💬 Sjors commented on pull request "mining: add early return to waitTipChanged()":
(https://github.com/bitcoin/bitcoin/pull/31297#issuecomment-2493619203)
See #31346.
(https://github.com/bitcoin/bitcoin/pull/31297#issuecomment-2493619203)
See #31346.
👍 rkrux approved a pull request: "contrib: skip missing binaries in gen-manpages"
(https://github.com/bitcoin/bitcoin/pull/30986#pullrequestreview-2454288885)
re-ACK 4bbd28baf33382231f4f1dab20681c05f9915af2
Minimal range diff.
<details>
<summary>git range-diff f080618...4bbd28b</summary>
```
1: 7060d64bf7 ! 1: 299e2220e9 gen-manpages: implement --skip-missing-binaries
@@ contrib/devtools/gen-manpages.py: BINARIES = [
+ "--skip-missing-binaries",
+ action="store_true",
+ default=False,
-+ help="skip generation for binaries that are not found",
++ help="skip generation for binaries that are no
...
(https://github.com/bitcoin/bitcoin/pull/30986#pullrequestreview-2454288885)
re-ACK 4bbd28baf33382231f4f1dab20681c05f9915af2
Minimal range diff.
<details>
<summary>git range-diff f080618...4bbd28b</summary>
```
1: 7060d64bf7 ! 1: 299e2220e9 gen-manpages: implement --skip-missing-binaries
@@ contrib/devtools/gen-manpages.py: BINARIES = [
+ "--skip-missing-binaries",
+ action="store_true",
+ default=False,
-+ help="skip generation for binaries that are not found",
++ help="skip generation for binaries that are no
...
💬 l0rinc commented on pull request "Set notifications m_tip_block in LoadChainTip()":
(https://github.com/bitcoin/bitcoin/pull/31346#discussion_r1853828160)
nit: at this point it may make more sense to just remove the `[[nodiscard]]` from `blockTip` instead
(https://github.com/bitcoin/bitcoin/pull/31346#discussion_r1853828160)
nit: at this point it may make more sense to just remove the `[[nodiscard]]` from `blockTip` instead
💬 l0rinc commented on pull request "util: Improve documentation and negation of args":
(https://github.com/bitcoin/bitcoin/pull/31212#discussion_r1853836054)
I'm not against that, my concern is just that reproducing old behavior before the fix isn't very useful activity - it's a lot more useful to see that a test fails before the fix with the correct error (i.e TDD), but that's not really possible currently - unless we add something like https://github.com/spockframework/spock/blob/master/spock-core/src/main/java/spock/lang/PendingFeature.java#L13-L17 (where we invert the test, making sure it fails (that's when the test is green) and remove the annot
...
(https://github.com/bitcoin/bitcoin/pull/31212#discussion_r1853836054)
I'm not against that, my concern is just that reproducing old behavior before the fix isn't very useful activity - it's a lot more useful to see that a test fails before the fix with the correct error (i.e TDD), but that's not really possible currently - unless we add something like https://github.com/spockframework/spock/blob/master/spock-core/src/main/java/spock/lang/PendingFeature.java#L13-L17 (where we invert the test, making sure it fails (that's when the test is green) and remove the annot
...
✅ Sjors closed an issue: "ci: how to run native arm job on Apple silicon?"
(https://github.com/bitcoin/bitcoin/issues/31344)
(https://github.com/bitcoin/bitcoin/issues/31344)
💬 Sjors commented on issue "ci: how to run native arm job on Apple silicon?":
(https://github.com/bitcoin/bitcoin/issues/31344#issuecomment-2493642543)
> It is not possible to run 32-bit arm binaries on a CPU that only supports 64-bit mode.
Ah, in that case it's probably pointless for a performance improvement.
> select qemu-arm in UTM. (Is there a reason why you haven't done this in the first place?)
My initial goal for that Ubuntu VM was to make Guix builds.
I could spin up a separate machine in 32 bit mode, but that seems like a world of pain on its own. Closing for now.
(https://github.com/bitcoin/bitcoin/issues/31344#issuecomment-2493642543)
> It is not possible to run 32-bit arm binaries on a CPU that only supports 64-bit mode.
Ah, in that case it's probably pointless for a performance improvement.
> select qemu-arm in UTM. (Is there a reason why you haven't done this in the first place?)
My initial goal for that Ubuntu VM was to make Guix builds.
I could spin up a separate machine in 32 bit mode, but that seems like a world of pain on its own. Closing for now.
💬 l0rinc commented on pull request "util: Improve documentation and negation of args":
(https://github.com/bitcoin/bitcoin/pull/31212#discussion_r1853838097)
what about `-noversion=0` :p
There has to be a better way, this is so confusing :/
(https://github.com/bitcoin/bitcoin/pull/31212#discussion_r1853838097)
what about `-noversion=0` :p
There has to be a better way, this is so confusing :/
💬 Sjors commented on issue "ci: how to run native arm job on Apple silicon?":
(https://github.com/bitcoin/bitcoin/issues/31344#issuecomment-2493644965)
I guess my confusion came from the use of `arm64` as the label for cirrus.
(https://github.com/bitcoin/bitcoin/issues/31344#issuecomment-2493644965)
I guess my confusion came from the use of `arm64` as the label for cirrus.
💬 Sjors commented on pull request "Set notifications m_tip_block in LoadChainTip()":
(https://github.com/bitcoin/bitcoin/pull/31346#discussion_r1853843127)
Maybe, or we could do something with the return value. I'm not familiar enough with init code and kernel notifications to have a strong opinion on this. cc @ryanofsky, @TheCharlatan
(https://github.com/bitcoin/bitcoin/pull/31346#discussion_r1853843127)
Maybe, or we could do something with the return value. I'm not familiar enough with init code and kernel notifications to have a strong opinion on this. cc @ryanofsky, @TheCharlatan
💬 maflcko commented on pull request "Set notifications m_tip_block in LoadChainTip()":
(https://github.com/bitcoin/bitcoin/pull/31346#issuecomment-2493675258)
This is incomplete. You'll have to update the docs as well: "If the tip was not connected on
* startup, this will wait.", and "which may
//! be true even long after startup, until shutdown.". Also, in init.cpp you can remove the if-guard?
(https://github.com/bitcoin/bitcoin/pull/31346#issuecomment-2493675258)
This is incomplete. You'll have to update the docs as well: "If the tip was not connected on
* startup, this will wait.", and "which may
//! be true even long after startup, until shutdown.". Also, in init.cpp you can remove the if-guard?
💬 orangesurf commented on pull request "miner: Reorg Testnet4 minimum difficulty blocks":
(https://github.com/bitcoin/bitcoin/pull/31117#issuecomment-2493677462)
A miner (earning coinbase rewards to [tb1q2dsc...](https://mempool.space/testnet4/address/tb1q2dsc94zq40nwnz27w5rxljwllutnwjtlxk44fz) is currently overwhelmingly dominant on testnet4.
Since block [000000000000000167286a4bea3bca60d6c3ab2cbbef79537bdf07b6420a2d2d / 54780](https://mempool.space/testnet4/block/000000000000000167286a4bea3bca60d6c3ab2cbbef79537bdf07b6420a2d2d) this miner has mined every bl
...
(https://github.com/bitcoin/bitcoin/pull/31117#issuecomment-2493677462)
A miner (earning coinbase rewards to [tb1q2dsc...](https://mempool.space/testnet4/address/tb1q2dsc94zq40nwnz27w5rxljwllutnwjtlxk44fz) is currently overwhelmingly dominant on testnet4.
Since block [000000000000000167286a4bea3bca60d6c3ab2cbbef79537bdf07b6420a2d2d / 54780](https://mempool.space/testnet4/block/000000000000000167286a4bea3bca60d6c3ab2cbbef79537bdf07b6420a2d2d) this miner has mined every bl
...
💬 maflcko commented on issue "ci: how to run native arm job on Apple silicon?":
(https://github.com/bitcoin/bitcoin/issues/31344#issuecomment-2493687553)
Yes, the cirrus machine is intentionally an aarch64 (also referred to as arm64) machine, because arm64 is generally available as performant hardware, compared to 32-bit arm. It was also intentionally selected to support 32-bit mode. (If you are looking for a cloud solution, IIRC only Hetzner cloud ships compatible VMs). Otherwise, you'll have to shop for compatible bare-metal yourself, or use qemu (or similar).
(https://github.com/bitcoin/bitcoin/issues/31344#issuecomment-2493687553)
Yes, the cirrus machine is intentionally an aarch64 (also referred to as arm64) machine, because arm64 is generally available as performant hardware, compared to 32-bit arm. It was also intentionally selected to support 32-bit mode. (If you are looking for a cloud solution, IIRC only Hetzner cloud ships compatible VMs). Otherwise, you'll have to shop for compatible bare-metal yourself, or use qemu (or similar).
