Because we don't have type checking for command-line/settings/config
args, strings are interpreted as 'false' for non-boolean args.
By convention, this "forces" us to interpret negated strings as 'true',
which conflicts with the negated option definition in all the settings
classes (they expect negated options to always be false and ignore any
other value preceding them). Consequently, when retrieving all "wallet"
values from the command-line/settings/config, we also fetch the negated
string boolean value, which is not of the expected 'string' type.
This mismatch leads to an internal fatal error, resulting in an unclean
shutdown during initialization. Furthermore, this error displays a poorly
descriptive error message:
"JSON value of type bool is not of expected type string"
This commit fixes the fatal error by ensuring that only string values are
returned in the "wallet" settings list, failing otherwise. It also improves
the clarity of the returned error message.
Co-authored-by: Ryan Ofsky <ryan@ofsky.org>
41ad84a00c seeds: Use fjahr's more up to date asmap (Ava Chow)
d8fd1e0faf seeds: Fixed seeds update (Ava Chow)
f1f24d7214 seeds: Add testnet4 fixed seeds file (Ava Chow)
8ace71c737 seeds: Remove manual onion and i2p seeds (Ava Chow)
ed5b86cbe4 seeds: Add testnet instructions (Ava Chow)
0676515397 seeds: Also pull from achow101 seeder (Ava Chow)
5bab3175a6 makeseeds: Configurable minimum blocks for testnet4's smaller chain (Ava Chow)
d2465dfac6 makeseeds: Shuffle ips after parsing (Ava Chow)
af550b3a0f makeseeds: Support CJDNS (Ava Chow)
d5a8c4c4bd makeseeds: Update user agent regex (Ava Chow)
Pull request description:
The [DNS seeder](https://github.com/achow101/dnsseedrs) that I wrote collects statistics on node reliability in the same way that sipa's seeder does, and also outputs this information in the same file format. Thus it can also be used in our fixed seeds update scripts. My seeder additionally crawls onion v3, i2p, and cjdns, so will now be able to set those fixed seeds automatically rather than curating manual lists.
In doing this update, I've found that `makeseeds.py` is missing newer versions from the regex as well as cjdns support; both of these have been updated.
I also noticed that the testnet fixed seeds are all manually curated and sipa's seeder does not appear to publish any testnet data. Since I am also running the seeder for testnet, I've added the commands to generate testnet fixed seeds from my seeder's data too.
Lastly, I've updated all of the fixed seeds. However, since my seeder has not found any cjdns nodes that met the reliability criteria (possibly due to connectivity issues present in those networks), I've left the previous manual seeds for that network.
ACKs for top commit:
fjahr:
re-ACK 41ad84a00c
virtu:
ACK [41ad84a](41ad84a00c)
Tree-SHA512: 6ba0141f053d9d6ae7d8c9574f061be38f3e65b28de1d6660c1885ab942623b5a0ec70754b4fcfc5d98fe970f5f179a940d5880b5061ed698f7932500e01d3ee
- Settings updates were not thread-safe, as they were executed in
three separate steps:
1) Obtain settings value while acquiring the settings lock.
2) Modify settings value.
3) Overwrite settings value while acquiring the settings lock.
This approach allowed concurrent threads to modify the same base value
simultaneously, leading to data loss. When this occurred, the final
settings state would only reflect the changes from the last thread
that completed the operation, overwriting updates from other threads.
Fix this by making the settings update operation atomic.
- Add test coverage for this behavior.
Co-authored-by: furszy <matiasfurszyfer@protonmail.com>
Accepting any Rng in RandMoney makes tests more flexible to use a
different Rng. Also, passing in the Rng clarifies the call sites, so
that they all use g_rand_ctx explicitly and consistently.
The global g_insecure_rand_ctx will be removed in the future, so
removing it from this helper is useful.
Also, tying the two concepts of the global internal RNGState and the
global test-only rng context is a bit confusing, because tests can
simply use the m_rng, if it exists. Also, tests may seed more than one
random context, or none at all, or a random context of a different type.
Fix all issues by moving the Reseed call to the two places where it is
used.
