| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
|
| |
See #2060.
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
|
| |
With the removal of the hard `rustls` dependency and preferring
`futures-rustls`, then `paste` crate gets downgraded to a too-low of a version
so we should update it for our code to be able to compile correctly.
|
| |
|
|
| |
Nightly clippy now suggests this.
|
| |
|
|
|
| |
The new `cargo-sort` version doesn't like the whitespace we're using to
logically group the workspace members.
|
| | |
|
| | |
|
| |
|
|
|
| |
Since it's using 1.86 now, there's no reason to downgrade these
packages.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
After upgrading to rust:1.86.0-alpine3.21, we got a new error:
```
osxcross: error: cannot find libc++ headers
```
To fix this, we change the linker to
`x86_64-apple-darwin16-clang++-stdc++`.
This invokes some osxcross magic that tells it to find the C++
headers in the stdc++ location... and possibly more.
|
| |
|
|
|
|
|
|
|
|
| |
This is sufficient to stop the osxcross errors
``` error: no member named 'lseek64' in the global namespace ```
that we encountered on our earlier attempt to increase our MSRV.
Closes #1335.
|
| |\
| |
| |
| |
| | |
shadow CI tweaks
See merge request tpo/core/arti!2989
|
| | |
| |
| |
| |
| | |
It probably makes sense to test against a version of c-tor that we
still support.
|
| |\ \
| |/
|/|
| |
| |
| |
| | |
Enable --features=full on reproducible builds.
Closes #1964
See merge request tpo/core/arti!2956
|
| | |
| |
| |
| |
| | |
This is a workaround for
https://github.com/eira-fransham/crunchy/pull/17
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
We generally recommend that people build with --features=full IRL,
so this better matches what we're hoping they come up with.
Additionally, it will help cases like #1964, where packagers would
like to use reproducible builds as inputs to their own packages.
Closes #1964
|
| |\ \
| | |
| | |
| | |
| | | |
Fix typos
See merge request tpo/core/arti!2961
|
| | | |
| | |
| | |
| | | |
Typos found with codespell
|
| |\ \ \
| | | |
| | | |
| | | |
| | | | |
maint: Remove obsolete cargo_audit warnings.
See merge request tpo/core/arti!2973
|
| | | | | |
|
| |\ \ \ \
| | | | |
| | | | |
| | | | |
| | | | | |
maint: Update check_license exceptions.
See merge request tpo/core/arti!2974
|
| | |/ / /
| | | |
| | | |
| | | |
| | | | |
webpki have fixed their situation, and tinystr is now Unicode-3.0
licensed.
|
| |\ \ \ \
| |/ / /
|/| | |
| | | |
| | | |
| | | |
| | | | |
maint: Check for wildcard requirements
Closes #1918
See merge request tpo/core/arti!2954
|
| | |/ /
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
This commit adds a script that checks all Cargo.toml files whether a
version requirement contains a wildcard, that is, a version containing a
"*" preceded by an arbitrary amount of digits and dots.
Fixes #1918
|
| | | |
| | |
| | |
| | |
| | | |
(This is going to be a _requirement_,
since rand 0.9.1 has a behavioral change from 0.9.0)
|
| | | |
| | |
| | |
| | |
| | | |
This seems to be compatible with our licenses, and is used by the webpki
crates.
|
| | |/
|/|
| |
| |
| | |
I updated everything except rand, because the new version of rand
interacts with #1903, thus requiring more care.
|
| |/
|
|
|
|
|
|
|
|
|
|
| |
(Previously, it was optimized for size, leading to problems like
\#1336.)
For any purposes that need the old optimize-for-size behavior,
I've added a new "release-small" target. I've also moved the
"strip=debuginfo" behavior from maint/binary_size to this new target,
since cargo started supporting "strip" in 1.59.
Closes #1954.
|
| |
|
|
|
|
|
|
|
| |
This doesn't need to be perfect, but we need to use it to see
if a consensus is new enough that we should obey its
recommendations.
This commit also adds a script to update or check our release date,
and calls this script from our cargo-publish script.
|
| |
|
|
| |
Fixes #1519
|
| |
|
|
|
| |
This is a workaround for
https://gitlab.torproject.org/tpo/core/arti/-/issues/1335
|
| | |
|
| | |
|
| |
|
|
|
| |
(This only needs to be used for DSA keys from ssh_key, which
we should never touch in production.)
|
| |
|
|
| |
- The rand crate's "getrandom" feature has been renamed to "os_rng"
|
| |
|
|
|
|
| |
(Per discussion at #1774, we think the changes are acceptable.)
This commit won't compile on its own; subsequent commits will fix it.
|
| |
|
|
| |
Crate humantime is no longer unmaintained
|
| | |
|
| | |
|
| |\
| |
| |
| |
| |
| |
| | |
maint: Make matrix-check run in parallel in CI.
Closes #1625
See merge request tpo/core/arti!2835
|
| | | |
|
| | | |
|
| | | |
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
We should ideally do this for every-crate as well, but since that's
written in bash, it's a little more annoying to do. The options there
are either to rewrite it in python (probably good to do anyways) or to
modify the list_crates script to move the parallelization in there
(useful since it lets us reuse code, but doing that with a generator is
likely not possible). For now I'm just going to file a ticket to do that
and go back to working on stuff i actually should be doing :)
Fixes: #1625
|
| | | |
|
| | | |
|
| | |
| |
| |
| |
| | |
This is the version that introduces `root_hint_subjects()`,
which we want our ServerCertVerifier to override.
|
| |/ |
|
| |
|
|
|
| |
This is a workaround for
https://gitlab.torproject.org/tpo/core/arti/-/issues/1335
|