d135c294764add81683ba47575f9a5dde7d7c07f [ci] make list of previous releases to download a setting (Sjors Provoost)
9c246b873c74834a121edba00fcaecf0cba6f9b4 [test] backwards compatibility: bump v0.19.0.1 to v0.19.1 (Sjors Provoost)
89a28e02fa46f3d5eb07ab02aa34aa95c6fcee11 [test] add v0.16.3 backwards compatibility test (Sjors Provoost)
Pull request description:
Thanks to #18774's `adjust_bitcoin_conf_for_pre_17` we can now test backwards compatibility for v0.16.3, both for sync and loading a recent wallet.
This PR bumps v0.19.0.1 to v0.19.1.
I also made the version list consistent for the `contrib/devtools/previous_release.sh` instruction, between both tests.
ACKs for top commit:
MarcoFalke:
ACK d135c294764add81683ba47575f9a5dde7d7c07f
Tree-SHA512: 5ff137a7a934237fa220f1c2807ce9abeeb75929266558bf3e4045bec7dfcd0a8747fa74d700065c568330b18badf58c60c308eb13d1eed444d4bbfe6decc48b
c456145b2c65f580683df03bf10cd39000cf24d5 [test] add 0.19 backwards compatibility tests (Sjors Provoost)
b769cd142deda74fe46e231cc7b687a86514f2f1 [test] add v0.17.1 wallet upgrade test (Sjors Provoost)
9d9390dab716f07057c94e8e21f3c7dd06192f35 [tests] add wallet backwards compatility tests (Sjors Provoost)
c7ca6308968b29a0e0edc485cd06e68e5edb7c7d [scripts] support release candidates of earlier releases (Sjors Provoost)
8b1460dbd1b732f06d4cebe1fa6844286c7a0056 [tests] check v0.17.1 and v0.18.1 backwards compatibility (Sjors Provoost)
ae379cf7d12943fc192d58176673bcfe7d53da53 [scripts] build earlier releases (Sjors Provoost)
Pull request description:
This PR adds binaries for 0.17, 0.18 and 0.19 to Travis and runs a basic block propagation test.
Includes test for upgrading v0.17.1 wallets and opening master wallets with older versions.
Usage:
```sh
contrib/devtools/previous_release.sh -f -b v0.19.0.1 v0.18.1 v0.17.1
test/functional/backwards_compatibility.py
```
Travis caches these earlier releases, so it should be able to run these tests with little performance impact.
Additional scenarios where it might be useful to run tests against earlier releases:
* creating a wallet with #11403's segwit implementation, copying it to an older node and making sure the user didn't lose any funds (although this PR doesn't support `v0.15.1`)
* future consensus changes
* P2P changes (e.g. to make sure we don't accidentally ban old nodes)
ACKs for top commit:
MarcoFalke:
ACK c456145b2c65f580683df03bf10cd39000cf24d5 🔨
Tree-SHA512: 360bd870603f95b14dc0cd629532cc147344f632b808617c18e1b585dfb1f082b401e5d493a48196b719e0aeaee533ae0a773dfc9f217f704aae898576c19232
Move funds from the coinbase, into the Asset Lock Pool. This is to incentivize MNs to upgrade to platform, because only MNs running platform will get these migrated rewards
fa4632c41714dfaa699bacc6a947d72668a4deef test: Move boost/stdlib includes last (MarcoFalke)
fa488f131fd4f5bab0d01376c5a5013306f1abcd scripted-diff: Bump copyright headers (MarcoFalke)
fac5c373006a9e4bcbb56843bb85f1aca4d87599 scripted-diff: Sort test includes (MarcoFalke)
Pull request description:
When writing tests, often includes need to be added or removed. Currently the list of includes is not sorted, so developers that write tests and have `clang-format` installed will either have an unrelated change (sorting) included in their commit or they will have to manually undo the sort.
This pull preempts both issues by just sorting all includes in one commit.
Please be aware that this is **NOT** a change to policy to enforce clang-format or any other developer guideline or process. Developers are free to use whatever tool they want, see also #18651.
Edit: Also includes a commit to bump the copyright headers, so that the touched files don't need to be touched again for that.
ACKs for top commit:
practicalswift:
ACK fa4632c41714dfaa699bacc6a947d72668a4deef
jonatack:
ACK fa4632c41714dfaa, light review and sanity checks with gcc build and clang fuzz build
Tree-SHA512: 130a8d073a379ba556b1e64104d37c46b671425c0aef0ed725fd60156a95e8dc83fb6f0b5330b2f8152cf5daaf3983b4aca5e75812598f2626c39fd12b88b180
c7437185589926ec8def2af6bede6a407b3d2e4a test: add further BIP37 size limit checks to p2p_filter.py (Sebastian Falbesoner)
Pull request description:
This is a follow-up PR to #18628. In addition to the hash-functions limit test introduced with commit fa4c29bc1d, it adds checks for the following size limits as defined in [BIP37](https://github.com/bitcoin/bips/blob/master/bip-0037.mediawiki):
ad message type `filterload`:
> The filter itself is simply a bit field of arbitrary byte-aligned size. The maximum size is **36,000 bytes**.
ad message type `filteradd`:
> The data field must be smaller than or equal to **520 bytes** in size (the maximum size of any potentially matched object).
Also introduces new constants for the limits (or reuses the max script size constant in case for the `filteradd` limit).
Also fixes#18711 by changing the misbehaviour check on "filteradd without filterset" (introduced with #18544) below to also use the more commonly used `assert_debug_log` method.
ACKs for top commit:
MarcoFalke:
ACK c7437185589926ec8def2af6bede6a407b3d2e4a
robot-visions:
ACK c7437185589926ec8def2af6bede6a407b3d2e4a
jonasschnelli:
utACK c7437185589926ec8def2af6bede6a407b3d2e4a. Seems to fix it: https://bitcoinbuilds.org/index.php?build=2524
Tree-SHA512: a03e7639263eb36a381922afb4e1d0ed2ae286f2ad2e7bbd922509a043ddf6cfd08747e01d54d29bfb8f54b66908f653974b9c347e4ca4f43332b586778893be
a9ecbdfcaa15499644d16e9c8ad2c63dfc45b37b test: add more inactive filter tests to p2p_filter.py (Sebastian Falbesoner)
5eae034996b340c19cebab9efb6c89d20fe051ef net: limit BIP37 filter lifespan (active between 'filterload' and 'filterclear') (Sebastian Falbesoner)
Pull request description:
This PR fixes https://github.com/bitcoin/bitcoin/issues/18483. On the master branch, there is currently _always_ a BIP37 filter set for every peer: if not a specific filter is set through a `filterload` message, a default match-everything filter is instanciated and pointed to via the `CBloomFilter` default constructor; that happens both initially, when the containing structure `TxRelay` is constructed:
c0b389b335/src/net.h (L812)
and after a loaded filter is removed again through a `filterclear` message:
c0b389b335/src/net_processing.cpp (L3201)
The behaviour was introduced by commit 37c6389c5a (an intentional covert fix for [CVE-2013-5700](https://github.com/bitcoin/bitcoin/pull/18515), according to gmaxwell).
