This avoids some false negatives for is_proxy_tor.
(previously we only set is_proxy_tor when the proxy settings were changed)
In particular, consider scenario:
- Tor browser not running
- user sets "localhost:9150" as proxy
- detection sets network.is_proxy_tor to False
- user starts Tor browser
- network, due to retries, finds proxy working and connects to some servers
- network.is_proxy_tor remains False
- `bitstring` started depending on `bitarray` in version 4.1 [0]
- that would mean one additional dependency for us (from yet another maintainer), which is not even pure python
- we only use bitstring for bolt11-parsing
- hence this PR rewrites the bolt11-parsing and removes `bitstring` as dependency
- note: I benchmarked lndecode using [1], and the new code performs better,
taking around 80% time needed for old code (when using bitstring 3.1.9, pure python).
Though the variance is quite large in both cases.
[0]: 95ee533ee4/release_notes.txt (L108)
[1]: d7597d96d0
- CoinGecko restricted its historical API to last 365 days
- we used to ask for, and get, the whole history, but now we can only ask for the last year
- so change HTTP request to only ask for 365 days
- we cache historical rates to disk
- previously we used to overwrite what is already stored, with the newly obtained data
- now this is changed so that we merge the new data into the already stored data
It might make sense to allow bump_fee on a local tx, and so wallet.get_tx_info() allows it.
For dscancel/cpfp, it does not allow it either. Still, I think this logic should be encapsulated in the wallet.
This gives more time for the client to come back online.
see https://github.com/spesmilo/electrum/issues/8940
- re note on submarine_swaps.py#L53:
lnpeer.Peer.maybe_fulfill_htlc only checks against MIN_FINAL_CLTV_DELTA_ACCEPTED(=144),
so this increased cltv_delta is not enforced when receiving the htlc on ln.
It is put in the invoice, so the sender is supposed to honour it ofc.
It would be nice to enforce it (make the check in maybe_fulfill_htlc dependent on
what was in the invoice).
related https://github.com/spesmilo/electrum/pull/8924
note: "forwarding_key" can be stored as False. I think this is ugly and
might be better to do a storage upgrade to change those values to None.
In maybe_fulfill_htlc, return two items: (preimage, (payment_key, callback)).
Rationale: The caller needs a payment_key only if we return a callback.
If we do not, the caller should use the payment_key that was previously
stored in channel.unfulfilled_htlcs
Note that for trampoline onions, payment_key may contain the inner
or outer secret.
