)]}'
{
  "log": [
    {
      "commit": "a5035c766cea5240708ef091ae24f8a4ca11bc9e",
      "tree": "99f18e7879fc41ff8b6712b1b3f4ef4f9775923e",
      "parents": [
        "73c655bd621684429675732aaacb330241f7ff7a"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Fri Nov 13 23:58:29 2026 -0800"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Fri Nov 13 23:58:29 2026 -0800"
      },
      "message": "[SPARK-61599] Remove the Export Control Notice copied from Apache Spark\n\n### What changes were proposed in this pull request?\n\nThis removes the Export Control Notice from `NOTICE` and `NOTICE-binary`. The rest of both files stays the same.\n\n### Why are the changes needed?\n\n`NOTICE` was copied from apache/spark when the code was imported, and Spark\u0027s Export Control Notice came along with it. The notice says the U.S. Bureau of Industry and Security (BIS) has classified the software as ECCN 5D002.C.1 and that it is eligible for License Exception TSU. That has not happened for the gateway. It has not gone through the ASF export process (https://infra.apache.org/crypto.html): it is not listed on https://www.apache.org/licenses/exports/, and BIS has not been notified for it.\n\nWith this change the gateway matches spark-kubernetes-operator. The operator\u0027s container image also bundles cryptographic code (Tink, nimbus-jose-jwt, commons-crypto, and a JRE with the JDK\u0027s crypto providers), and its `NOTICE` has no Export Control Notice of its own.\n\nRemoving the notice does not answer whether the gateway needs to go through the export process. That question is for the PMC, on dev. If the answer is yes, the notice comes back once the process is done.\n\n### Does this PR introduce _any_ user-facing change?\n\n`NOTICE` in the source tree and `/NOTICE` in the container image no longer contain the Export Control Notice. Nothing else changes.\n\n### How was this patch tested?\n\n`dev/check-license-binary.py` passes. It also checks that `NOTICE-binary` is identical to `NOTICE`.\n\n### Was this patch authored or co-authored using generative AI tooling?\n\nYes, co-authored with Claude Code.\n\nCloses #64 from viirya/SPARK-61599.\n\nAuthored-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\n"
    },
    {
      "commit": "73c655bd621684429675732aaacb330241f7ff7a",
      "tree": "50f7f0e374860ffa0857c050d3c2dab17bd4b3f4",
      "parents": [
        "5d801fdd1d77295039c8cc1feda4111ba7d5959c"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Wed Nov 11 15:03:02 2026 -0800"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Wed Nov 11 15:03:02 2026 -0800"
      },
      "message": "[SPARK-61432] Add LICENSE-binary and NOTICE-binary for the container image\n\n### What changes were proposed in this pull request?\n\nThis adds the LICENSE and NOTICE for the container image.\n\n- `LICENSE-binary` is `LICENSE` followed by an appendix that lists the 232 third-party crates linked into the gateway binary on Linux x86_64. The crates are grouped by the license each one is used under. Where a crate offers Apache-2.0 as a choice, that is the license used.\n- `licenses-binary/\u003ccrate\u003e/` holds the license files of the 81 crates that are not used under Apache-2.0 alone. The files are copied unchanged from the published crates:\n  - MIT: 53\n  - Unicode-3.0: 15\n  - BSD-3-Clause: 5\n  - ISC: 3\n  - one each: MIT AND BSD-3-Clause (matchit), Apache-2.0 AND ISC (ring), Zlib, CDLA-Permissive-2.0 (webpki-roots) and BSL-1.0\n- The appendix also covers the Rust standard library, which is statically linked. It is used under Apache-2.0, apart from two parts:\n  - `compiler_builtins` is MIT AND Apache-2.0 WITH LLVM-exception. Its license file is in `licenses-binary/compiler_builtins/`.\n  - `memchr` is MIT and is already in the list.\n- `NOTICE-binary` is the same as `NOTICE`, because no linked crate ships a NOTICE file.\n- `dev/check-license-binary.py` regenerates the crate list from `cargo tree` and compares the result with `LICENSE-binary`. It also compares each `licenses-binary/` file byte-for-byte with the crate\u0027s own file. With `--fix`, it rewrites both and prints what it copied.\n- The Dockerfile installs the files as `/LICENSE`, `/NOTICE` and `/licenses/`.