This default match-everything filter leads to some unintended side-effects:
1. `getdata` request for filtered blocks (i.e. type `MSG_FILTERED_BLOCK`) are always responded to with `merkleblock`s, even if no filter was set by the peer, see issue #18483 (strictly speaking, this is a violation of BIP37) c0b389b335/src/net_processing.cpp (L1504-L1507)
2. if a peer sends a `filteradd` message without having loaded a filter via `filterload` before, the intended increasing of the banscore never happens (triggered if `bad` is set to true, a few lines below) c0b389b335/src/net_processing.cpp (L3182-L3186)
This PR basically activates the `else`-branch code paths for all checks of `pfilter` again (on the master branch, they are dead code) by limiting the pointer's lifespan: instead of always having a filter set, the `pfilter` is only pointing to a `CBloomFilter`-instance after receiving a `filterload` message and the instance is destroyed again (and the pointer nullified) after receiving a `filterclear` message.
Here is a before/after comparison in behaviour:
| code part / scenario | master branch | PR branch |
| --------------------------------------------- | ---------------------------------- | ---------------------------------------------------- |
| `getdata` processing for `MSG_FILTERED_BLOCK` | always responds with `merkleblock` | only responds if filter was set via `filterload` |
| `filteradd` processing, no filter was loaded | nothing | peer's banscore increases by 100 (i.e. disconnect) |
On the other code parts where `pfilter` is checked there is no change in the logic behaviour (except that `CBloomFilter::IsRelevantAndUpdate()` is unnecessarily called and immediately returned in the master branch).
Note that the default constructor of `CBloomFilter` is only used for deserializing the received `filterload` message and nowhere else. The PR also contains a functional test checking that sending `getdata` for filtered blocks is ignored by the node if no bloom filter is set.
ACKs for top commit:
MarcoFalke:
re-ACK a9ecbdfcaa, only change is in test code 🕙
Tree-SHA512: 1a656a6d74ccaf628e7fdca063ba63fbab2089e0b6d0a11be9bbd387c2ee6d3230706ff8ffc1a55711481df3d4547137dd7c9d9184d89eaa43ade4927792d0b6
## Issue being fixed or feature implemented
## What was done?
Renaming of all classes/variables/functions/rpcs from `hpmn` to `evo`.
## How Has This Been Tested?
All unit and func tests are passing.
Sync of Testnet.
## Breaking Changes
All protx RPCs ending with `_hpmn` were converted to `_evo`.
`_hpmn` RPCs are now deprecated.
Although, they can still be enabled by adding `-deprecatedrpc=hpmn`.
## Checklist:
- [x] I have performed a self-review of my own code
- [ ] I have commented my code, particularly in hard-to-understand areas
- [x] I have added or updated relevant unit/integration/functional/e2e
tests
- [x] I have made corresponding changes to the documentation
- [x] I have assigned this pull request to a milestone _(for repository
code-owners and collaborators only)_
---------
Co-authored-by: thephez <thephez@users.noreply.github.com>
Co-authored-by: UdjinM6 <UdjinM6@users.noreply.github.com>
## Issue being fixed or feature implemented
Execute command when the best chainlock changes (`%s` in cmd is replaced
by chainlocked block hash). Same as `-blocknotify` but for chainlocks.
Let `-instantsendnotify` replace `%w` with wallet name like
`-walletnotify` does.
## What was done?
## How Has This Been Tested?
run tests
## Breaking Changes
n/a
## Checklist:
- [x] I have performed a self-review of my own code
- [ ] I have commented my code, particularly in hard-to-understand areas
- [ ] I have added or updated relevant unit/integration/functional/e2e
tests
- [ ] I have made corresponding changes to the documentation
- [x] I have assigned this pull request to a milestone
fab46b34f4b13abbb0af276c3fb548f25ccc28bd test: Fix restart node race (MarcoFalke)
Pull request description:
It is not allowed to start a node before it has been fully stopped. Otherwise it could lead to intermittent issues due to access issues (e.g. cookie file https://cirrus-ci.com/task/6409665024098304?command=ci#L4793)
Fix that by waiting for the node to fully stop.
ACKs for top commit:
laanwj:
code review ACK fab46b34f4b13abbb0af276c3fb548f25ccc28bd
Tree-SHA512: 7605cac0573a7b04f05ff110d0131e8940d87f7baf6d698505ed16b363d4d15b1e552c5ffd1a187c8fe5639f7e265c3122734c85283275746e46bd789614fd21
fa918dd537fea775c19a590e5f9161bf51a5839b test: Use Popen.wait instead of RPC in assert_start_raises_init_error (MarcoFalke)
Pull request description:
Using RPC (`wait_for_rpc_connection`) has several issue:
* It polls in a loop, which might be slow
* It tries to read the RPC cookie file, which might not be present, thus leading to intermittent issues
Fix both by using `Popen.wait`
ACKs for top commit:
laanwj:
Code review ACK ~~faf7b05be9c86ee61c39e5314511fe2410128a6b~~ fa918dd537fea775c19a590e5f9161bf51a5839b
darosior:
ACK fa918dd537fea775c19a590e5f9161bf51a5839b
Tree-SHA512: 5368ad0d0ea2deb0af9582a42667c9290efe8f2705f37a236afc2c7908b04265ab342e2dd356a57156e99389f4a27ab6da9fa7bf9161fb7568240aa005e693b9
faede1b293354560317b67f0b4e6874dcac6ef41 test: Properly raise FailedToStartError when rpc shutdown before warmup finished (MarcoFalke)
Pull request description:
Should fix issues such as https://travis-ci.org/github/bitcoin/bitcoin/jobs/671910152#L7034
Top commit has no ACKs.
