be using those any more, should you?)
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
7c977a7633
Move signature <-> proto conversions to signature.c, check lower bit.
Make sure signature can be used by bitcoin.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
eec612af9f
Use raw 64-byte signatures in protocol.
DER encoding introduces problems for non-canonical encodings; we should
do that only at the lightning<->bitcoin interface.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
64b443dcd9
Add doubts on me & you sig on commit transaction.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
c04d20e019
commit_tx: new file containing logic to create initial commitment transaction.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
bf38ca052e
open-commit-sig: create signature for commit tx.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
63986e5b2d
Rename perturb to permute.
Perturb is a bit stretched for this, permute is better.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
10 years ago
Rusty Russell