fix: drop ring from TLS path; bump aarch64 wheel base to manylinux_2_28#277
Merged
Conversation
The aarch64 linux-cross job built wheels in the CentOS 7 / GCC 4.8 toolchain that ships with manylinux2014. Switching to manylinux_2_28 (AlmaLinux 8, GCC 8+) gives the build a newer assembler, which avoids a class of miscompilations in crates that ship hand-written aarch64 asm.
reqwest's rustls-tls feature implicitly enables rustls/ring, which pulls ring v0.17 into the TLS path. ring's aarch64 build produces TLS signature-verify failures on some glibc builds of our wheels. Swap every reqwest TLS feature to the *-no-provider variant and install aws-lc-rs as the process-wide default CryptoProvider at CLI startup. The install is also called at the tower-uv reqwest entry point so its standalone unit test (which bypasses App::new) has a provider installed before the first TLS handshake. After this change, `cargo tree -i ring` is empty.
|
Important Review skippedAuto reviews are disabled on base/target branches other than the default branch. Please check the settings in the CodeRabbit UI or the ⚙️ Run configurationConfiguration used: Path: .coderabbit.yaml Review profile: CHILL Plan: Pro Run ID: You can disable this status message by setting the Use the checkbox below for a quick retry:
✨ Finishing Touches🧪 Generate unit tests (beta)
Comment |
scripts/rust-client-templates/Cargo.mustache hardcodes `rustls-tls` in the generated reqwest dep. Patch the template so the next generate-rust-api-client.sh run preserves the aws-lc-rs feature choice instead of reverting tower-api's Cargo.toml. The CFLAGS_aarch64_unknown_linux_gnu workaround in build-binaries.yml existed only to coax ring's build script into detecting ARMv8. With ring gone from the dep tree, the workaround is dead.
bradhe
approved these changes
May 14, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
tower loginfails insidepython:3.11-slim-bookwormon aarch64. The handshake to api.tower.dev errors out. Same wheel works on Alpine and on x86_64.Cause: rustls uses ring as its crypto provider (because
reqwest/rustls-tlspulls it in), and ring's aarch64 asm gets miscompiled by manylinux2014's toolchain (CentOS 7, GCC 4.8) for glibc 2.36+.Two changes, either one would fix it:
manylinux2014tomanylinux_2_28(AlmaLinux 8, GCC 8+).reqwest/rustls-tls*feature for the-no-providervariant and installaws_lc_rs::default_provider()at startup.Also patches
scripts/rust-client-templates/Cargo.mustacheso the next regen oftower-apikeeps the aws-lc-rs feature, and removes the now-deadCFLAGS_aarch64_unknown_linux_gnuring workaround from the workflow.Repro
Before, on an arm64 host:
After, wheel built from this branch in the same container: