Notably verifymessage and decrypt(message) were silently ignoring trailing garbage
or inserted non-base64 characters present in signatures/ciphertext.
(both the CLI commands and in the GUI)
I think it is much cleaner and preferable to treat such signatures/ciphertext as invalid.
In fact I find it surprising that base64.b64decode(validate=False) is the default.
Perhaps we should create a helper function for it that set validate=True and use that.
This new `Transaction.verify_sig_for_txin` function is an instance method of `Transaction` instead of `PartialTransaction`.
It takes a complete txin, a pubkey and a signature, and verifies the signature.
- `get_preimage_script` is renamed to `get_scriptcode_for_sighash` and now effectively has two implementations:
- the old impl became `PartialTxInput.get_scriptcode_for_sighash`
- this assumes we are the ones constructing a spending txin and can have knowledge beyond what will be revealed onchain
- the new impl is in the base class, `TxInput.get_scriptcode_for_sighash`
- this assumes the txin is already "complete", and mimics a consensus-verifier by extracting the required fields
from the already complete witness/scriptSig and the scriptpubkey of the funding utxo
- `serialize_preimage` now does not require a PartialTransaction, it also works on the base class Transaction
-----
I intend to use this for debugging only atm: I noticed TxBatcher sometimes creates invalid signatures by seeing
that bitcoind rejects txs with `mandatory-script-verify-flag-failed (Signature must be zero for failed CHECK(MULTI)SIG operation)`.
However the txs in question have multiple txins, with some txins containing multiple signatures, and bitcoind does not tell us
which txin/signature is invalid. Knowing which signature is invalid would be a start, but I can now add some temp debug logging
to `serialize_preimage` to compare the message being signed with the message being verified.
As can be seen from the tests, the signature and the pubkey needs to be manually extracted from the txin to be verified:
we still don't have a script interpreter so we don't have logic to "verify a txin". However this new code adds logic
to verify a signature for a txin/pubkey combo (which is a small part of an interpreter/verifier).
This removes support for Ledger HW.1 and "Nano" (non-S) devices.
These were manufactured/sold around 2015-2016, and are long unsupported by the upstream vendor.
We previously added a deprecation warning to the GUI [0] released in 4.3.3 (2023-01-02), to warn owners of these devices.
This PR now fully removes support.
As a consequence, the unmaintained btchip-python dependency can now be removed, which solves [1].
[0]: 9b82eb6d06
[1]: https://github.com/spesmilo/electrum/issues/9370#issuecomment-2593675364
Traceback (most recent call last):
File "...\electrum\electrum\gui\qt\main_window.py", line 1797, in toggle_search
self.search_box.setFocus(1)
TypeError: arguments did not match any overloaded call:
setFocus(self): too many arguments
setFocus(self, reason: Qt.FocusReason): argument 1 has unexpected type 'int'
Instead of some functions operating with hex strings,
and others using bytes, this consolidates most things to use bytes.
This mainly focuses on bitcoin.py and transaction.py,
and then adapts the API usages in other files.
Notably,
- scripts,
- pubkeys,
- signatures
should be bytes in almost all places now.
A new config API is introduced, and ~all of the codebase is adapted to it.
The old API is kept but mainly only for dynamic usage where its extra flexibility is needed.
Using examples, the old config API looked this:
```
>>> config.get("request_expiry", 86400)
604800
>>> config.set_key("request_expiry", 86400)
>>>
```
The new config API instead:
```
>>> config.WALLET_PAYREQ_EXPIRY_SECONDS
604800
>>> config.WALLET_PAYREQ_EXPIRY_SECONDS = 86400
>>>
```
The old API operated on arbitrary string keys, the new one uses
a static ~enum-like list of variables.
With the new API:
- there is a single centralised list of config variables, as opposed to
these being scattered all over
- no more duplication of default values (in the getters)
- there is now some (minimal for now) type-validation/conversion for
the config values
closes https://github.com/spesmilo/electrum/pull/5640
closes https://github.com/spesmilo/electrum/pull/5649
Note: there is yet a third API added here, for certain niche/abstract use-cases,
where we need a reference to the config variable itself.
