6c7a17a8e0 psbt: support externally provided preimages for Miniscript satisfaction (Antoine Poinsot)
840a396029 qa: add a "smart" Miniscript fuzz target (Antoine Poinsot)
17e3547241 qa: add a fuzz target generating random nodes from a binary encoding (Antoine Poinsot)
611e12502a qa: functional test Miniscript signing with key and timelocks (Antoine Poinsot)
d57b7f2021 refactor: make descriptors in Miniscript functional test more readable (Antoine Poinsot)
0a8fc9e200 wallet: check solvability using descriptor in AvailableCoins (Antoine Poinsot)
560e62b1e2 script/sign: signing support for Miniscripts with hash preimage challenges (Antoine Poinsot)
a2f81b6a8f script/sign: signing support for Miniscript with timelocks (Antoine Poinsot)
61c6d1a844 script/sign: basic signing support for Miniscript descriptors (Antoine Poinsot)
4242c1c521 Align 'e' property of or_d and andor with website spec (Pieter Wuille)
f5deb41780 Various additional explanations of the satisfaction logic from Pieter (Pieter Wuille)
22c5b00345 miniscript: satisfaction support (Antoine Poinsot)
Pull request description:
This makes the Miniscript descriptors solvable.
Note this introduces signing support for much more complex scripts than the wallet was previously able to solve, and the whole tooling isn't provided for a complete Miniscript integration in the wallet. Particularly, the PSBT<->Miniscript integration isn't entirely covered in this PR.
ACKs for top commit:
achow101:
ACK 6c7a17a8e0
sipa:
utACK 6c7a17a8e0 (to the extent that it's not my own code).
Tree-SHA512: a71ec002aaf66bd429012caa338fc58384067bcd2f453a46e21d381ed1bacc8e57afb9db57c0fb4bf40de43b30808815e9ebc0ae1fbd9e61df0e7b91a17771cc
dee8549be3 test: simplify and speedup mempool_updatefromblock.py by using MiniWallet (Sebastian Falbesoner)
Pull request description:
This PR simplifies the functional test mempool_updatefromblock.py by using MiniWallet in order to avoid manual low-level tx creation (signing, outputs selection, fee calculation). Most of the tedious work is done by the method `MiniWallet.send_self_transfer_multi` (calling `create_self_transfer_multi` internally) which supports spending a given set of UTXOs and creating a certain number of outputs.
As a nice side-effect, the test's performance increases significantly (~3.5x on my system):
```
master
1m56.80s real 1m50.10s user 0m06.36s system
PR
0m32.34s real 0m30.26s user 0m01.41s system
```
The arguments `start_input_txid` and `end_address` have been removed from the `transaction_graph_test` method, as they are currently unused and I don't see them being needed for future tests.
ACKs for top commit:
brunoerg:
crACK dee8549be3
MarcoFalke:
lgtm ACK dee8549be3🚏
Tree-SHA512: 9f6da634bdc8c272f9a2af1cddaa364ee371d4e95554463a066249eecebb668d8c6cb123ec8a5404c41b3291010c0c8806a8a01dd227733cec03e73aa93b0103
511aa4f1c7 Add unit test for ChaCha20's new caching (Pieter Wuille)
fb243d25f7 Improve test vectors for ChaCha20 (Pieter Wuille)
93aee8bbda Inline ChaCha20 32-byte specific constants (Pieter Wuille)
62ec713961 Only support 32-byte keys in ChaCha20{,Aligned} (Pieter Wuille)
f21994a02e Use ChaCha20Aligned in MuHash3072 code (Pieter Wuille)
5d16f75763 Use ChaCha20 caching in FastRandomContext (Pieter Wuille)
38eaece67b Add fuzz test for testing that ChaCha20 works as a stream (Pieter Wuille)
5f05b27841 Add xoroshiro128++ PRNG (Martin Leitner-Ankerl)
12ff72476a Make unrestricted ChaCha20 cipher not waste keystream bytes (Pieter Wuille)
6babf40213 Rename ChaCha20::Seek -> Seek64 to clarify multiple of 64 (Pieter Wuille)
e37bcaa0a6 Split ChaCha20 into aligned/unaligned variants (Pieter Wuille)
Pull request description:
This is an alternative to #25354 (by my benchmarking, somewhat faster), subsumes #25712, and adds additional test vectors.