60055f1abc test: replace deprecated secp256k1 context flags usage (Sebastian Falbesoner)
Pull request description:
The flags `SECP256K1_CONTEXT_{SIGN,VERIFY}` have been marked as deprecated since libsecp256k1 version 0.2 (released in December 2022), with the recommendation to use SECP256K1_CONTEXT_NONE instead, see https://github.com/bitcoin-core/secp256k1/pull/1126 and 1988855079/CHANGELOG.md (L132). Note that in contrast to other deprecated functions/variables, these defines don't have a deprecated attribute and hence don't lead to a compiler warning (see https://github.com/bitcoin-core/secp256k1/pull/1126#discussion_r922105271), so they are not easily detected.
ACKs for top commit:
TheCharlatan:
ACK 60055f1abc
ismaelsadeeq:
utACK 60055f1abc
tdb3:
light CR and test ACK 60055f1abc
Tree-SHA512: d93cf49e018a58469620c0d2f50242141f22dabc70afb2a7cd64e416f4f55588714510ae5a877376dd1e6b6f7494261969489af4b18a1c9dff0d0dfdf93f1fa8
Removes dependency on unsafe and deprecated uint256S.
This makes parsing more strict, by requiring RANDOM_CTX_SEED
to be a string of up to 64 hex digits (optionally prefixed with
"0x"), whereas previously any string would be accepted, with
non-hex characters silently ignored and input longer than
64 characters (ignoring "0x" prefix) silently trimmed.
Can be tested with:
$ RANDOM_CTX_SEED=z ./src/test/test_bitcoin --log_level=all --run_test=timeoffsets_tests/timeoffsets_warning -- -printtoconsole=1 | grep RANDOM_CTX_SEED
RANDOM_CTX_SEED must consist of up to 64 hex digits ("0x" prefix allowed), it was set to: 'z'.
Co-Authored-By: MarcoFalke <*~=`'#}+{/-|&$^_@721217.xyz>
Removes dependency on unsafe and deprecated uint256S.
This makes parsing more strict, by returning an error
when the input contains non-hex characters, or when it
contains more than 64 hex digits.
Also make feature_assumevalid.py more robust by using CBlock.hash
which is guaranteed to be 64 characters long, as opposed to the
variable-length hex(CBlock.sha256)
Removes dependency on unsafe and deprecated uint256S.
This makes parsing more strict, by returning an error
when the input contains more than 64 hex digits.
FromUserHex will be used in future commits to construct
uint256 instances from user hex input without being
unnecessarily restrictive on formatting by allowing
0x-prefixed input that is shorter than 64 characters.
59ff17e5af miner: adjust clock to timewarp rule (Sjors Provoost)
e929054e12 Add timewarp attack mitigation test (Sjors Provoost)
e85f386c4b consensus: enable BIP94 on regtest (Sjors Provoost)
dd154b0568 consensus: lower regtest nPowTargetTimespan to 144 (Sjors Provoost)
Pull request description:
Because #30647 reduced the timewarp attack threshold from 7200s to 600s, our miner code will fail to propose a block template (on testnet4) if the last block of the previous period has a timestamp two hours in the future. This PR fixes that and also adds a test.
The non-test changes in the last commit should be in v28, otherwise miners have to patch it themselves. If necessary I can split that out into a separate PR, but I prefer to get the tests in as well.
In order to add the test, we activate BIP94 on regtest.
In order for the test to run faster, we reduce its difficulty retarget period to 144, the same number that's already used for softfork activation logic. Regtest does not actually adjust its difficulty, so this change has no effect (except for `getnetworkhashps`, see commit).
An alternative approach would be to run this test on testnet4, by hardcoding its first 2015 in the test suite. But since the timewarp mitigation is a serious candidate for a future mainnet softfork, it seems better to just deploy it on regtest.
The next commits add a test and fix the miner code.
The `MAX_TIMEWARP` constant is moved to `consensus.h` so both validation and miner code have access to it.