It should only be used when needed:
```
>>> var = config.cv.WALLET_PAYREQ_EXPIRY_SECONDS
>>> var
<ConfigVarWithConfig key='request_expiry'>
>>> var.get()
604800
>>> var.set(3600)
>>> var.get_default_value()
86400
>>> var.is_set()
True
>>> var.is_modifiable()
True
```
fix signing txs when using old "bitcoin app" (pre-2.1) on the ledger device
```
33.36 | W | transaction | heyheyhey. cp1. include_sigs=True force_legacy=False use_segwit_ser=True
33.36 | W | transaction | heyheyhey. cp2. branch1
33.37 | E | plugins.ledger.ledger |
Traceback (most recent call last):
File "...\electrum\electrum\plugins\ledger\ledger.py", line 669, in sign_transaction
rawTx = tx.serialize_to_network()
File "...\electrum\electrum\transaction.py", line 945, in serialize_to_network
witness = ''.join(self.serialize_witness(x, estimate_size=estimate_size) for x in inputs)
File "...\electrum\electrum\transaction.py", line 945, in <genexpr>
witness = ''.join(self.serialize_witness(x, estimate_size=estimate_size) for x in inputs)
File "...\electrum\electrum\transaction.py", line 839, in serialize_witness
sol = desc.satisfy(allow_dummy=estimate_size, sigdata=txin.part_sigs)
File "...\electrum\electrum\descriptor.py", line 378, in satisfy
sol = self._satisfy_inner(sigdata=sigdata, allow_dummy=allow_dummy)
File "...\electrum\electrum\descriptor.py", line 574, in _satisfy_inner
raise MissingSolutionPiece(f"no sig for {pubkey.hex()}")
electrum.descriptor.MissingSolutionPiece: no sig for 033e92e55923ea25809790f292ee9bd410355ee02492472d9a1ff1b364874d0679
33.38 | I | plugins.ledger.ledger | no sig for 033e92e55923ea25809790f292ee9bd410355ee02492472d9a1ff1b364874d0679
```
fixes https://github.com/spesmilo/electrum/issues/8365
regression from https://github.com/spesmilo/electrum/pull/8230
in particular, ledger: fix sign_message for some wallets
```
156.02 | E | plugins.ledger.ledger |
Traceback (most recent call last):
File "...\electrum\electrum\plugins\ledger\ledger.py", line 1265, in sign_message
result = base64.b64decode(self.client.sign_message(message, address_path))
File "...\Python310\site-packages\ledger_bitcoin\client.py", line 230, in sign_message
sw, response = self._make_request(self.builder.sign_message(message_bytes, bip32_path), client_intepreter)
File "...\Python310\site-packages\ledger_bitcoin\command_builder.py", line 176, in sign_message
bip32_path: List[bytes] = bip32_path_from_string(bip32_path)
File "...\Python310\site-packages\ledger_bitcoin\common.py", line 68, in bip32_path_from_string
return [int(p).to_bytes(4, byteorder="big") if "'" not in p
File "...\Python310\site-packages\ledger_bitcoin\common.py", line 68, in <listcomp>
return [int(p).to_bytes(4, byteorder="big") if "'" not in p
ValueError: invalid literal for int() with base 10: '84h'
```
Regression from df2bd61de6, where the
default hardened char was changed from "'" to "h". Note that there was
no corresponding wallet db upgrade, so some files use one char and
others use the other.
B023 Function definition does not bind loop variable 'already_selected_buckets_value_sum'
in keepkey/qt.py, looks like this was an actual bug
(fixed in trezor plugin already: 52a4810752 )
Required a much higher mental load to parse the name "convert_bip32_path_to_list_of_uint32"
than to parse "convert_bip32_strpath_to_intpath".
And we already have the ~inverse: "convert_bip32_intpath_to_strpath".