Tree-SHA512: ac659f29c5ec91985c916b734e24911cbf4e2c5c4b1f1891a7e6c2d2511ec285167550fb03848eee4a7a3cbc9f8cdb0c766f4e881d9e44368c7415d007006368
fa40168ab3102b9ad850f967a0e7fa22dbfbd0c6 Remove unused bits from service flags enum (MarcoFalke)
Pull request description:
Remove service bits that haven't been observed on the active network for years and won't ever be observed on the network with this meaning. Keeping this dead assignment in our source code forever doesn't add any value.
I somehow forgot to do this in commit fa0d0ff6e1bee60fde63724ae28a51aac5a94d4a.
ACKs for top commit:
laanwj:
Code review ACK fa40168ab3102b9ad850f967a0e7fa22dbfbd0c6
practicalswift:
cr ACK fa40168ab3102b9ad850f967a0e7fa22dbfbd0c6
fanquake:
ACK fa40168ab3102b9ad850f967a0e7fa22dbfbd0c6
Tree-SHA512: 376e5ac05940493cf2209fea60515c843e978c4b476f2524f6bf7a37a646d237c3ddcf6c0fa23641f9ba550f625609703d9b51b4be631a7f2a90e1092b557232
## Issue being fixed or feature implemented
Since v19, Evo nodes are paid 4x blocks in a row.
This needs to be reverted when MN Reward Reallocation activates.
## What was done?
Starting from MN Reward Reallocation activation, Evo nodes are paid one
block in a row (like regular masternodes).
In addition, `nConsecutivePayments` isn't incremented anymore for Evo
nodes.
## How Has This Been Tested?
`feature_llmq_hpmn.py` with MN Reward Reallocation activation.
## Breaking Changes
no
## Checklist:
- [x] I have performed a self-review of my own code
- [ ] I have commented my code, particularly in hard-to-understand areas
- [x] I have added or updated relevant unit/integration/functional/e2e
tests
- [ ] I have made corresponding changes to the documentation
- [x] I have assigned this pull request to a milestone _(for repository
code-owners and collaborators only)_
---------
Co-authored-by: Konstantin Akimov <knstqq@gmail.com>
bfa9309ad606102f24c9bd3c33dfe78949f09418 Use COINBASE_MATURITY constant in functional tests. (Kiminuo)
525448df9dc2ab6b7e960ff138956ae3e2efdf60 Move COINBASE_MATURITY from `feature_nulldummy` test to `blocktools`. (Kiminuo)
Pull request description:
`COINBASE_MATURITY` constant was added to `feature_nulldummy` test in #21373. This PR moves the constant to `blocktools.py` file and uses the constant in more tests as suggested [here](https://github.com/bitcoin/bitcoin/pull/21373#discussion_r605418462).
Edit: Goal of this PR is to replace integer constants with `COINBASE_MATURITY` but not necessarily in *all* cases because that would mean to read and fully understand all tests. That's out of my time constraints. Any reports where `COINBASE_MATURITY` should be used are welcome though!
ACKs for top commit:
theStack:
ACK bfa9309ad606102f24c9bd3c33dfe78949f09418 🌇
Tree-SHA512: 01f04645f05a39028681f355cf3d42dd63ea3303f76d93c430e0fdce441934358a2d847a54e6068d61932f1b75e1d406f51859b057b3e4b569f7083915cb317f
## Issue being fixed or feature implemented
Current implementation of MnEhfTx is not matched with DIP-0023, this PR
fixes it. It is a prior work for
https://github.com/dashpay/dash/pull/5469
## What was done?
- requestID is fixed from `clsig{quorumHeight}` to `mnhf{versionBit}` +
fixes for signature validation properly
- v20 is minimal height to accept MnEHF special transactions
- versionBit is not BLS version - removed unrelated wrong code and
validations
- TxMempool will accept MnEHF transaction even if inputs/outputs are
zeroes and no fee
- implemented python's serialization/deserialization of MnEHF
transactions for future using in functional tests
## How Has This Been Tested?
Run functional/unit tests. Beside that there's new functional test in
https://github.com/dashpay/dash/pull/5469 that actually test format of
transaction and signature validation - to be merged later.
## Breaking Changes
Payload of MnEhf tx is changed, related consensus rules are changed.
## Checklist:
- [x] I have performed a self-review of my own code
- [x] I have commented my code, particularly in hard-to-understand areas
- [x] I have added or updated relevant unit/integration/functional/e2e
tests
- [x] I have made corresponding changes to the documentation
- [x] I have assigned this pull request to a milestone
8f5dc8800aeb524eee2fa2451cd22883b7b2bfec test: display command line options passed to send_cli() in debug log (Jon Atack)
Pull request description:
as per https://github.com/bitcoin/bitcoin/pull/18691#discussion_r411382589, and revert two cli calls changed in #18691 from rpc commands back to command line options (these were the only occurrences).
ACKs for top commit:
MarcoFalke:
ACK 8f5dc8800aeb524eee2fa2451cd22883b7b2bfec
Tree-SHA512: fcb3eca00aa4099066028c90d5e50a02e074366e09a17f5f5b937d9f7562dd054ff65681aa0ad4c94f6de1e98b1e2b9ac4cd084ddc297010253989a80483b1b9
## Issue being fixed or feature implemented
This is an implementation of DIP0027 "Credit Asset Locks".
It's a mechanism to fluidly exchange between Dash and credits.
## What was done?
This pull request includes:
- Asset Lock transaction
- Asset Unlock transaction (withdrawal)
- Credit Pool in coinbase
- Unit tests for Asset Lock/Unlock tx
- New functional test `feature_asset_locks.py`
RPC: currently locked amount (credit pool) is available through rpc call
`getblock`.
## How Has This Been Tested?
There added new unit tests for basic checks of transaction validity
(asset lock/unlock).
Also added new functional test "feature_asset_locks.py" that cover
typical cases, but not all corner cases yet.
## Breaking Changes
This feature should be activated as hard-fork because:
- It adds 2 new special transaction and one of them [asset unlock tx]
requires update consensus rulels
- It adds new data in coinbase tx (credit pool)
## Checklist:
- [x] I have performed a self-review of my own code
- [x] I have commented my code, particularly in hard-to-understand areas
- [x] I have added or updated relevant unit/integration/functional/e2e
tests
- [ ] I have made corresponding changes to the documentation
**To release DIP 0027**
- [x] I have assigned this pull request to a milestone
---------
Co-authored-by: UdjinM6 <UdjinM6@users.noreply.github.com>
0956e46bff7f0b6da65a4de6d4f8261fe9d7055c test: use zero-argument super() shortcut (Python 3.0+) (Sebastian Falbesoner)
Pull request description:
This mini-PR replaces all calls to `super(...)` with arguments with the zero-argument shortcut `super()` where applicable. See [PEP 3135](https://www.python.org/dev/peps/pep-3135/#specification):
> The new syntax:
>
> super()
>
> is equivalent to:
>
> super(__class__, <firstarg>)
>
> where __class__ is the class that the method was defined in, and <firstarg> is
> the first parameter of the method (normally self for instance methods, and cls
> for class methods).
ACKs for top commit:
fanquake:
ACK 0956e46bff7f0b6da65a4de6d4f8261fe9d7055c
Tree-SHA512: 4ac66fe7ab2be2e8a514e5fcfc41dbb298f21b23ebb7b7b0310d704b0b3cef8adf287a8d80346d1ea9418998c597b4f0ff1f66148d0d806bb43db6607e0fe1cf
## Issue being fixed or feature implemented
Implementation of Randomness Beacon Part 3.
Starting from v20 activation fork, members for quorums are sorted using
(if available) the best CL signature found in Coinbase.
If no CL signature is present yet, then the usual way is used (By using
Blockhash instead)
The actual new way to shuffle is already implemented in
https://github.com/dashpay/dash/pull/5366.
SPV clients also need to calculate members, but they only know block
headers.
Since Coinbase is in the actual block, then they lack the required
information to correctly calculate quorum members.
## What was done?
- Message `MNLISTIDFF` is enriched with a new field `quorumsCLSigs`.
This field holds the Chainlock Signature required for each set of
indexes corresponding to quorums in field `newQuorums`.
- Protocol version has been bumped to `70230`.
- Clients with protocol version greater or equal to `70230` will receive
the new field `quorumsCLSigs`.
- The same field is returned in `protx diff` RPC.
Note:
- Field `quorumsCLSigs` will populated only after v20 activation
- If for one or more quorums, no non-null CL sig was found in CbTx then
a null signature is returned in `quorumsCLSigs`.
## How Has This Been Tested?
- Functional test mininode's protocol version was bumped to `70230`.
- `feature_llmq_rotation.py` checks that `quorumsCLSigs` match in both
P2P and RPC messages.
## Breaking Changes
No
## Checklist:
- [x] I have performed a self-review of my own code
- [x] I have commented my code, particularly in hard-to-understand areas
- [x] I have added or updated relevant unit/integration/functional/e2e
tests
- [x] I have made corresponding changes to the documentation
- [x] I have assigned this pull request to a milestone _(for repository
code-owners and collaborators only)_
---------
Co-authored-by: thephez <thephez@users.noreply.github.com>
Co-authored-by: UdjinM6 <UdjinM6@users.noreply.github.com>
Co-authored-by: pasta <pasta@dashboost.org>
## Issue being fixed or feature implemented
Version field should always be the first field of a message for better
readibility.
## What was done?
- Introduced new protocol version `MNLISTDIFF_VERSION_ORDER` (`70229`).
- `nVersion` serialisation order is changed for clients with protocol
version greater than or equal to `70229`.
- For clients with protocol version >= `70225` and < `70229` the old
order is used: can be deprecated in the future.
- Increased functional test P2P mininode's protocol version to `70229`.
## How Has This Been Tested?
`feature_llmq_rotation.py` with new protocol version.
## Breaking Changes
## Checklist:
- [x] I have performed a self-review of my own code
- [ ] I have commented my code, particularly in hard-to-understand areas
- [x] I have added or updated relevant unit/integration/functional/e2e
tests
- [ ] I have made corresponding changes to the documentation
- [x] I have assigned this pull request to a milestone _(for repository
code-owners and collaborators only)_
## Issue being fixed or feature implemented
Mobile wallets would have to convert 4k+ pubkeys at the V19 fork point
and it's a pretty hard job for them that can easily take 10-15 seconds
if not more. Also after the HF, if a masternode list is requested from
before the HF, the operator keys come in basic scheme, but the
merkelroot was calculated with legacy. From mobile team work it wasn't
possible to convert all operator keys to legacy and then calculate the
correct merkleroot.
~This PR builds on top of ~#5392~ #5403 (changes that belong to this PR:
26f7e966500bdea4c604f1d16716b40b366fc707 and
4b42dc8fcee3354afd82ce7e3a72ebe1659f5f22) and aims to solve both of
these issues.~
cc @hashengineering @QuantumExplorer
## What was done?
Introduce `nVersion` on p2p level for every CSimplifiedMNListEntry. Set
`nVersion` to the same value we have it in CDeterministicMNState i.e.
pubkey serialization would not be via basic scheme only after the V19
fork, it would match the way it’s serialized on-chain/in
CDeterministicMNState for that specific MN.
## How Has This Been Tested?
run tests
## Breaking Changes
NOTE: `testnet` is going to re-fork at v19 forkpoint because
`merkleRootMNList` is not going to match
## Checklist:
- [x] I have performed a self-review of my own code
- [ ] I have commented my code, particularly in hard-to-understand areas
- [ ] I have added or updated relevant unit/integration/functional/e2e
tests
- [ ] I have made corresponding changes to the documentation
- [ ] I have assigned this pull request to a milestone _(for repository
code-owners and collaborators only)_
## Issue being fixed or feature implemented
fix a couple of issues in helpers, extend feature_dip3_v19.py to check
more after v19 fork
## What was done?
pls see individual PRs
## How Has This Been Tested?
run tests
## Breaking Changes
n/a
## Checklist:
- [x] I have performed a self-review of my own code
- [x] I have commented my code, particularly in hard-to-understand areas
- [ ] I have added or updated relevant unit/integration/functional/e2e
tests
- [ ] I have made corresponding changes to the documentation
- [x] I have assigned this pull request to a milestone _(for repository
code-owners and collaborators only)_
## Issue being fixed or feature implemented
Currently, Chainlocks are either enabled or disabled. This PR adds a
third state: enabled but we will not sign new ones.
Should probably backport this to v19.x
## What was done?
Spork state != 0 but active will now result in chain locks being
enforced but not created.
