|
| 1 | +--- |
| 2 | +title: 'Bitcoin Optech Newsletter #15' |
| 3 | +permalink: /en/newsletters/2018/10/02/ |
| 4 | +name: 2018-10-02-newsletter |
| 5 | +type: newsletter |
| 6 | +layout: newsletter |
| 7 | +lang: en |
| 8 | +--- |
| 9 | +This week's newsletter includes a notice of Bitcoin Core 0.17's |
| 10 | +impending release, links to the backport releases of Bitcoin Core 0.15 |
| 11 | +and 0.14 to fix the CVE-2018-17144 duplicate inputs bug for those users |
| 12 | +unable to run more recent releases, a brief description of a chainsplit |
| 13 | +on testnet, and links to notable merges in Bitcoin infrastructure |
| 14 | +projects. |
| 15 | + |
| 16 | +## Action items |
| 17 | + |
| 18 | +- **Upgrade to Bitcoin Core 0.17:** the new release has been tagged and |
| 19 | + several people have begun reproducing builds of the software, so the |
| 20 | + binaries and formal release announcement are likely to become |
| 21 | + available Tuesday or Wednesday on [BitcoinCore.org][]. The |
| 22 | + announcement will include a copy of the release notes detailing major |
| 23 | + changes to the software since the 0.16.0 release. |
| 24 | + |
| 25 | +## News |
| 26 | + |
| 27 | +- **Bitcoin Core [0.15.2][] and [0.14.3][] released:** although source code has |
| 28 | + been available for these older branches since the public announcement |
| 29 | + of the [CVE-2018-17144][] duplicate inputs bug, getting enough people |
| 30 | + to certify a reproducible build took extra time before the |
| 31 | + [binaries][bcco /bin] could be made available. |
| 32 | + |
| 33 | +- **CVE-2018-17144 duplicate inputs bug exploited on testnet:** |
| 34 | + last Thursday a block was created on testnet containing a transaction |
| 35 | + that spent the same input twice. As expected, nodes believed to be |
| 36 | + vulnerable to the bug accepted the block and all other nodes rejected |
| 37 | + it, leading to a consensus failure (chainsplit) where the chain with |
| 38 | + the most proof of work contained the duplicate inputs and a weaker |
| 39 | + chain did not. |
| 40 | + |
| 41 | + Eventually, the chain without the duplicate inputs gained more proof |
| 42 | + of work and the vulnerable nodes attempted to switch to it. This |
| 43 | + caused the vulnerable nodes to attempt to re-add the duplicate input |
| 44 | + to the UTXO database twice, triggering an assert and causing them to |
| 45 | + shutdown. When restarted, operators of the vulnerable nodes needed |
| 46 | + to manually trigger a lengthy reindex procedure to fix their nodes' |
| 47 | + database inconsistencies. (This side-effect of recovering from a |
| 48 | + duplicate inputs chainsplit was previously known to developers.) |
| 49 | + |
| 50 | + Nodes upgraded to Bitcoin Core 0.16.3, 0.17.0RC4, or running other |
| 51 | + software that wasn't vulnerable had no reported problems. However, |
| 52 | + many block explorers with a testnet mode did accept the vulnerable |
| 53 | + block, providing a reminder that users should be careful about using |
| 54 | + third-parties to determine whether or not transactions are valid. |
| 55 | + |
| 56 | +## Notable code changes |
| 57 | + |
| 58 | +*Notable code changes this week in [Bitcoin Core][core commits], |
| 59 | +[LND][lnd commits], and [C-lightning][cl commits].* |
| 60 | + |
| 61 | +{% include linkers/github-log.md |
| 62 | + refname="core commits" |
| 63 | + repo="bitcoin/bitcoin" |
| 64 | + start="920c090f63f4990bf0f3b3d1a6d3d8a8bcd14ba0" |
| 65 | + end="c9327306b580bb161d1732c0a0260b46c0df015c" |
| 66 | +%} |
| 67 | +{% include linkers/github-log.md |
| 68 | + refname="lnd commits" |
| 69 | + repo="lightningnetwork/lnd" |
| 70 | + start="f4305097e1638f6f8958dfa9eec941d8bf80246e" |
| 71 | + end="79ed4e8b600e4834f058cbf3cb8b93f5aa5ab3d4" |
| 72 | +%} |
| 73 | +{% include linkers/github-log.md |
| 74 | + refname="cl commits" |
| 75 | + repo="ElementsProject/lightning" |
| 76 | + start="3ce53ab9eddd397d57b6afc5faefe6703e56ac26" |
| 77 | + end="d6fcfe00c722f7e6f4b691cd47743ed593aeea0e" |
| 78 | +%} |
| 79 | + |
| 80 | +- [Bitcoin Core #14305][]: after the discovery of a few cases where |
| 81 | + Python-based tests were passing incorrectly as a result of using |
| 82 | + misnamed variables, a variable name whitelist was implemented using |
| 83 | + Python 3's `__slots__` feature for classes. |
| 84 | + |
| 85 | +- [LND #1987][]: the `NewWitnessAddress` RPC has been removed and the |
| 86 | + `NewAddress` RPC now only supports generating addresses for |
| 87 | + P2SH-wrapped P2WKH and native P2WPKH. |
| 88 | + |
| 89 | +- [C-Lightning #1982][]: The `invoice` RPC now implements [RouteBoost][] |
| 90 | + by including a [BOLT11][] `r` parameter in the invoice that provides |
| 91 | + routing information to the payer for an already-open channel that has |
| 92 | + the capacity to support paying the invoice. This parameter was originally |
| 93 | + intended to help support private routes, but it can also be used this |
| 94 | + way to support nodes that no longer want to accept new incoming |
| 95 | + channels. Alternatively, if no available channel can support |
| 96 | + payment of the invoice, C-Lightning will emit a warning. |
| 97 | + |
| 98 | +{% include references.md %} |
| 99 | +{% include linkers/issues.md issues="14305,1987,1982" %} |
| 100 | + |
| 101 | +[0.16.3]: https://bitcoincore.org/en/2018/09/18/release-0.16.3/ |
| 102 | +[0.15.2]: https://github.com/bitcoin/bitcoin/releases/tag/v0.15.2 |
| 103 | +[0.14.3]: https://github.com/bitcoin/bitcoin/releases/tag/v0.14.3 |
| 104 | +[cve-2018-17144]: https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2018-17144 |
| 105 | +[bcc 0.17]: https://bitcoincore.org/bin/bitcoin-core-0.17.0/ |
| 106 | +[bcco /bin]: https://bitcoincore.org/bin/ |
| 107 | +[routeboost]: https://lists.linuxfoundation.org/pipermail/lightning-dev/2018-September/001417.html |
0 commit comments