It separates the multiple-of-64-bytes-only "core" logic (which becomes simpler) from a layer around which performs caching/slicing to support arbitrary byte amounts. Both have their uses (in particular, the MuHash3072 code can benefit from multiple-of-64-bytes assumptions), plus the separation results in more readable code. Also, since FastRandomContext effectively had its own (more naive) caching on top of ChaCha20, that can be dropped in favor of ChaCha20's new built-in caching.
I thought about rebasing #25712 on top of this, but the changes before are fairly extensive, so redid it instead.
ACKs for top commit:
ajtowns:
ut reACK 511aa4f1c7
dhruv:
tACK crACK 511aa4f1c7
Tree-SHA512: 3aa80971322a93e780c75a8d35bd39da3a9ea570fbae4491eaf0c45242f5f670a24a592c50ad870d5fd09b9f88ec06e274e8aa3cefd9561d623c63f7198cf2c7
772671245d test: p2p: check that headers message with invalid proof-of-work disconnects peer (Sebastian Falbesoner)
Pull request description:
One of the earliest anti-DoS checks done after receiving and deserializing a `headers` message from a peer is verifying whether the proof-of-work is valid (called in method `PeerManagerImpl::ProcessHeadersMessage`):
f227e153e8/src/net_processing.cpp (L2752-L2762)
The called method `PeerManagerImpl::CheckHeadersPoW` calls `Misbehaving` with a score of 100, i.e. leading to an immediate disconnect of the peer:
f227e153e8/src/net_processing.cpp (L2368-L2372)
This PR adds a simple test for both the misbehaving log and the resulting disconnect. For creating a block header with invalid proof-of-work, we first create one that is accepted by the node (the difficulty field `nBits` is copied from the genesis block) and based on that the nonce is modified until we have block header hash prefix that is too high to fulfill even the minimum difficulty.
ACKs for top commit:
Sjors:
ACK 772671245d
achow101:
ACK 772671245d
brunoerg:
crACK 772671245d
furszy:
Code review ACK 77267124 with a non-blocking speedup.
Tree-SHA512: 680aa7939158d1dc672b90aa6554ba2b3a92584b6d3bcb0227776035858429feb8bc66eed18b47de0fe56df7d9b3ddaee231aaeaa360136603b9ad4b19e6ac11
ab4efad51b test: fix immediate tx relay in wallet_groups.py (Sebastian Falbesoner)
Pull request description:
In the functional test wallet_groups.py we whitelist peers on all nodes (`-whitelist=noban@127.0.0.1`) to enable immediate tx relay for fast mempool synchronization. However, considering that this setting only applies to inbound peers and the default test topology looks like this:
```
node0 <--- node1 <---- node2 <--- ... <-- nodeN
```
txs propagate fast only from lower- to higher-numbered nodes (i.e. "left to right" in the above diagram) and take long from higher- to lower-numbered nodes ("right to left") since in the latter direction we only have outbound peers, where the trickle relay is still active. As a consequence, if a tx is submitted from any node other than node0, the mempool synchronization can take quite long.
This PR fixes this by simply adding another connection from node0 to the last node, leading to a ~2-3x speedup (5 runs measured via `time ./test/functional/wallet_groups.py` are shown):
```
master:
0m53.31s real 0m08.22s user 0m05.60s system
0m32.85s real 0m07.44s user 0m04.08s system
0m46.40s real 0m09.18s user 0m04.23s system
0m46.96s real 0m11.10s user 0m05.74s system
0m57.23s real 0m10.53s user 0m05.59s system
PR:
0m19.64s real 0m09.58s user 0m05.50s system
0m18.05s real 0m07.77s user 0m04.03s system
0m18.99s real 0m07.90s user 0m04.25s system
0m17.49s real 0m07.56s user 0m03.92s system
0m18.11s real 0m07.74s user 0m03.88s system
```
Note that in most tests this is not a problem since txs very often originate from node0.
ACKs for top commit:
brunoerg:
utACK ab4efad51b
Tree-SHA512: 12675357e6eb5a18383f2bfe719a184c0790863b37a98749d8e757dd5dc3a36212e16a81f0a192340c11b793eda00db359c7011f46f7c27e3a093af4f5b62147
We'll need a better integration of the hash preimages PSBT fields to
satisfy Miniscript with such challenges from the RPC.
Thanks to Greg Sanders for his examples and suggestions to improve this
test.