## How Has This Been Tested?
## Breaking Changes
None
## Checklist:
_Go over all the following points, and put an `x` in all the boxes that
apply._
- [ ] I have performed a self-review of my own code
- [ ] I have commented my code, particularly in hard-to-understand areas
- [ ] I have added or updated relevant unit/integration/functional/e2e
tests
- [ ] I have made corresponding changes to the documentation
- [x] I have assigned this pull request to a milestone _(for repository
code-owners and collaborators only)_
---------
Co-authored-by: UdjinM6 <UdjinM6@users.noreply.github.com>
0ed2d8e07d3806d78d03a77d2153f22f9d733a07 test: add BIP37 remote crash bug [CVE-2013-5700] test to p2p_filter.py (Sebastian Falbesoner)
Pull request description:
Integrates the missing message type `filteradd` to the test framework and checks that the BIP37 implementation is not vulnerable to the "remote crash bug" [CVE-2013-5700](https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2013-5700) anymore. Prior to v.0.8.4, it was possible to trigger a division-by-zero error on the following line in the function `CBloomFilter::Hash()`:
f0d6487e29/src/bloom.cpp (L45)
By setting a zero-length filter via `filterload`, `vData.size()` is 0, so the modulo operation above, called on any .insert() or .contains() operation then crashed the node. The test uses the approach of just sending an arbitrary `filteradd` message after, which calls `CBloomFilter::insert()` (and in turn `CBloomFilter::Hash()`) on the node. The vulnerability was fixed by commit 37c6389c5a (an intentional covert fix, [according to gmaxwell](https://github.com/bitcoin/bitcoin/issues/18483#issuecomment-608224095)), which introduced flags `isEmpty`/`isFull` that wouldn't call the `Hash()` member function if `isFull` is true (set to true by default constructor).
To validate that the test fails if the implementation is vulnerable, one can simply set the flags to false in the member function `UpdateEmptyFull()` (that is called after a filter received via `filterload` is constructed), which activates the vulnerable code path calling `Hash` in any case on adding or testing for data in the filter:
```diff
diff --git a/src/bloom.cpp b/src/bloom.cpp
index bd6069b..ef294a3 100644
--- a/src/bloom.cpp
+++ b/src/bloom.cpp
@@ -199,8 +199,8 @@ void CBloomFilter::UpdateEmptyFull()
full &= vData[i] == 0xff;
empty &= vData[i] == 0;
}
- isFull = full;
- isEmpty = empty;
+ isFull = false;
+ isEmpty = false;
}
```
Resulting in:
```
$ ./p2p_filter.py
[...]
2020-04-03T14:38:59.593000Z TestFramework (INFO): Check that division-by-zero remote crash bug [CVE-2013-5700] is fixed
2020-04-03T14:38:59.695000Z TestFramework (ERROR): Assertion failed
[...]
[... some exceptions following ...]
```
ACKs for top commit:
naumenkogs:
utACK 0ed2d8e07d3806d78d03a77d2153f22f9d733a07
Tree-SHA512: 02d0253d13eab70c4bd007b0750c56a5a92d05d419d53033523eeb3ed80318bc95196ab90f7745ea3ac9ebae7caee3adbf2a055a40a4124e0915226e49018fe8
00559229588feb19de2a0cb7506f70c483a1f433 test: add BIP37 'filterclear' test to p2p_filter.py (Sebastian Falbesoner)
Pull request description:
Integrates the message type `filterclear` to the test framework and adds a simple test to `p2p_filter.py`, checking that arbitrary txs get relayed again after deleting the filter.
ACKs for top commit:
naumenkogs:
utACK 00559229588feb19de2a0cb7506f70c483a1f433
Tree-SHA512: fe64e99a526865770707d8077b9968d3923f248045ec7fa56cd380dba85ac77a71a473d244ef3aede2fc0d287b8d7c6bc0156b6033b0c949c2058cc08e255697
e09c701e0110350f78366fb837308c086b6503c0 scripted-diff: Bump copyright of files changed in 2020 (MarcoFalke)
6cbe6209646db8914b87bf6edbc18c6031a16f1e scripted-diff: Replace CCriticalSection with RecursiveMutex (MarcoFalke)
Pull request description:
`RecursiveMutex` better clarifies that the mutex is recursive, see also the standard library naming: https://en.cppreference.com/w/cpp/thread/recursive_mutex
For that reason, and to avoid different people asking me the same question repeatedly (e.g. https://github.com/bitcoin/bitcoin/pull/15932#pullrequestreview-339175124 ), remove the outdated alias `CCriticalSection` with a scripted-diff
## Issue being fixed or feature implemented
Speed thing up: 8075fc0c61
Unify things: ff1a390224 (and _probably_
fix issues like https://gitlab.com/dashpay/dash/-/jobs/4304343867),
876f5c3a9f,
ed58cdda13
Let tsan tests finish on smaller/slower machines:
ba1e3360f9
## What was done?
pls see individual commits
## How Has This Been Tested?
run tests locally and my in gitlab ci
https://gitlab.com/UdjinM6/dash/-/jobs/4319419014
## Breaking Changes
n/a
## Checklist:
- [x] I have performed a self-review of my own code
- [x] I have commented my code, particularly in hard-to-understand areas
- [x] I have added or updated relevant unit/integration/functional/e2e
tests
- [ ] I have made corresponding changes to the documentation
- [x] I have assigned this pull request to a milestone _(for repository
code-owners and collaborators only)_
## Issue being fixed or feature implemented
I think the logic in activate_by_name is broken
## What was done?
fix it
## How Has This Been Tested?
run tests
## Breaking Changes
n/a
## Checklist:
- [x] I have performed a self-review of my own code
- [x] I have commented my code, particularly in hard-to-understand areas
- [ ] I have added or updated relevant unit/integration/functional/e2e
tests
- [ ] I have made corresponding changes to the documentation
- [x] I have assigned this pull request to a milestone _(for repository
code-owners and collaborators only)_
## Issue being fixed or feature implemented
## What was done?
- Bumped version of `CbTx`. Added fields `bestCLHeightDiff`,
`bestCLSignature`
- Miner starting from v20 fork, includes best CL signature in `CbTx` (if
available) or null signature.
- All nodes should verify included CL signature before accepting the
block.
## How Has This Been Tested?
Basically, activated v20 on in the beginning of
`feature_llmq_chainlocks.py`
## Breaking Changes
Yes, new version of CbTx
## Checklist:
- [x] I have performed a self-review of my own code
- [ ] I have commented my code, particularly in hard-to-understand areas
- [x] I have added or updated relevant unit/integration/functional/e2e
tests
- [ ] I have made corresponding changes to the documentation
**For repository code-owners and collaborators only**
- [x] I have assigned this pull request to a milestone
---------
Co-authored-by: UdjinM6 <UdjinM6@users.noreply.github.com>
## Issue being fixed or feature implemented
Functional tests in CI and locally often fails without a good reason
(pretty randomly)
## What was done?
It was re-implemented `get_recovered_sig` and updated `create_raw_tx`
for better selection/change output.