⚠️ rkrux opened an issue: "Add support for creating v3 raw transactions in `createrawtransaction` RPC"
(https://github.com/bitcoin/bitcoin/issues/31348)
### Please describe the feature you'd like to see added.
Currently, `createrawtransaction` RPC creates only v2 raw transaction, i.e. the first byte of the serialised transaction hex is `02`. It'd be helpful for the RPC to conditionally create V3 raw transactions if such intent is passed in the arguments of the RPC call.
### Is your feature related to a problem, if so please describe it.
_No response_
### Describe the solution you'd like
- A new argument to the RPC call can be passed that al
...
(https://github.com/bitcoin/bitcoin/issues/31348)
### Please describe the feature you'd like to see added.
Currently, `createrawtransaction` RPC creates only v2 raw transaction, i.e. the first byte of the serialised transaction hex is `02`. It'd be helpful for the RPC to conditionally create V3 raw transactions if such intent is passed in the arguments of the RPC call.
### Is your feature related to a problem, if so please describe it.
_No response_
### Describe the solution you'd like
- A new argument to the RPC call can be passed that al
...
💬 ryanofsky commented on pull request "util: Improve documentation and negation of args":
(https://github.com/bitcoin/bitcoin/pull/31212#discussion_r1853870363)
> I'm not against that, my concern is just that reproducing old behavior before the fix isn't very useful activity
Strongly disagree :). In 2ca1460ae3a7217eaa8c5972515bf622bedadfce if tests had not been added *before* that commit checking for incorrect HTTP error codes, you would not know from looking at the change what the behavior was before the bugfix, or how new tests were connected to it. It is very useful to have test coverage for bugs before they are fixed.
> it's a lot more useful
...
(https://github.com/bitcoin/bitcoin/pull/31212#discussion_r1853870363)
> I'm not against that, my concern is just that reproducing old behavior before the fix isn't very useful activity
Strongly disagree :). In 2ca1460ae3a7217eaa8c5972515bf622bedadfce if tests had not been added *before* that commit checking for incorrect HTTP error codes, you would not know from looking at the change what the behavior was before the bugfix, or how new tests were connected to it. It is very useful to have test coverage for bugs before they are fixed.
> it's a lot more useful
...
✅ maflcko closed a pull request: "ci: Place datadirs for tests under tmpfs"
(https://github.com/bitcoin/bitcoin/pull/31182)
(https://github.com/bitcoin/bitcoin/pull/31182)
💬 maflcko commented on pull request "ci: Place datadirs for tests under tmpfs":
(https://github.com/bitcoin/bitcoin/pull/31182#issuecomment-2493700147)
I tested it on a Pi 4B, with 4GB of memory, Ubuntu 24.04, using 1 CPU thread: `CCACHE_MAXSIZE=1500M MAKEJOBS="-j1" FILE_ENV="./ci/test/00_setup_env_native_nowallet_libbitcoinkernel.sh" /usr/bin/time --format='%U+%S %e (%E)' ./ci/test_run_all.sh`. However, I couldn't find any difference either. Moreover, with `MAKEJOBS="-j4"`, this pull performed worse than the commit prior to it (55min vs 1h5min).
Not sure why, but I guess modern Linux will cache the file IO sufficiently to not make a differe
...
(https://github.com/bitcoin/bitcoin/pull/31182#issuecomment-2493700147)
I tested it on a Pi 4B, with 4GB of memory, Ubuntu 24.04, using 1 CPU thread: `CCACHE_MAXSIZE=1500M MAKEJOBS="-j1" FILE_ENV="./ci/test/00_setup_env_native_nowallet_libbitcoinkernel.sh" /usr/bin/time --format='%U+%S %e (%E)' ./ci/test_run_all.sh`. However, I couldn't find any difference either. Moreover, with `MAKEJOBS="-j4"`, this pull performed worse than the commit prior to it (55min vs 1h5min).
Not sure why, but I guess modern Linux will cache the file IO sufficiently to not make a differe
...
🤔 glozow reviewed a pull request: "Package validation: accept packages of size 1"
(https://github.com/bitcoin/bitcoin/pull/31096#pullrequestreview-2454400142)
I still don't really like that the `IsChildWithParents()` function returns true for a tx that isn't a child / doesn't have any parents. Also dislike that we run the whole input-loading thing for the child-with-unconfirmed-parents check even for a singleton tx. I guess one way around this is to rename the 2 `IsChildWithParents` functions to something along the lines of "is topology we can handle" and to add a `package.size() > 1` gate to the child-with-unconfirmed-parents check.