Try to solve a script using the Miniscript satisfier if the legacy
solver fails under P2WSH context. Only solve public key and public key
hash challenges for now.
We don't entirely replace the raw solver and especially rule out trying to
solve CHECKMULTISIG-based multisigs with the Miniscript satisfier since
some features, such as the transaction input combiner, rely on the
specific behaviour of the former.
741c215b5f test: remove unused vars in `feature_block` (brunoerg)
Pull request description:
There is no need to assign `self.next_block` to variables if we're not using its return value. Most cases touched here, we're reassigning it right after with the value from `self.update_block`.
Top commit has no ACKs.
Tree-SHA512: 25bbea2a09f38c3a3483fa363f024d2a8edd06a00cccc93cef99e489b9a3485d58bbd6a1ed2dddc00f1cebec7e63aed8ad95701a2645ce20a0db9b69573c20a7
c9ba4f9ecb test: Add test for file system permissions (Hennadii Stepanov)
581f16ef34 Apply default umask in `SetupEnvironment()` (Hennadii Stepanov)
8a6219e543 Remove `-sysperms` option (Hennadii Stepanov)
Pull request description:
On master (1e7564eca8) docs say:
```
$ ./src/bitcoind -help | grep -A 3 sysperms
-sysperms
Create new files with system default permissions, instead of umask 077
(only effective with disabled wallet functionality)
```
Basing on that, one could expect that running `bitcoind` first time will create data directory and `wallets/` subdirectory with safe 0700 permissions.
But that is not the case:
```
$ stat .bitcoin | grep id
Access: (0775/drwxrwxr-x) Uid: ( 1000/ hebasto) Gid: ( 1000/ hebasto)
$ stat .bitcoin/wallets | grep id
Access: (0775/drwxrwxr-x) Uid: ( 1000/ hebasto) Gid: ( 1000/ hebasto)
```
Both directories, in fact, are created with system default permissions.
With this PR:
```
$ stat .bitcoin/wallets | grep id
Access: (0700/drwx------) Uid: ( 1000/ hebasto) Gid: ( 1000/ hebasto)
$ stat .bitcoin/wallets | grep id
Access: (0700/drwx------) Uid: ( 1000/ hebasto) Gid: ( 1000/ hebasto)
```
---
This PR:
- is alternative to bitcoin/bitcoin#13389
- fixesbitcoin/bitcoin#15902
- fixesbitcoin/bitcoin#22595
- closesbitcoin/bitcoin#13371
- reverts bitcoin/bitcoin#4286
Changes in behavior: removed `-sysperms` command-line argument / configure option. The related discussions are here:
- https://github.com/bitcoin/bitcoin/pull/13389#issuecomment-395306690
- https://github.com/bitcoin/bitcoin/pull/13389#issuecomment-539906114
- https://github.com/bitcoin/bitcoin/pull/13389#discussion_r279160472
If users rely on non-default access permissions, they could use `chmod`.
ACKs for top commit:
john-moffett:
ACK c9ba4f9ecb
willcl-ark:
ACK c9ba4f9ecb
Tree-SHA512: 96c745339e6bd0e4d7bf65daf9a721e2e1945b2b0ab74ca0f66576d0dc358b5de8eb8cdb89fe2160f3b19c39d2798bb8b291784316085dc73a27102d3415bd57
b8032293e6 Remove use of snprintf and simplify (John Moffett)
Pull request description:
These are the only remaining uses of `snprintf` in our project, and they can cause unexpected issues -- for example, see https://github.com/bitcoin/bitcoin/issues/27014. Change them to use our `ToString` (which uses a locale-independent version of `std::to_string`) to convert an `int` to `std::string`. Also remove resulting unused parts of `StringContentsSerializer`.
Closes https://github.com/bitcoin/bitcoin/issues/27014
ACKs for top commit:
Sjors:
tACK b8032293e6, fixes#27014.
Tree-SHA512: c903977e654711929decafe8887d0de13b38a340d7082875acc5d41950d834dcfde074e9cabecaf5f9a760f62c34322297b4b156af29761650ef5803b1a54b59
One test case uses snprintf to convert an
int to a string. Change it to use ToString
(which uses a locale-independent version of
std::to_string). Also remove unnecessary
parts of StringContentsSerializer.
The descriptor inference logic would previously always use a dummy
signing provider and would never analyze the witness script of a P2WSH
scriptPubKey.
Note even a valid Miniscript might not always be decodable from Script
without more contextual information (for instance the key preimage for a
pk_h).
6d31900e52 wallet: migrate wallet, exit early if no legacy data exist (furszy)
Pull request description:
The process first creates a backup file then return an error,
without removing the recently created file, when notices that
the db is already running sqlite.
ACKs for top commit:
john-moffett:
ACK 6d31900e52
achow101:
ACK 6d31900e52
ishaanam:
crACK 6d31900e52
Tree-SHA512: 9fb52e80de96e129487ab91bef13647bc4570a782003b1e37940e2a00ca26283fd24ad39bdb63a984ae0a56140b518fd0d74aa2fc59ab04405b2c179b7d3c54a
b530d9605d test: refactor: introduce `replace_in_config` helper (Sebastian Falbesoner)
Pull request description:
Currently two functional tests (p2p_permissions.py and wallet_crosschain.py) include quite similar code for substituting strings in a TestNode's bitcoind configuration file, so refactoring that out to a dedicated helper method seems to make sense (probably other tests could need that too in the future).
ACKs for top commit:
kouloumos:
ACK b530d9605d
Tree-SHA512: 5ca65a2ef3292460e5720d5c6acf7326335001858e8f71ab054560117f9479dbadb1dacd8c9235f67f3fcfd08dbc761b62704f379cbf619bba8804f16a25bc7b
Xoroshiro128++ is a fast non-cryptographic random generator.
Reference implementation is available at https://prng.di.unimi.it/
Co-Authored-By: Pieter Wuille <pieter@wuille.net>
b0fa5989e1 test: Check that orphaned coinbase unconf spend is still abandoned (Andrew Chow)
9addbd7890 wallet: Automatically abandon orphaned coinbases and their children (Andrew Chow)
Pull request description:
When a block is reorged out of the main chain, any descendants of the coinbase will no longer be valid. Currently they are only marked as inactive, which means that our balance calculations will still include them. In order to be excluded from the balance calculation, they need to either be abandoned or conflicted. This PR goes with the abandoned method.
Note that even when they are included in balance calculations, coin selection will not select outputs belonging to these transactions because they are not in the mempool.
Fixes#14148
ACKs for top commit:
furszy:
ACK b0fa5989 with a not-blocking nit.
aureleoules:
reACK b0fa5989e1
ishaanam:
ACK b0fa5989e1
Tree-SHA512: 68f12e7aa8df392d8817dc6ac0becce8dbe83837bfa538f47027e6730e5b2e1b1a090cfcea2dc598398fdb66090e02d321d799f087020d7e1badcf96e598c3ac
39b93649c4 test: add functional test for IBD stalling logic (Martin Zumsande)
0565951f34 p2p: Make block stalling timeout adaptive (Martin Zumsande)
Pull request description:
During IBD, there is the following stalling mechanism if we can't proceed with assigning blocks from a 1024 lookahead window because all of these blocks are either already downloaded or in-flight: We'll mark the peer from which we expect the current block that would allow us to advance our tip (and thereby move the 1024 window ahead) as a possible staller. We then give this peer 2 more seconds to deliver a block (`BLOCK_STALLING_TIMEOUT`) and if it doesn't, disconnect it and assign the critical block we need to another peer.
Now the problem is that this second peer is immediately marked as a potential staller using the same mechanism and given 2 seconds as well - if our own connection is so slow that it simply takes us more than 2 seconds to download this block, that peer will also be disconnected (and so on...), leading to repeated disconnections and no progress in IBD. This has been described in #9213, and I have observed this when doing IBD on slower connections or with Tor - sometimes there would be several minutes without progress, where all we did was disconnect peers and find new ones.
The `2s` stalling timeout was introduced in #4468, when blocks weren't full and before Segwit increased the maximum possible physical size of blocks - so I think it made a lot of sense back then.
But it would be good to revisit this timeout now.
This PR makes the timout adaptive (idea by sipa):
If we disconnect a peer for stalling, we now double the timeout for the next peer (up to a maximum of 64s). If we connect a block, we half it again up to the old value of 2 seconds. That way, peers that are comparatively slower will still get disconnected, but long phases of disconnecting all peers shouldn't happen anymore.