## How Has This Been Tested?
Run functional times with bug amount of parrallel jobs:
```
test/functional/test_runner.py -j 20 feature_llmq_is_cl_conflicts.py feature_llmq_is_cl_conflicts.py feature_llmq_is_cl_conflicts.py feature_llmq_is_cl_conflicts.py feature_llmq_is_cl_conflicts.py feature_llmq_is_cl_conflicts.py feature_llmq_is_cl_conflicts.py feature_llmq_is_cl_conflicts.py feature_llmq_is_cl_conflicts.py feature_llmq_is_cl_conflicts.py feature_llmq_is_cl_conflicts.py feature_llmq_is_cl_conflicts.py feature_llmq_is_cl_conflicts.py feature_llmq_is_cl_conflicts.py feature_llmq_is_cl_conflicts.py feature_llmq_is_cl_conflicts.py feature_llmq_is_cl_conflicts.py feature_llmq_is_cl_conflicts.py feature_llmq_is_cl_conflicts.py feature_llmq_is_cl_conflicts.py feature_llmq_is_cl_conflicts.py
```
Without these changes usually 2-3 instance fails.
With these changes all failures happened only for `p2p_addrv2_relay.py`
and `mempool_unbroadcast.py`. Beside feature_llmq_is_conflicts.py
improved stability of `interface_zmq_dash.py` also.
## Breaking Changes
No breaking changes
## Checklist:
- [x] I have performed a self-review of my own code
- [ ] I have commented my code, particularly in hard-to-understand areas
- [ ] I have added or updated relevant unit/integration/functional/e2e
tests
- [ ] I have made corresponding changes to the documentation
- [x] I have assigned this pull request to a milestone
---------
Co-authored-by: UdjinM6 <UdjinM6@users.noreply.github.com>
e78aaf41f43d0e2ad78fa6d8dad61032c8ef73d0 [docs] Add release notes for burying bip 9 soft fork deployments (John Newbery)
8319e738f9f118025b332e4fa804d4c31e4113f4 [tests] Add coverage for the content of getblockchaininfo.softforks (James O'Beirne)
0328dcdcfcb56dc8918697716d7686be048ad0b3 [Consensus] Bury segwit deployment (John Newbery)
1c93b9b31c2ab7358f9d55f52dd46340397c906d [Consensus] Bury CSV deployment height (John Newbery)
3862e473f0cb71a762c0306b171b591341d58142 [rpc] Tidy up reporting of buried and ongoing softforks (John Newbery)
Pull request description:
This hardcodes CSV and segwit activation heights, similar to the BIP 90 buried deployments for BIPs 34, 65 and 66.
CSV and segwit have been active for over 18 months. Hardcoding the activation height is a code simplification, makes it easier to understand segwit activation status, and reduces technical debt.
This was originally attempted by jl2012 in #11398 and again by me in #12360.
ACKs for top commit:
ajtowns:
ACK e78aaf41f43d0e2ad78fa6d8dad61032c8ef73d0 ; checked diff to previous acked commit, checked tests still work
ariard:
ACK e78aaf4, check diff, run the tests again and successfully activated csv/segwit heights on mainnet as expected.
MarcoFalke:
ACK e78aaf41f43d0e2ad78fa6d8dad61032c8ef73d0 (still didn't check if the mainnet block heights are correct, but the code looks good now)
Tree-SHA512: 7e951829106e21a81725f7d3e236eddbb59349189740907bb47e33f5dbf95c43753ac1231f47ae7bee85c8c81b2146afcdfdc11deb1503947f23093a9c399912
8925df86c4df16b1070343fef8e4d238f3cc3bd1 doc: update release notes (Jon Atack)
8bb405bbadf11391ccba7b334b4cfe66dc85b390 test: getaddressinfo labels purpose deprecation test (Jon Atack)
60aba1f2f11529add115d963d05599130288ae28 rpc: simplify getaddressinfo labels, deprecate previous behavior (Jon Atack)
7851f14ccf2bcd1e9b2ad48e5e08881be06d9d21 rpc: incorporate review feedback from PR 17283 (Jon Atack)
Pull request description:
This PR builds on #17283 (now merged) and is followed by #17585.
It modifies the value returned by rpc getaddressinfo `labels` to an array of label name strings and deprecates the previous behavior of returning an array of JSON hash structures containing label `name` and address `purpose` key/value pairs.
before
```
"labels": [
{
"name": "DOUBLE SPEND",
"purpose": "receive"
}
```
after
```
"labels": [
"DOUBLE SPEND"
]
```
The deprecated behavior can be re-enabled by starting bitcoind with `-deprecatedrpc=labelspurpose`.
For context, see:
- https://github.com/bitcoin/bitcoin/pull/17283#issuecomment-554458001
- http://www.erisian.com.au/bitcoin-core-dev/log-2019-12-13.html#l-425 (lines 425-427)
- http://www.erisian.com.au/bitcoin-core-dev/log-2019-11-22.html#l-622
Reviewers: This PR may be tested manually by building, then running bitcoind with and without the `-deprecatedrpc=labelspurpose` flag while verifying the rpc getaddressinfo help text and `labels` output.
Next steps: deprecate the rpc getaddressinfo `label` field (EDIT: done in #17585) and add support for multiple labels per address. This PR will unblock those.
ACKs for top commit:
jnewbery:
reACK 8925df8
promag:
Code review ACK 8925df86c4df16b1070343fef8e4d238f3cc3bd1.
meshcollider:
Code review ACK 8925df86c4df16b1070343fef8e4d238f3cc3bd1
Tree-SHA512: c2b717209996da32b6484de7bb8800e7048410f9ce6afdb3e02a6866bd4a8f2c730f905fca27b10b877b91cf407f546e69e8c4feb9cd934325a6c71c166bd438
33f5fc32e5bfbe1e89c4d20ce455bcc6dc194151 test: add rpc getaddressinfo labels test coverage (Jon Atack)
0f3539ac6d772fc646b5f184fa1efe77bf632f6a test: add listlabels test in wallet_labels.py (Jon Atack)
1388de83900eaced906d369fe9e8887ae74b2dcf rpc: add getaddressinfo code documentation (Jon Atack)
2ee0cb3330ccf70f0540cb42370796e32eff1569 rpc: update getaddressinfo RPCExamples to bech32 (Jon Atack)
8d1ed0c263f8cdff7189f02040b5d02238d93da0 rpc: clarify label vs labels in getaddressinfo RPCHelpman (Jon Atack)
5a0ed850700dfb19167d40b38f80313bd5e427ca rpc: improve getaddressinfo RPCHelpman content (Jon Atack)
70cda342cd20d0e0cd9f28405457544036968f2d rpc: improve getaddressinfo RPCHelpman formatting (Jon Atack)
Pull request description:
This PR is a continuation of the work in https://github.com/bitcoin/bitcoin/pull/12892.
Main motivations:
- There is currently no test coverage for the getaddressinfo `labels` response. Coverage here is a prerequisite before deprecating the `label` response or adding multiple labels per address.
- `bitcoin-cli help getaddressinfo` returns a few content errors, difficult-to-read formatting, and no explanation why it returns both `label` and `labels` and how they relate, which can be confusing for application developers.
Changes by order of commits:
- [x] improve/fix getaddressinfo RPCHelpman layout formatting
- [x] improve/fix getaddressinfo RPCHelpman content
- [x] clarify the `label` and `labels` fields in getaddressinfo RPCHelpman
- [x] update getaddressinfo RPCExamples addresses to bech32
- [x] add getaddressinfo code docs
- [x] add a `listlabels` test assertion in wallet_labels.py
- [x] add missing getaddressinfo `labels` test coverage and improve the existing `label` tests
Here are gists of the CLI help output:
[`bitcoin-cli help getaddressinfo` before this PR](https://gist.github.com/jonatack/022af5221a85c069780359a22643c810)
[`bitcoin-cli help getaddressinfo` after this PR](https://gist.github.com/jonatack/4ee5f6abc62a3d99269570206a5f90ba)
It seems we ought to begin a deprecation process for the getaddressinfo `label` field? If yes, I have a follow-up ready. _--> EDIT: Deprecation follow-ups #17578 and #17585 now build on this PR._
ACKs for top commit:
fjahr:
Re-ACK 33f5fc32e5bfbe1e89c4d20ce455bcc6dc194151
jnewbery:
ACK 33f5fc32e5bfbe1e89c4d20ce455bcc6dc194151.
Tree-SHA512: a001aa863090ec2566a31059477945b1c303ebeb430b33472f8b150e420fa5742fc33bca9d95571746395b607f43f6078dd5b53e238ac1f3fc648b51c8f79a07
ea4cc3a7b36a9c77dbf0aff439da3ef0ea58e6e4 Truly decouple wallet from chainparams for -fallbackfee (Jorge Timón)
Pull request description:
Before it was 0 by default for main and 20000 for test and regtest.
Now it is 0 by default for all chains, thus there's no need to call Params().
Also now the default for main is properly documented.
Suggestion for release notes:
-fallbackfee was 0 (disabled) by default for the main chain, but 20000 by default for the test chains. Now it is 0 by default for all chains. Testnet and regtest users will have to add fallbackfee=20000 to their configuration if they weren't setting it and they want it to keep working like before.
Should I propose them to the wiki for the release notes or only after merge?
For more context, see https://github.com/bitcoin/bitcoin/pull/16402#issuecomment-515701042
ACKs for top commit:
MarcoFalke:
ACK ea4cc3a7b36a9c77dbf0aff439da3ef0ea58e6e4
Tree-SHA512: fdfaba5d813da4221e405e0988bef44f3856d10f897a94f9614386d14b7716f4326ab8a6646e26d41ef3f4fa61b936191e216b1b605e9ab0520b0657fc162e6c
----
Co-Authored-By: UdjinM6 <UdjinM6@users.noreply.github.com>
## Issue being fixed or feature implemented
This was reported/requested by @HashEngineering:
> Older versions of our App won't sync due to if (obj.nVersion ==
BASIC_BLS_VERSION) . Older versions don't know what version a SML Entry
is. As such, they will never read the type field. On the android client
this causes an offset problem when reading the mnlistdiff and it will
throw an exception that bans the peer that supplied it. Soon enough, no
peers will be left to connect to because they will all give the android
client bad data.
## What was done?
With this PR, SML will serialise the new v19 fields (`nType`,
`platformHTTPPort`, `platformNodeID`) if the client's version is at
least equal to `70227`.
Note: Serialisation for hashing skips the above rule.
Also, functional test mininode protocol version is set to `70227`.
## How Has This Been Tested?
## Breaking Changes
## Checklist:
- [x] I have performed a self-review of my own code
- [ ] I have commented my code, particularly in hard-to-understand areas
- [x] I have added or updated relevant unit/integration/functional/e2e
tests
- [ ] I have made corresponding changes to the documentation
**For repository code-owners and collaborators only**
- [x] I have assigned this pull request to a milestone
## Issue being fixed or feature implemented
should hopefully fix some sporadic ci test failures (like
https://gitlab.com/dashpay/dash/-/jobs/4052206622#L1962)
## What was done?
tweaked dynamically_add/update functions to make checks more consistent
and avoid some edge cases, pls see individual commits
## How Has This Been Tested?
`feature_llmq_hpmn.py` and `feature_dip3_v19.py` still work locally,
let's see if ci is now (constantly) happy about these too...
## Breaking Changes
n/a
## Checklist:
- [x] I have performed a self-review of my own code
- [ ] I have commented my code, particularly in hard-to-understand areas
- [x] I have added or updated relevant unit/integration/functional/e2e
tests
- [ ] I have made corresponding changes to the documentation
**For repository code-owners and collaborators only**
- [x] I have assigned this pull request to a milestone
3ebde2143aa98af213872b98b474d904e55056f7 [test] Fix wait condition in disconnect_p2ps (Amiti Uttarwar)
Pull request description:
#19315 currently has a [test failure](https://cirrus-ci.com/task/4545582645641216) because of a race. `disconnect_p2ps` is intended to have a `wait_until` clause that prevents this race, but the conditional doesn't match since its comparing two different object types. `MY_SUBVERSION` is defined in messages.py as a byte string, but is compared to the value returned by the RPC. This PR simply converts types to ensure they match, which should prevent the race from occurring.
