Bump MSRV to 1.85.0#1046
Conversation
c038223 to
2305f34
Compare
jaoleal
left a comment
There was a problem hiding this comment.
All those formatting commits, it could be squashed
|
|
Tip to reviwers: to check those it should show this diff: This means the rest was generated by |
|
If you guys prefer, I can squash the formatting commits now. Just let me know. |
I think its better to review, a single big commit that the changes are only formatting. |
8d97f14 to
b2249eb
Compare
jaoleal
left a comment
There was a problem hiding this comment.
Just a question, i think everything is good
b2249eb to
43515de
Compare
| ServiceFlags::NONE, | ||
| port, | ||
| rand::random(), | ||
| rand::random::<u64>() as usize, |
There was a problem hiding this comment.
I went on a search, it looks like this cast can lead to data-loss on 32-bit devices. Is this something to be worried ?
There was a problem hiding this comment.
I don't think this is a problem. These random numbers are just random ID's, and the loss doesn't really matter, as this value is only used internally.
I could always just replace them with u32's, but I think it's best to drop those usizes from the structs themselves in favor of u32 or u64. Then we can drop the casts completely.
There was a problem hiding this comment.
This can be done on a follow up, as proposed by yourself on #1014
43515de to
a0d7e24
Compare
|
cc @moisesPompilio this will be needed for your BDK PR |
|
Rebased needed |
- Bump the workspace MSRV to 1.85.0 - Add `edition = 2024` to the workspace's `Cargo.toml`
Rust 1.85.0 is stricter when it comes to `unsafe`s. Before, unsafe functions could be called from within another functions without wrapping them in an `unsafe` block. Now, `rustc` 1.85.0 requires that we wrap each call of an unsafe function in an `unsafe` block.
…test_gbt_with_conflict` Add the missing `ConflictingTransaction` to the `did_conflict` arm. This test used to pass pass due to how `rand` v0.8 built the transactions with the `21` seed. Since `rand`'s RNG behavior is unstable across minor versions, it now builds a conflicting transaction, which causes a `ConflictingTransaction` error.
…cpkg.json` I'm guessing that GitHub very recently changed something about their `windows-latest` CI runner, which broke Boost installation on the Windows cross-testing job. Fix it by pinning `vcpkg` to the version declared on `rust-bitcoinkernel` v0.2.0 vendored `libbitcoinkernel-sys` `vcpkg.json` (120deac3062162151622ca4860575a33844ba10b), which can be found at: https://github.com/sedited/rust-bitcoinkernel/blob/0eadd82ba2c0bdb07b8321f4da5636ff502c0ea6/libbitcoinkernel-sys/bitcoin/vcpkg.json
3a7df87 to
4cc3d44
Compare
|
@JoseSK999 rebased |
Closes #1010
This PR bumps MSRV to 1.85.0 in preparation for using
bdk_walletas the wallet backend forfloresta-watch-only[#950].Changelog
I guess MSFT recently changed something about the Windows runner, which broke Boost installation due to out-of-sync
vcpkgversions. This is fixed by pinningvcpkginstallation using the version thatrust-bitcoinkernel's vendoredlibbitcoinkernel-sysuses.The version hash can be found at https://github.com/sedited/rust-bitcoinkernel/blob/0eadd82ba2c0bdb07b8321f4da5636ff502c0ea6/libbitcoinkernel-sys/bitcoin/vcpkg.json#L4