chore(deps): bump rama-core from 0.3.0-alpha.4 to 0.3.0 in /codex-rs

[dependabot]

Bumps rama-core from 0.3.0-alpha.4 to 0.3.0.

Release notes

Sourced from rama-core's releases.

rama-0.3.0

Rama 0.3.0 is here. This is a big release: the 0.3 series started with the first alpha on 2025-07-07 and carried a long line of protocol work, API redesign, platform hardening, and real-world proxy features all the way to this final release.

For people coming from 0.2: expect breaking changes. The central Context model is gone, service inputs and outputs have been generalized, extensions are now the main way to carry request, response, connection and transport metadata, and Rama owns more of its HTTP and URI surface directly. The migration cost is real, but it buys a cleaner service model, stronger protocol fidelity, better proxy ergonomics, and much more room for the next release trains.

This also closes the long alpha era. From here on we aim to ship regular Rama releases on a 2 to 8 week cadence. Versioning will continue to follow SemVer semantics.

Community

Thank you to everyone who contributed code, reviews, documentation, testing, bug reports, ideas, and production feedback during the 0.3 cycle. This release includes work from Glen De Cauwsemaecker, Brecht Stamper, Abdelkader Boudih, Nicolas Trippar, M-Kusumgar, Ali Tariq, Camille Louédoc-Eyriès, sim-hash, Irfan - ئىرفان, Yukun Wang, hafihaf123, Aydan Pirani, Kenny Lau, MeerKatDev, Maarten Deprez, Dominic Lindsay, Xavier Lambein, Stijn De Clercq, Shabbir Hasan, Antoine Bernardeau, Darshil Patel, Gautham Venkataraman, FS, Elias, Azzam S.A, Nikita, Ian Wagner, bitterpanda, Elizabeth Gonzales Belsuzarri, 0x676e67, KoHcoJlb, and everyone else who helped shape the release.

Thank you also to our GitHub Sponsors for directly funding Rama development, and to our commercial partners for funding a significant part of the work in this cycle. If you want to support Rama, you can become a GitHub Sponsor. If your organisation is looking for a long-term partner around proxies, protocol work, support, training, or custom integrations, see ramaproxy.com.

Alpha Train Recap

• 0.3.0-alpha.1 (2025-07-07) opened the cycle with the first versions of rama-ws, rama-socks5, rama-unix, built-in Datastar support, SSE work, richer observability, PeetPrint and TLS fingerprinting improvements, protocol peek routing, and a much larger proxy example set.
• 0.3.0-alpha.2 (2025-08-05) added ACME support, the first rama-crypto crate with JOSE/JWK/JWA/JWS support, connection pooling, TCP connector pools, WebSocket fingerprinting support, target HTTP version enforcement, and the first anti-bot examples.
• 0.3.0-alpha.3 (2025-08-29) promoted Windows to tier-1, added signed Windows CLI releases and winget packaging, custom X.509 stores, WebSocket extensions, HTTP-to-HTTPS upgrade redirects, include-dir serving, router quality-of-life improvements, and the unified HTTP/SOCKS proxy

... (truncated)

Changelog

Sourced from rama-core's changelog.

0.3.0

Planned release date: 2026-07-07

Rama 0.3.0 is here. This is a big release: the 0.3 series started with the first alpha on 2025-07-07 and carried a long line of protocol work, API redesign, platform hardening, and real-world proxy features all the way to this final release.

For people coming from 0.2: expect breaking changes. The central Context model is gone, service inputs and outputs have been generalized, extensions are now the main way to carry request, response, connection and transport metadata, and Rama owns more of its HTTP and URI surface directly. The migration cost is real, but it buys a cleaner service model, stronger protocol fidelity, better proxy ergonomics, and much more room for the next release trains.

This also closes the long alpha era. From here on we aim to ship regular Rama releases on a 2 to 8 week cadence. Versioning will continue to follow SemVer semantics.