Fixes#9213
ACKs for top commit:
achow101:
ACK 39b93649c4
RandyMcMillan:
Strong Concept ACK 39b93649c4
vasild:
ACK 39b93649c4
naumenkogs:
ACK 39b93649c4
Tree-SHA512: 85bc57093b2fb1d28d7409ed8df5a91543909405907bc129de7c6285d0810dd79bc05219e4d5aefcb55c85512b0ad5bed43a4114a17e46c35b9a3f9a983d5754
8609f24be2 test: refactor: simplify p2p_eviction.py by using MiniWallet (Sebastian Falbesoner)
7aa4b32cd4 test: refactor: simplify p2p_tx_download.py by using MiniWallet (Sebastian Falbesoner)
Pull request description:
Similar to #26892, this PR simplies the functional tests p2p_tx_download.py and p2p_eviction.py by using MiniWallet in order to avoid manual low-level tx creation. For the latter, rather than mining 100 blocks manually, the pre-mined chain of the test framework is used.
These instances were found via `$ git grep signrawtransactionwithkey ./test/functional`. AFAICT, there are no other instances where MiniWallet could replace tx creation trivially.
ACKs for top commit:
MarcoFalke:
review ACK 8609f24be2
Tree-SHA512: dfb0103fe7f0625d125e8e4408baed8bfc1ff579954af17d0ead5277e05f933b2c2d98a0093e8109e947635f1718d5c58d837ab825f26077fac0a40575bd3e6f
fafeddfe0e rpc: Throw more user friendly arg type check error (MarcoFalke)
Pull request description:
The arg type check error doesn't list which arg (position or name) failed. Fix that.
ACKs for top commit:
stickies-v:
ACK fafeddfe0e - although I think the functional test isn't in a logical place (but not blocking)
Tree-SHA512: 17425aa145aab5045940ec74fff28f0e3b2b17ae55f91c4bb5cbcdff0ef13732f8e31621d85964dc2c04333ea37dbe528296ac61be27541384b44e37957555c8
b9d5674541 init: Remove sensitive flag from rpcbind (Andrew Chow)
Pull request description:
`-rpcbind` is currently flagged as a sensitive option which means that its value will be masked when the command line args are written to the debug.log file. However this is not useful as if `rpcbind` is actually activated, the bound IP addresses will be written to the log anyways. The test `feature_config_args.py` did not catch this contradiction as the test node was not started with `rpcallowip` and so `rpcbind` was not acted upon.
This also brings `rpcbind` inline with `bind` as that is not flagged as sensitive either.
ACKs for top commit:
Sjors:
re-utACK b9d5674541
willcl-ark:
ACK b9d5674
theStack:
ACK b9d5674541
Tree-SHA512: 50ab5ad2e18ae70649deb1ac429d404b5f5c41f32a4943b2041480580152df22e72d4aae493379d0b23fcb649ab342376a82119760fbf6dfdcda659ffd3e244a
b68e5a7fef lint: specify the right commit range when running locally (James O'Beirne)
dff7ed5732 test: add an easy way to run linters locally (James O'Beirne)
Pull request description:
Adds a Dockerfile configuration ~~(originally written mostly by fanquake)~~ that allows straightforward running of linters with compatible versions locally. This removes a ton of annoyance when trying to appease CI, because many of the linter versions are quite old and difficult to maintain locally.
I realize that people may not be thrilled to add more ancillary tooling to the repo, but I think this makes a lot of sense given the linter versions listed in this container configuration are dictated by this repo (within the CI configuration), so having these things live in two separate places is a recipe for version mismatches.
Eventually we can likely just use this container on CI directly to avoid any chance of inconsistencies between local dev experience and CI.
ACKs for top commit:
aureleoules:
ACK b68e5a7fef
stickies-v:
ACK b68e5a7fe
john-moffett:
ACK b68e5a7fef
Tree-SHA512: 7ef7a5dae023d81fdb6296d5d92dfa074ee321c7993e607c9f014d0f21c91558611aa00fc3ce1edc7b5e68371aea0d27fa1931291a79bb867a6c783bb536775f
d96d97ad30 doc: Add release note for shutdownnotify. (klementtan)
0bd73e2c45 util: Add -shutdownnotify option. (klementtan)
Pull request description:
**Description**: Similar to `-startupnotify`, this PR adds a new option to allow users to specify a command to be executed when Bitcoin Core shuts down.
**Note**: The `shutdownnotify` commands will not be executed if bitcoind shut down due to *unexpected* reasons (ie `killall -9 bitcoind`).