HUGE PROPS TO jnewbery for discovering the issue 🔎
ACKs for top commit:
jnewbery:
ACK 3ebde2143aa98af213872b98b474d904e55056f7
glozow:
Code review ACK 3ebde2143a
Tree-SHA512: ca096b80a3e4d757a645f38846d6dc89d6a3d35c3435513a72d278e305faddd4aff9e75a767941b51b2abbf59c82679bac1e9a0140d6f285efe3053e51bcc2a8
638441928a446726ce3a7fb20433a5478e7585bb test: add parameterized constructor for msg_sendcmpct() (Sebastian Falbesoner)
Pull request description:
While working on the test for #19776 I noticed that creating a `sendcmpct` message is quite cumbersome -- due to the lack of a parameterized constructor, one needs to create an empty (that is, initialized with default values) object and then set the two fields one by one. This PR replaces the default constructor with a parameterized constructor and uses it in the test `p2p_compactblocks.py`, reducing LOC. No need to pollute the namespace with temporary throw-away message objects anymore.
ACKs for top commit:
guggero:
Code review ACK 638441928a446726ce3a7fb20433a5478e7585bb.
epson121:
Code review ACK 638441928a446726ce3a7fb20433a5478e7585bb
Tree-SHA512: 3b58d276d714b73abc6cc98d1d52dec5f6026b33f03faaeb7dcbc5d83ac377555179f98b159b2b9ecc8957999c35a1dc082e3c69299c5fde4e35f1bd0587ce9d
854382885f18aa9a95cdde3d11591b05c305ad3f refactor: test: improve wait_for{header,merkleblock} interface (Sebastian Falbesoner)
1356a45ef042e7bd3d539fbb606d6b1be547d00f test: complete impl. of msg_merkleblock and wait_for_merkleblock (Sebastian Falbesoner)
Pull request description:
Implements the missing initialization/serialization methods for `msg_merkleblock`, based on the already present class `CMerkleBlock`. Also changes the method `wait_for_merkleblock()` to be more precise by waiting for a merkleblock with a specified blockhash instead of an arbitrary one.
In the BIP37 test `p2p_filter.py`, this new method is used to make the test of receiving merkleblock and tx if a filter is set to be more precise, by checking if they also arrive in the right order.
In the course of this PR, also the interface for the methods `wait_for_merkleblock()` and `wait_for_header()` are improved to take a hex string instead of an integer, which is more typesafe and less of a burden to the caller.
ACKs for top commit:
MarcoFalke:
ACK 854382885f18aa9a95cdde3d11591b05c305ad3f
Tree-SHA512: adaf0ac728ef0b9929cb417a7a7b4c1346c400b2d365bf6914515c67b6cfe8f4a7ecc62fb514afdce9792f0bed833416f6bca6b9620f3d5dcdc66e4d5b0b7ea3
## Issue being fixed or feature implemented
1. we need to move time forward to let invs being relayed
2. nNextInvSend in SendMessages can be bumped up to 30+ seconds into the
future in rare cases
make sure timeouts in tests are high enough to relay tx inv/wait for
corresponding islock
## What was done?
tl;dr: bump mocktime while waiting, wait longer
extracted fixes from https://github.com/dashpay/dash/pull/5288 but I
expect this to fix other sporadic test failures too
## How Has This Been Tested?
tests are ok locally and in https://github.com/dashpay/dash/pull/5288
## Breaking Changes
n/a
## Checklist:
- [x] I have performed a self-review of my own code
- [ ] I have commented my code, particularly in hard-to-understand areas
- [ ] I have added or updated relevant unit/integration/functional/e2e
tests
- [ ] I have made corresponding changes to the documentation
**For repository code-owners and collaborators only**
- [ ] I have assigned this pull request to a milestone
19a354b11f85a3c6c81ff83bf702bf7a40cf5046 Output a descriptor in createmultisig and addmultisigaddress (Andrew Chow)
Pull request description:
Give a descriptor from `createmultisig` and `addmultisigaddress`.
Extracted from #16528 with `addmultisgaddress` and tests added.
ACKs for top commit:
Sjors:
tACK 19a354b11f85a3c6c81ff83bf702bf7a40cf5046
MarcoFalke:
ACK 19a354b11f85a3c6c81ff83bf702bf7a40cf5046
promag:
Code review ACK 19a354b11f85a3c6c81ff83bf702bf7a40cf5046.
meshcollider:
utACK 19a354b11f85a3c6c81ff83bf702bf7a40cf5046
Tree-SHA512: e813125fbbc358ea8d45b1748de16a29a94efd83175b748fb8fa3b0bfc8e783ed36b6c554d84f5d4ead1ba252a83a3e937b6c3f75da7b8d3b4e55f94d6013771
e4f4ea47ebf7774fb6f445adde7bf7ea71fa05a1 lint: Catch use of [] or {} as default parameter values in Python functions (practicalswift)
25dd86715039586d92176eee16e9c6644d2547f0 Avoid using mutable default parameter values (practicalswift)
Pull request description:
Avoid common Python default parameter gotcha when mutable `dict`/`list`:s are used as default parameter values.
Examples of this gotcha caught during review:
* https://github.com/bitcoin/bitcoin/pull/16673#discussion_r317415261
* https://github.com/bitcoin/bitcoin/pull/14565#discussion_r241942304
Perhaps surprisingly this is how mutable list and dictionary default parameter values behave in Python:
```
>>> def f(i, j=[], k={}):
... j.append(i)
... k[i] = True
... return j, k
...
>>> f(1)
([1], {1: True})
>>> f(1)
([1, 1], {1: True})
>>> f(2)
([1, 1, 2], {1: True, 2: True})
```
In contrast to:
```
>>> def f(i, j=None, k=None):
... if j is None:
... j = []
... if k is None:
... k = {}
... j.append(i)
... k[i] = True
... return j, k
...
>>> f(1)
([1], {1: True})
>>> f(1)
([1], {1: True})
>>> f(2)
([2], {2: True})
```
The latter is typically the intended behaviour.
This PR fixes two instances of this and adds a check guarding against this gotcha going forward :-)
ACKs for top commit:
Sjors:
Oh Python... ACK e4f4ea47ebf7774fb6f445adde7bf7ea71fa05a1. Testing tip: swap the two commits.
Tree-SHA512: 56e14d24fc866211a20185c9fdb274ed046c3aed2dc0e07699e58b6f9fa3b79f6d0c880fb02d72b7fe5cc5eb7c0ff6da0ead33123344e1a872209370c2e49e3f