Community

Thank you to everyone who contributed code, reviews, documentation, testing, bug reports, ideas, and production feedback during the 0.3 cycle. This release includes work from Glen De Cauwsemaecker, Brecht Stamper, Abdelkader Boudih, Nicolas Trippar, M-Kusumgar, Ali Tariq, Camille Louédoc-Eyriès, sim-hash, Irfan - ئىرفان, Yukun Wang, hafihaf123, Aydan Pirani, Kenny Lau, MeerKatDev, Maarten Deprez, Dominic Lindsay, Xavier Lambein, Stijn De Clercq, Shabbir Hasan, Antoine Bernardeau, Darshil Patel, Gautham Venkataraman, FS, Elias, Azzam S.A, Nikita, Ian Wagner, bitterpanda, Elizabeth Gonzales Belsuzarri, 0x676e67, KoHcoJlb, and everyone else who helped shape the release.

Thank you also to our GitHub Sponsors for directly funding Rama development, and to our commercial partners for funding a significant part of the work in this cycle. If you want to support Rama, you can become a GitHub Sponsor. If your organisation is looking for a long-term partner around proxies, protocol work, support, training, or custom integrations, see ramaproxy.com.

Alpha Train Recap

• 0.3.0-alpha.1 (2025-07-07) opened the cycle with the first versions of rama-ws, rama-socks5, rama-unix, built-in Datastar support, SSE work, richer observability, PeetPrint and TLS fingerprinting improvements, protocol peek routing, and a much larger proxy example set.
• 0.3.0-alpha.2 (2025-08-05) added ACME support, the first rama-crypto crate with JOSE/JWK/JWA/JWS support, connection pooling, TCP connector pools, WebSocket fingerprinting support, target HTTP version enforcement, and the first anti-bot examples.
• 0.3.0-alpha.3 (2025-08-29) promoted Windows to tier-1, added signed

... (truncated)

Commits

• 1e094e4 add docsrs-check (just + daily ci) + 0.3.0 notes touch-ups
• 8ef735d update deps (includes crossbeam security fix)
• 4f48d82 fix no_std dep leaks in utils/net/core + guard w/ check-nostd
• 4040ff5 enable serde feature in ipnet (rama-net)
• 958566a ci: clippy --fix auto-fixes (2026-07-06) (#1060)
• 37bc64c apple xpc + ne tproxy lifecycle fixes, uri query utils (#1059)
• 3f6754b update deps + embedded profiles (20260706)
• 97a8255 ne tproxy: stop callable from async contexts without fault
• 67b4abe ne tproxy: engine stop bounded even on a wedged runtime
• aa95355 uri: drop Uri PartialEq<str> impls; compare as_str instead
• Additional commits viewable in compare view

---

^{Need help on this PR? Tag @codesmith-bot with what you need. Autofix is disabled.}

[andrei-hasna]

[REVIEW] NO_GO — #565 @ 91f267b — lens: correctness+security+gates, reviewer unresolved-account002 (1 of 1)

Reviewed exact candidate

• Base: origin/main at 2fd98cf.
• Head: 91f267b.
• git log --oneline origin/main..HEAD: exit 0; 1 commit.
• git diff origin/main...HEAD --stat: exit 0; 2 files, 83 insertions, 60 deletions.
• Read the complete diff for codex-rs/Cargo.lock and codex-rs/network-proxy/Cargo.toml, plus the affected network-proxy Rama imports/call sites and the resolved Rama/Tokio lockfile graph.

Commands and declared gates

• bun install: exit 0; setup only, 521 packages installed. This is not the repository test result.
• The root package.json declares no test script. Therefore this repository declares no test gate; pass/fail count: 0/0 because no such gate exists.
• The root package.json declares no typecheck script. No typecheck gate was invented or run.
• No lane-chosen local Rust build/test command was run; this repository routes codex-rs build/test verification through its forge/Testbox lanes.
• gh pr view 565 --repo hasna/codewith --json statusCheckRollup: exit 0; 27 checks read: 9 SUCCESS, 11 FAILURE, 4 IN_PROGRESS, 3 SKIPPED.
• gh api repos/hasna/codewith/actions/jobs/94021151491/logs: exit 0; the failed Linux CLI build log records Cargo exit 101 after 14 compiler errors.
• gh api repos/hasna/codewith/actions/jobs/94021151510/logs: exit 0; the failed SDK build independently records the same Cargo exit 101 after 14 compiler errors.