### Testing:
**Normal shutdown commands**
```
# start bitcoind with shutdownnotify optioin
./src/bitcoind -signet -shutdownnotify="touch foo.txt"
# shutdown bitcoind
./src/bitcoin-cli -signet stop
# check that foo.txt has been created
```
**Final RPC call**
Commands:
```
$ ./src/bitcoind -signet -nolisten -noconnect -shutdownnotify="./src/bitcoin-cli -signet getblockchaininfo > tmp.txt"
$ ./src/bitcoin-cli stop
$ cat tmp.txt
```
<details>
<summary>Screen Shot</summary>
</details>
ACKs for top commit:
achow101:
ACK d96d97ad30
1440000bytes:
ACK d96d97ad30
theStack:
re-ACK d96d97ad30
Tree-SHA512: 16f7406fd232e8b97aea5e58854c84755b0c35c88cb3ef9ee123b29a1475a376122b1e100da860cc336d4d657e6046a70e915fdb9b70c9fd097c6eef1b028161
Adds a Dockerfile configuration
that allows straightforward running of linters with compatible versions
locally. This removes a ton of annoyance when trying to appease CI,
because many of the linter versions are quite old and difficult to
maintain locally.
I realize that people may not be thrilled to more ancillary tooling to
the repo, but I think this makes a lot of sense given the linter
versions listed in this container configuration are dictated by this
repo (within the CI configuration), so having these things live in
two separate places is a recipe for version mismatches.
Eventually we can likely just use this container on CI directly to avoid
any chance of inconsistencies between local dev experience and CI.
fa6b402114 test: Run mempool_packages.py with MiniWallet (MarcoFalke)
fa448c27d2 test: Return fee from MiniWallet (MarcoFalke)
faec09f240 test: Return chain of MiniWallet txs from MiniWallet chain method (MarcoFalke)
faa12d4ccd test: Refactor MiniWallet sign_tx (MarcoFalke)
fa2d82103f test: Return wtxid from create_self_transfer_multi (MarcoFalke)
Pull request description:
This allows to run the test even when no wallet is compiled in.
Also, it is a lot nicer to read now.
ACKs for top commit:
glozow:
reACK fa6b402
Tree-SHA512: de0338068fd51db01d64ce270f94fd2982a63a6de597325cd1e1f11127e9075bd4aeacace0ed76d09a2db8b962b27237cf11edb4c1fe1a01134d397f8a11bd05
603d295199 test: wallet: add coverage for `-spendzeroconfchange` setting (Sebastian Falbesoner)
50112034bc test: remove `-spendzeroconfchange` setting from mempool_limit.py (Sebastian Falbesoner)
Pull request description:
This PR adds missing test coverage for the `-spendzeroconfchange` setting (in particular the non-default case `=0`). Note that in contrast to the name, the setting does not only apply to change outputs, but in fact to _all_ unconfirmed outputs that we sent to ourselves, i.e. we can trigger the testing path simply with a single recipient address. The first commit removes the setting from the functional test mempool_limit.py, where it doesn't have any effect, since the test was changed to use MiniWallet in commit dddca3899c.
ACKs for top commit:
brunoerg:
crACK 603d295199
Tree-SHA512: 15d9c8bd5eb37c6b228bf887eb27debee0a391c82356662785da4553ee2558e611834c3936ef7136812b46f877bab7aa5f3088bbd278b81f296bdda96cc8e1c3
fa9f6d7bcd rpc: Run type check against RPCArgs (MarcoFalke)
faf96721a6 test: Fix wrong types passed to RPCs (MarcoFalke)
Pull request description:
It seems brittle to require `RPCTypeCheck` being called inside the code logic. Without compile-time enforcement this will lead to places where it is forgotten and thus to inconsistencies and bugs. Fix this by removing the calls to `RPCTypeCheck` and doing the check internally.
The changes should be reviewed as refactoring changes. However, if they change behavior, it will be a bugfix. For example the changes here happen to also detect/fix bugs like the one fixed in commit 3b5fb6e77a.
ACKs for top commit:
ajtowns:
ACK fa9f6d7bcd
Tree-SHA512: fb2c0981fe6e24da3ca7dbc06898730779ea4e02ea485458505a281cf421015e44dad0221a04023fc547ea2c660d94657909843fc85d92b847611ec097532439