ACKs for top commit:
achow101:
ACK 59ff17e5af
fjahr:
ACK 59ff17e5af
glozow:
ACK 59ff17e5af
Tree-SHA512: 50af9fdcba9b0d5c57e1efd5feffd870bd11b5318f1f8b0aabf684657f2d33ab108d5f00b1475fe0d38e8e0badc97249ef8dda20c7f47fcc1698bc1008798830
Integer promotion will already turn the `signed` into `unsigned` in
those lines. However, make the `unsigned` explicit so that the code is
clearer and a compiler warning is avoided when switching to m_rng:
| test/validation_block_tests.cpp: warning: comparison of integers of different signs: 'int' and 'const unsigned int' [-Wsign-compare]
| 136 | bool gen_invalid = m_rng.randrange(100) < invalid_rate;
| | ~~~~~~~~~~~~~~~~~~~~~~~~~ ^ ~~~~~~~~~~~~
| 137 | bool gen_fork = m_rng.randrange(100) < branch_rate;
| | ~~~~~~~~~~~~~~~~~~~~~~~~~ ^ ~~~~~~~~~~~
| 2 warnings generated.
rand_cache is unused since commit
16329224e7, so it can be removed
rand_seed is wrong since commit
022cf47dd7, because it is no longer
printing the seed that was used to seed the global random context in
tests. Instead, it prints a (random-ish) value derived from the global
random context via InsecureRand256().
Finally, the for loop creating new prevector_tester objects will always
use the same seed since commit fae43a97ca,
because repeated calls to SeedInsecureRand/SeedRandomForTest will always
reseed the global with the same "static const" seed.
Fix all issues by
* removing the unused rand_cache,
* removing the call to SeedRandomForTest which restored the same seed on
every call in the process, and
* Reseeding the global random context with the (random-ish) rand_seed.
fa899fb7aa fuzz: Speed up utxo_snapshot fuzz target (MarcoFalke)
fa386642b4 fuzz: Speed up utxo_snapshot by lazy re-init (MarcoFalke)
fa645c7a86 fuzz: Remove unused DataStream object (MarcoFalke)
fae8c73d9e test: Disallow fee_estimator construction in ChainTestingSetup (MarcoFalke)
Pull request description:
Two commits to speed up unit and fuzz tests.
Can be tested by running the fuzz target and looking at the time it took, or by looking at the flamegraph. For example:
```
FUZZ=utxo_snapshot perf record -g --call-graph dwarf ./src/test/fuzz/fuzz -runs=100
hotspot ./perf.data
ACKs for top commit:
TheCharlatan:
Re-ACK fa899fb7aa
marcofleon:
Re ACK fa899fb7aa
brunoerg:
ACK fa899fb7aa
Tree-SHA512: d3a771bb12d7ef491eee61ca47325dd1cea5c20b6ad42554babf13ec98d03bef8e7786159d077e59cc7ab8112495037b0f6e55edae65b871c7cf1708687cf717
The flags SECP256K1_CONTEXT_{SIGN,VERIFY} have been deprecated since
libsecp256k1 version 0.2 (released in December 2022), with the
recommendation to use SECP256K1_CONTEXT_NONE instead.
This currently has no effect due to fPowNoRetargeting,
except for the getnetworkhashps when called with -1.
It will when the next commit enforces the timewarp attack mitigation on regtest.
16e95bda86 Move maximum timewarp attack threshold back to 600s from 7200s (Matt Corallo)
Pull request description:
In 6bfa26048d the testnet4 timewarp attack fix block time variation was increased from the Great Consensus Cleanup value of 600s to 7200s on the thesis that this allows miners to always create blocks with the current time. Sadly, doing so does allow for some nonzero inflation, even if not a huge amount.
While it could be that some hardware ignores the timestamp provided to it over Stratum and forces the block header timestamp to the current time, I'm not aware of any such hardware, and it would also likely suffer from random invalid blocks due to relying on NTP anyway, making its existence highly unlikely.
This leaves the only concern being pools, but most of those rely on work generated by Bitcoin Core (in one way or another, though when spy mining possibly not), and it seems likely that they will also not suffer any lost work. While its possible that a pool does generate invalid work due to spy mining or otherwise custom logic, it seems unlikely that a substantial portion of hashrate would do so, making the difference somewhat academic (any pool that screws this up will only do so once and the network would come out just fine).
Further, while we may end up deciding these assumptions were invalid and we should instead use 7200s, it seems prudent to try with the value we "want" on testnet4, giving us the ability to learn if the compatibility concerns are an issue before we go to mainnet.
ACKs for top commit:
fjahr:
tACK 16e95bda86
achow101:
ACK 16e95bda86
murchandamus:
crACK 16e95bda86
Tree-SHA512: ae46d03b728b6e23cb6ace64c9813bc01c01e38dd7f159cf0fab53b331ef84b3b811edab225453ccdfedb53b242f55b0efd69829782657490fe393d24dacbeb2