Forge checks read

• Bazel test on ubuntu-24.04 for x86_64-unknown-linux-gnu: FAILURE
• Build Codewith Linux CLI: FAILURE
• cla: SUCCESS
• Check for spelling errors: SUCCESS
• Blob size policy: SUCCESS
• cargo-deny: SUCCESS
• build-test: SUCCESS
• Detect changed areas: SUCCESS
• python-sdk: SUCCESS
• Bazel test on ubuntu-24.04 for x86_64-unknown-linux-musl: IN_PROGRESS
• Bazel test on windows-latest for x86_64-pc-windows-gnullvm shard 1/4: FAILURE
• sdks: FAILURE
• Format / etc: SUCCESS
• Bazel test on windows-latest for x86_64-pc-windows-gnullvm shard 2/4: FAILURE
• Bazel test on windows-latest for x86_64-pc-windows-gnullvm shard 3/4: FAILURE
• Bazel test on windows-latest for x86_64-pc-windows-gnullvm shard 4/4: FAILURE
• Bazel test on windows-latest for x86_64-pc-windows-gnullvm (native main): SKIPPED
• cargo shear: SUCCESS
• Bazel clippy on ubuntu-24.04 for x86_64-unknown-linux-gnu: IN_PROGRESS
• Argument comment lint package: SKIPPED
• Bazel clippy on windows-latest for x86_64-pc-windows-gnullvm: FAILURE
• Verify release build on ubuntu-24.04 for x86_64-unknown-linux-gnu: IN_PROGRESS
• Argument comment lint - Linux: IN_PROGRESS
• Verify release build on windows-latest for x86_64-pc-windows-gnullvm: FAILURE
• Argument comment lint - Windows: FAILURE
• Bazel test on windows-latest for x86_64-pc-windows-gnullvm: FAILURE
• [code]smith: SKIPPED

Blocking P0/P1 findings

1. P1 — the dependency graph is source-incompatible and fails the required build path. The PR pins only rama-core to 0.3.0 while rama-http-types and the rest of the direct Rama family remain at 0.3.0-alpha.4. Both independent Cargo build jobs report the actual lines error[E0432]: unresolved import rama_error::OpaqueError and error[E0432]: unresolved import rama_core::extensions::ExtensionsMut, followed by error: could not compile rama-http-types (lib) due to 14 previous errors and exit 101. This is currently reachable on every Codewith build that compiles network-proxy.
2. P1 — the repository-required Bazel dependency lock gate is broken. Check MODULE.bazel.lock is up to date concluded FAILURE, and this dependency-changing PR does not update MODULE.bazel.lock.

Required remedy
Upgrade the Rama dependency family as one compatible set (or retain the prior all-alpha set), adapt network-proxy call sites for the chosen APIs, regenerate Cargo.lock and MODULE.bazel.lock, and rerun the affected Cargo/SDK/Bazel build lanes. This is real compatibility work, not a one-line lockfile fix, so the PR remains open.

Non-blocking follow-ups

• The root package declares neither a test gate nor a typecheck gate. That absence is recorded as a repository-level gate gap, but it is pre-existing and is not the reason for this NO_GO.
• No additional security or correctness P0/P1 finding was found in the reviewed diff.

[andrei-hasna]

[REVIEW] NO_GO — #565 @ 77d3220 — lens: correctness+security+gates, reviewer unresolved-account011 (1 of 1)

Reviewed exact candidate

• Head: 77d32209b8731785e785f2b8ef909c5a9ecddda7; fetched base: origin/main at 0f5990beb98f0e7a2c2a49201a1f830144b5f9af.
• git log --oneline origin/main..HEAD, git diff origin/main...HEAD --stat, and the complete diff of codex-rs/Cargo.lock plus codex-rs/network-proxy/Cargo.toml: exit 0.
• Read surrounding dependency and call-site context in the network proxy manifest, lock graph, Bazel target/workflow/lock rules, and the rama_core service boundaries in connect_policy.rs, upstream.rs, http_proxy.rs, mitm.rs, and socks5.rs.
• Change size: 61 insertions, 38 deletions; dependency-only and below the review-size threshold.