I spent some t
...
(https://github.com/bitcoin/bitcoin/pull/31096#pullrequestreview-2454400142)
I still don't really like that the `IsChildWithParents()` function returns true for a tx that isn't a child / doesn't have any parents. Also dislike that we run the whole input-loading thing for the child-with-unconfirmed-parents check even for a singleton tx. I guess one way around this is to rename the 2 `IsChildWithParents` functions to something along the lines of "is topology we can handle" and to add a `package.size() > 1` gate to the child-with-unconfirmed-parents check.
I spent some t
...
💬 l0rinc commented on pull request "util: Improve documentation and negation of args":
(https://github.com/bitcoin/bitcoin/pull/31212#discussion_r1853891120)
> if tests had not been added before that commit checking for incorrect HTTP error codes, you would not know from looking at the change what the behavior was before the bugfix
The test would fail before the fix with the exact reason, not sure we're in a disagreement here.
> If the bugfix commit show a line of test code changing from [...]
> if the bugfix commit just included an entirely new [...]
But my point was exactly that if there's a failing test in the commit before the fix, whic
...
(https://github.com/bitcoin/bitcoin/pull/31212#discussion_r1853891120)
> if tests had not been added before that commit checking for incorrect HTTP error codes, you would not know from looking at the change what the behavior was before the bugfix
The test would fail before the fix with the exact reason, not sure we're in a disagreement here.
> If the bugfix commit show a line of test code changing from [...]
> if the bugfix commit just included an entirely new [...]
But my point was exactly that if there's a failing test in the commit before the fix, whic
...
💬 fanquake commented on pull request "build: increase minimum supported Windows to 10.0":
(https://github.com/bitcoin/bitcoin/pull/31172#issuecomment-2493742473)
Guix Build:
```bash
606d0909c4591fc7dac3759e230e7bd3de00555c1c535d437ca8bc19df85fc70 guix-build-ee1128ead846/output/aarch64-linux-gnu/SHA256SUMS.part
4aca1c476b6824d485044c6636ce2ecf45f542a89bc501493f4868cf16c13d1f guix-build-ee1128ead846/output/aarch64-linux-gnu/bitcoin-ee1128ead846-aarch64-linux-gnu-debug.tar.gz
46d17a50226b60af12124b9c2b70cdab1c257a2034463999c0340d8a8b573cdd guix-build-ee1128ead846/output/aarch64-linux-gnu/bitcoin-ee1128ead846-aarch64-linux-gnu.tar.gz
d82c381c8d4e888e
...
(https://github.com/bitcoin/bitcoin/pull/31172#issuecomment-2493742473)
Guix Build:
```bash
606d0909c4591fc7dac3759e230e7bd3de00555c1c535d437ca8bc19df85fc70 guix-build-ee1128ead846/output/aarch64-linux-gnu/SHA256SUMS.part
4aca1c476b6824d485044c6636ce2ecf45f542a89bc501493f4868cf16c13d1f guix-build-ee1128ead846/output/aarch64-linux-gnu/bitcoin-ee1128ead846-aarch64-linux-gnu-debug.tar.gz
46d17a50226b60af12124b9c2b70cdab1c257a2034463999c0340d8a8b573cdd guix-build-ee1128ead846/output/aarch64-linux-gnu/bitcoin-ee1128ead846-aarch64-linux-gnu.tar.gz
d82c381c8d4e888e
...
💬 maflcko commented on pull request "util: Improve documentation and negation of args":
(https://github.com/bitcoin/bitcoin/pull/31212#discussion_r1853899185)
> > if tests had not been added before that commit checking for incorrect HTTP error codes, you would not know from looking at the change what the behavior was before the bugfix
>
> The test would fail before the fix with the exact reason, not sure we're in a disagreement here.
The test would be reproducing the bug, thus pass. The next commit would fix the pre-existing bug in the code and the test, and would thus pass as well.
(https://github.com/bitcoin/bitcoin/pull/31212#discussion_r1853899185)
> > if tests had not been added before that commit checking for incorrect HTTP error codes, you would not know from looking at the change what the behavior was before the bugfix
>
> The test would fail before the fix with the exact reason, not sure we're in a disagreement here.
The test would be reproducing the bug, thus pass. The next commit would fix the pre-existing bug in the code and the test, and would thus pass as well.