Commands and declared gates

• bun install: exit 0; 521 packages installed. Setup only, not a test gate. It generated an untracked bun.lock, which I removed; the worktree is clean.
• package.json declares no typecheck script.
• package.json declares no test script. This repository therefore declares no package test gate; I did not invent one.
• No lane-chosen build or test command was run. Repository policy routes Rust/Bazel validation to remote CI.
• gh pr view 565 --repo hasna/codewith --json statusCheckRollup --jq '.statusCheckRollup | map(if .__typename == "CheckRun" then {name: .name, status: .status, conclusion: .conclusion} else {name: .context, status: .state, conclusion: .state} end)': exit 0. Rollup count: 8 SUCCESS, 9 FAILURE, 7 IN_PROGRESS, 3 SKIPPED.

Forge checks read by name

Check	Conclusion
Bazel test on ubuntu-24.04 for x86_64-unknown-linux-gnu	FAILURE
cargo-deny	FAILURE
Build Codewith Linux CLI	FAILURE
cla	SUCCESS
Check for spelling errors	SUCCESS
Blob size policy	SUCCESS
build-test	SUCCESS
Detect changed areas	SUCCESS
python-sdk	SUCCESS
Bazel test on ubuntu-24.04 for x86_64-unknown-linux-musl	IN_PROGRESS
Bazel test on windows-latest for x86_64-pc-windows-gnullvm shard 1/4	FAILURE
sdks	FAILURE
Format / etc	SUCCESS
Bazel test on windows-latest for x86_64-pc-windows-gnullvm shard 2/4	FAILURE
Bazel test on windows-latest for x86_64-pc-windows-gnullvm shard 3/4	FAILURE
Bazel test on windows-latest for x86_64-pc-windows-gnullvm shard 4/4	FAILURE
Bazel test on windows-latest for x86_64-pc-windows-gnullvm (native main)	SKIPPED
cargo shear	SUCCESS
Bazel clippy on ubuntu-24.04 for x86_64-unknown-linux-gnu	IN_PROGRESS
Argument comment lint package	SKIPPED
Bazel clippy on windows-latest for x86_64-pc-windows-gnullvm	IN_PROGRESS
Verify release build on ubuntu-24.04 for x86_64-unknown-linux-gnu	IN_PROGRESS
Argument comment lint - Linux	IN_PROGRESS
Verify release build on windows-latest for x86_64-pc-windows-gnullvm	IN_PROGRESS
Argument comment lint - Windows	IN_PROGRESS
Bazel test on windows-latest for x86_64-pc-windows-gnullvm	FAILURE
[code]smith	SKIPPED

Blocking P0/P1 findings

1. P1 required-gate failure: codex-rs/network-proxy/Cargo.toml and codex-rs/Cargo.lock move rama-core to 0.3.0, but MODULE.bazel.lock still contains rama-core_0.3.0-alpha.4. The authoritative Linux Bazel job failed at the named step Check MODULE.bazel.lock is up to date, before bazel test //... could run. This is the repository's explicit dependency-lock gate, not an invented lane command.
2. P1 required build/security gates are red at this exact head: cargo-deny, Build Codewith Linux CLI, sdks at Build Codewith with Cargo, all four Windows Bazel test shards, and their Windows aggregate. Seven other checks remain in progress. The candidate is not mergeable while these required gates fail or have not completed.

Required remedy

Regenerate and commit MODULE.bazel.lock from this dependency graph, resolve the current Cargo/cargo-deny/Windows Bazel failures, and rerun the current-head forge matrix. This is cross-platform release-gate work rather than a bounded one-line reviewer fix, so the PR remains open.

Non-blocking follow-ups

• The diff contains no direct secret exposure, unsafe mutation, or independently reachable security regression found by this review.
• The absent test and typecheck package scripts are recorded as the repository's declared-gate gap; no undeclared command is being used as a blocker.