\n  - `.dockerignore` gets an exception for `licenses-binary/`, because its `*.md` rule would otherwise drop `pem`\u0027s `LICENSE.md`.\n  - `.gitattributes` marks `licenses-binary/` as `-text`, so two CRLF files (generic-array, nu-ansi-term) are stored unchanged. Otherwise the repo-wide `eol\u003dlf` rule would rewrite them.\n\nSome licenses are not on the resolved list, and some packaging was decided case by case:\n\n- CDLA-Permissive-2.0 is Category A per LEGAL-732.\n- The OS packages in the distroless base image are the \"system\" layer and stay out of the list, per LEGAL-437 and LEGAL-695. The appendix points to their copyright files in the image.\n- `tonic-prost` publishes no license file. It uses the LICENSE of `tonic`, which comes from the same repository and has the same MIT license.\n\n### Why are the changes needed?\n\nThe first release will ship the container image as a convenience binary. The gateway binary statically links its Rust dependencies, so ASF release policy requires the image\u0027s LICENSE and NOTICE to cover that third-party code.\n\n### Does this PR introduce _any_ user-facing change?\n\nThe container image now contains `/LICENSE`, `/NOTICE` and `/licenses/`. Nothing else changes.\n\n### How was this patch tested?\n\n- `dev/check-license-binary.py` passes: `OK: 232 crates, 81 with license files`. It also passes in a fresh worktree of the commit. Before the `.gitattributes` change, it failed there on the two CRLF files.\n- Before writing the files, I checked the crate list against a separate census of `cargo tree -p gateway -e normal,no-proc-macro --target x86_64-unknown-linux-gnu`. I also read the license file of every crate that is not used under Apache-2.0, to confirm it matches the declared license.\n- I built an image from the three new `COPY` lines with this repo as the build context. `/licenses/` had all 82 directories, including `pem/LICENSE.md`.\n- The CI e2e jobs build the full image from the Dockerfile.\n\n### Was this patch authored or co-authored using generative AI tooling?\n\nYes, co-authored with Claude Code.\n\nCloses #62 from viirya/SPARK-61432.\n\nAuthored-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\n"
    },
    {
      "commit": "5d801fdd1d77295039c8cc1feda4111ba7d5959c",
      "tree": "e094a3f8bcee06ff4578e0208af84ee9625d2e0f",
      "parents": [
        "d46d176ede404372474ef8ffde84c3b621599501"
      ],
      "author": {
        "name": "dependabot[bot]",
        "email": "49699333+dependabot[bot]@users.noreply.github.com",
        "time": "Tue Nov 10 14:43:18 2026 -0800"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Tue Nov 10 14:43:18 2026 -0800"
      },
      "message": "Bump redis from 1.7.0 to 1.8.1\n\nBumps [redis](https://github.com/redis-rs/redis-rs) from 1.7.0 to 1.8.1.\n\u003cdetails\u003e\n\u003csummary\u003eRelease notes\u003c/summary\u003e\n\u003cp\u003e\u003cem\u003eSourced from \u003ca href\u003d\"https://github.com/redis-rs/redis-rs/releases\"\u003eredis\u0027s releases\u003c/a\u003e.\u003c/em\u003e\u003c/p\u003e\n\u003cblockquote\u003e\n\u003ch2\u003eredis-1.8.1\u003c/h2\u003e\n\u003ch4\u003eChanges \u0026amp; Bug fixes\u003c/h4\u003e\n\u003cul\u003e\n\u003cli\u003efix(streams): read radix_tree_keys from XINFO STREAM\u0027s radix-tree-keys (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/pull/2522\"\u003e#2522\u003c/a\u003e by \u003ca href\u003d\"https://github.com/LindseyZ1205\"\u003e\u003ccode\u003e​LindseyZ1205\u003c/code\u003e\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eSkip pending pubsub messages when leaving RESP2 pubsub mode (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/pull/2535\"\u003e#2535\u003c/a\u003e by \u003ca href\u003d\"https://github.com/varun3257\"\u003e\u003ccode\u003e​varun3257\u003c/code\u003e\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eApply every key and flush invalidation to the client-side cache (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/pull/2536\"\u003e#2536\u003c/a\u003e by \u003ca href\u003d\"https://github.com/varun3257\"\u003e\u003ccode\u003e​varun3257\u003c/code\u003e\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eDon\u0027t let push messages consume skipped replies after a read timeout (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/pull/2538\"\u003e#2538\u003c/a\u003e by \u003ca href\u003d\"https://github.com/varun3257\"\u003e\u003ccode\u003e​varun3257\u003c/code\u003e\u003c/a\u003e)\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch2\u003eredis-1.8.0\u003c/h2\u003e\n\u003ch4\u003eChanges \u0026amp; Bug fixes\u003c/h4\u003e\n\u003cul\u003e\n\u003cli\u003eUse a const chunking function in tuple impl of FromRedisValue (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/pull/2424\"\u003e#2424\u003c/a\u003e by \u003ca href\u003d\"https://github.com/nihohit\"\u003e\u003ccode\u003e​nihohit\u003c/code\u003e\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003elint: Add \u003ccode\u003eFromStr\u003c/code\u003e for \u003ccode\u003eNodeAddress\u003c/code\u003e (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/pull/2444\"\u003e#2444\u003c/a\u003e by \u003ca href\u003d\"https://github.com/somechris\"\u003e\u003ccode\u003e​somechris\u003c/code\u003e\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eAdd Debug implementation to AsyncConnectionConfig, PubSubSink and dependent types (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/pull/2450\"\u003e#2450\u003c/a\u003e by \u003ca href\u003d\"https://github.com/nihohit\"\u003e\u003ccode\u003e​nihohit\u003c/code\u003e\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eAdd Debug implementation to Sentinel configuration types (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/pull/2451\"\u003e#2451\u003c/a\u003e by \u003ca href\u003d\"https://github.com/nihohit\"\u003e\u003ccode\u003e​nihohit\u003c/code\u003e\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eAdd debug impl for cluster config types (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/pull/2453\"\u003e#2453\u003c/a\u003e by \u003ca href\u003d\"https://github.com/nihohit\"\u003e\u003ccode\u003e​nihohit\u003c/code\u003e\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eAdd Debug impls for command input types (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/pull/2454\"\u003e#2454\u003c/a\u003e by \u003ca href\u003d\"https://github.com/nihohit\"\u003e\u003ccode\u003e​nihohit\u003c/code\u003e\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eAdd Debug impls for return types (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/pull/2455\"\u003e#2455\u003c/a\u003e by \u003ca href\u003d\"https://github.com/nihohit\"\u003e\u003ccode\u003e​nihohit\u003c/code\u003e\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eAdd FLAT vector field support for \u003ccode\u003eFT.CREATE\u003c/code\u003e [5/8] (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/pull/2319\"\u003e#2319\u003c/a\u003e by \u003ca href\u003d\"https://github.com/StefanPalashev\"\u003e\u003ccode\u003e​StefanPalashev\u003c/code\u003e\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eReplace assert with an error when a multi-slot shard returns a wrong result count (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/pull/2471\"\u003e#2471\u003c/a\u003e by \u003ca href\u003d\"https://github.com/17krishna8\"\u003e\u003ccode\u003e​17krishna8\u003c/code\u003e\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eHandle TLS handshake errors when connecting with a timeout (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/pull/2473\"\u003e#2473\u003c/a\u003e by \u003ca href\u003d\"https://github.com/17krishna8\"\u003e\u003ccode\u003e​17krishna8\u003c/code\u003e\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eAdd HNSW vector field support for \u003ccode\u003eFT.CREATE\u003c/code\u003e [6/8] (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/pull/2467\"\u003e#2467\u003c/a\u003e by \u003ca href\u003d\"https://github.com/StefanPalashev\"\u003e\u003ccode\u003e​StefanPalashev\u003c/code\u003e\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eAdd SVS-VAMANA vector field support for \u003ccode\u003eFT.CREATE\u003c/code\u003e [7/8] (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/pull/2468\"\u003e#2468\u003c/a\u003e by \u003ca href\u003d\"https://github.com/StefanPalashev\"\u003e\u003ccode\u003e​StefanPalashev\u003c/code\u003e\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eRoute \u003ccode\u003eFT.CREATE\u003c/code\u003e as a keyless command and add cluster integration tests [8/8] (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/pull/2469\"\u003e#2469\u003c/a\u003e by \u003ca href\u003d\"https://github.com/StefanPalashev\"\u003e\u003ccode\u003e​StefanPalashev\u003c/code\u003e\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eReject mismatched shard response lengths in logical_aggregate (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/pull/2472\"\u003e#2472\u003c/a\u003e by \u003ca href\u003d\"https://github.com/17krishna8\"\u003e\u003ccode\u003e​17krishna8\u003c/code\u003e\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eRedact secrets in Debug for BasicAuth and ClientCertificate (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/pull/2487\"\u003e#2487\u003c/a\u003e by \u003ca href\u003d\"https://github.com/varun3257\"\u003e\u003ccode\u003e​varun3257\u003c/code\u003e\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eAsyncPubsub: await all confirmations of subscription commands (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/pull/2475\"\u003e#2475\u003c/a\u003e by \u003ca href\u003d\"https://github.com/nihohit\"\u003e\u003ccode\u003e​nihohit\u003c/code\u003e\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eMultiplexed Connection: await all confirmations of subscription commands (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/pull/2491\"\u003e#2491\u003c/a\u003e by \u003ca href\u003d\"https://github.com/nihohit\"\u003e\u003ccode\u003e​nihohit\u003c/code\u003e\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003efix(streams): parse the fields of XINFO STREAM\u0027s first and last entries (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/pull/2494\"\u003e#2494\u003c/a\u003e by \u003ca href\u003d\"https://github.com/LindseyZ1205\"\u003e\u003ccode\u003e​LindseyZ1205\u003c/code\u003e\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eSupport custom root certificates under \u003ccode\u003etls-native-tls\u003c/code\u003e (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/pull/2488\"\u003e#2488\u003c/a\u003e by \u003ca href\u003d\"https://github.com/MohitKS5\"\u003e\u003ccode\u003e​MohitKS5\u003c/code\u003e\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eReject mixed TLS initial nodes regardless of order in ClusterClientBuilder (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/pull/2496\"\u003e#2496\u003c/a\u003e by \u003ca href\u003d\"https://github.com/varun3257\"\u003e\u003ccode\u003e​varun3257\u003c/code\u003e\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003efix(types): forward num_of_args and describe_numeric_behavior for Cow (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/pull/2500\"\u003e#2500\u003c/a\u003e by \u003ca href\u003d\"https://github.com/LindseyZ1205\"\u003e\u003ccode\u003e​LindseyZ1205\u003c/code\u003e\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003efix(connection-manager, cluster): await all confirmations of subscription commands (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/pull/2504\"\u003e#2504\u003c/a\u003e by \u003ca href\u003d\"https://github.com/nihohit\"\u003e\u003ccode\u003e​nihohit\u003c/code\u003e\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003esentinel: ask the next sentinel when one has no valid master for the name (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/pull/2427\"\u003e#2427\u003c/a\u003e by \u003ca href\u003d\"https://github.com/1jeffchristensen\"\u003e\u003ccode\u003e​1jeffchristensen\u003c/code\u003e\u003c/a\u003e)\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch4\u003eCI \u0026amp; operational improvements\u003c/h4\u003e\n\u003cul\u003e\n\u003cli\u003eci: Fix Valkey Bloom module caching (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/pull/2420\"\u003e#2420\u003c/a\u003e by \u003ca href\u003d\"https://github.com/nkaradzhov\"\u003e\u003ccode\u003e​nkaradzhov\u003c/code\u003e\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eAdd test harness and CI support for the Search module [4/8] (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/pull/2318\"\u003e#2318\u003c/a\u003e by \u003ca href\u003d\"https://github.com/StefanPalashev\"\u003e\u003ccode\u003e​StefanPalashev\u003c/code\u003e\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003edoc: Add the environment variable for the search module (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/pull/2429\"\u003e#2429\u003c/a\u003e by \u003ca href\u003d\"https://github.com/somechris\"\u003e\u003ccode\u003e​somechris\u003c/code\u003e\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003efeat: add a seeded randomized test to ensure that interleaved MOVED/broken_connection does not cause a deadlock (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/pull/2425\"\u003e#2425\u003c/a\u003e by \u003ca href\u003d\"https://github.com/virratanasangpunth\"\u003e\u003ccode\u003e​virratanasangpunth\u003c/code\u003e\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eFix new nightly lints lints (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/pull/2460\"\u003e#2460\u003c/a\u003e by \u003ca href\u003d\"https://github.com/nihohit\"\u003e\u003ccode\u003e​nihohit\u003c/code\u003e\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eci: Switch \u003ccode\u003elint-and-check\u003c/code\u003e to stabilized dependencies format (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/pull/2477\"\u003e#2477\u003c/a\u003e by \u003ca href\u003d\"https://github.com/somechris\"\u003e\u003ccode\u003e​somechris\u003c/code\u003e\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003ePin all github runners (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/pull/2501\"\u003e#2501\u003c/a\u003e by \u003ca href\u003d\"https://github.com/nihohit\"\u003e\u003ccode\u003e​nihohit\u003c/code\u003e\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eAdd asserts on error content in tests. (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/pull/2498\"\u003e#2498\u003c/a\u003e by \u003ca href\u003d\"https://github.com/nihohit\"\u003e\u003ccode\u003e​nihohit\u003c/code\u003e\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eImprove caching (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/pull/2497\"\u003e#2497\u003c/a\u003e by \u003ca href\u003d\"https://github.com/nihohit\"\u003e\u003ccode\u003e​nihohit\u003c/code\u003e\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003ePick the parser microbenchmark harnesses over from \u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/issues/2199\"\u003e#2199\u003c/a\u003e (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/pull/2484\"\u003e#2484\u003c/a\u003e by \u003ca href\u003d\"https://github.com/nihohit\"\u003e\u003ccode\u003e​nihohit\u003c/code\u003e\u003c/a\u003e)\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch2\u003eredis-1.7.1\u003c/h2\u003e\n\u003ch4\u003eChanges \u0026amp; Bug fixes\u003c/h4\u003e\n\u003c/blockquote\u003e\n\u003cp\u003e... (truncated)\u003c/p\u003e\n\u003c/details\u003e\n\u003cdetails\u003e\n\u003csummary\u003eCommits\u003c/summary\u003e\n\u003cul\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/redis-rs/redis-rs/commit/d8863e6001bc94423135a2258311c65f2163c6ab\"\u003e\u003ccode\u003ed8863e6\u003c/code\u003e\u003c/a\u003e Prepare version (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/issues/2540\"\u003e#2540\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/redis-rs/redis-rs/commit/f716fd85f7a7ec9b22a93c5dca07f7185516d8d6\"\u003e\u003ccode\u003ef716fd8\u003c/code\u003e\u003c/a\u003e build(deps): bump rustls from 0.23.45 to 0.23.46 (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/issues/2534\"\u003e#2534\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/redis-rs/redis-rs/commit/9af617c962301d4a655420b19134bee1ba8d38dc\"\u003e\u003ccode\u003e9af617c\u003c/code\u003e\u003c/a\u003e Don\u0027t let push messages consume skipped replies after a read timeout (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/issues/2517\"\u003e#2517\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/redis-rs/redis-rs/commit/e358353fe02d9134c0cf1bd6724aef40959ef2dc\"\u003e\u003ccode\u003ee358353\u003c/code\u003e\u003c/a\u003e Apply every key and flush invalidation to the client-side cache (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/issues/2508\"\u003e#2508\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/redis-rs/redis-rs/commit/055768b179e656de3cd8d573bb3680a4d5d69afd\"\u003e\u003ccode\u003e055768b\u003c/code\u003e\u003c/a\u003e skip pending pubsub messages when leaving RESP2 pubsub mode (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/issues/2531\"\u003e#2531\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/redis-rs/redis-rs/commit/164fab9ebb5c28b816dd3d2979f78004ab90ff2f\"\u003e\u003ccode\u003e164fab9\u003c/code\u003e\u003c/a\u003e fix(streams): read radix_tree_keys from XINFO STREAM\u0027s radix-tree-keys (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/issues/2522\"\u003e#2522\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/redis-rs/redis-rs/commit/fec5ef0127a34a3e2096bbc014f18305859223e3\"\u003e\u003ccode\u003efec5ef0\u003c/code\u003e\u003c/a\u003e Prepare version 1.8.0 (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/issues/2507\"\u003e#2507\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/redis-rs/redis-rs/commit/9758f38650eabaca06688d5739252ff2f327fd41\"\u003e\u003ccode\u003e9758f38\u003c/code\u003e\u003c/a\u003e sentinel: ask the next sentinel when one has no valid master for the name (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/issues/2\"\u003e#2\u003c/a\u003e...\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/redis-rs/redis-rs/commit/1d666d18747905dbfa4640e24e28ffb55a93f9b3\"\u003e\u003ccode\u003e1d666d1\u003c/code\u003e\u003c/a\u003e fix(connection-manager, cluster): await all confirmations of subscription com...\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/redis-rs/redis-rs/commit/a61f7358e9b0644a4acb96a332bb14dd5fa26efd\"\u003e\u003ccode\u003ea61f735\u003c/code\u003e\u003c/a\u003e fix(types): forward num_of_args and describe_numeric_behavior for Cow (\u003ca href\u003d\"https://redirect.github.com/redis-rs/redis-rs/issues/2500\"\u003e#2500\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eAdditional commits viewable in \u003ca href\u003d\"https://github.com/redis-rs/redis-rs/compare/redis-1.7.0...redis-1.8.1\"\u003ecompare view\u003c/a\u003e\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/details\u003e\n\u003cbr /\u003e\n\nCloses #61 from dependabot[bot]/dependabot/cargo/redis-1.8.1.\n\nAuthored-by: dependabot[bot] \u003c49699333+dependabot[bot]@users.noreply.github.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\n"
    },
    {
      "commit": "d46d176ede404372474ef8ffde84c3b621599501",
      "tree": "f830cd14f64cb48b186dc1f83f2d7e28c8a25970",
      "parents": [
        "11e9cd9a9ce775174567c942ed74b8f7908e77f5"
      ],
      "author": {
        "name": "dependabot[bot]",
        "email": "49699333+dependabot[bot]@users.noreply.github.com",
        "time": "Thu Nov 05 15:28:21 2026 -0800"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Thu Nov 05 15:28:21 2026 -0800"
      },
      "message": "Bump tokio from 1.53.3 to 1.54.1\n\nBumps [tokio](https://github.com/tokio-rs/tokio) from 1.53.3 to 1.54.1.\n\u003cdetails\u003e\n\u003csummary\u003eRelease notes\u003c/summary\u003e\n\u003cp\u003e\u003cem\u003eSourced from \u003ca href\u003d\"https://github.com/tokio-rs/tokio/releases\"\u003etokio\u0027s releases\u003c/a\u003e.\u003c/em\u003e\u003c/p\u003e\n\u003cblockquote\u003e\n\u003ch2\u003eTokio v1.54.1\u003c/h2\u003e\n\u003ch1\u003e1.54.1 (October 21st, 2026)\u003c/h1\u003e\n\u003ch3\u003eFixed\u003c/h3\u003e\n\u003cul\u003e\n\u003cli\u003etime: revert \u0026quot;remove the runtime handle from timers\u0026quot; (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8692\"\u003e#8692\u003c/a\u003e)\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003e\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8692\"\u003e#8692\u003c/a\u003e: \u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/pull/8692\"\u003etokio-rs/tokio#8692\u003c/a\u003e\u003c/p\u003e\n\u003ch2\u003eTokio v1.54.0\u003c/h2\u003e\n\u003ch1\u003e1.54.0 (October 19th, 2026)\u003c/h1\u003e\n\u003cp\u003eThe 1.54.0 release includes all changes up to 1.53.3.\u003c/p\u003e\n\u003cp\u003eThe MSRV is increased to 1.85. (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8514\"\u003e#8514\u003c/a\u003e)\u003c/p\u003e\n\u003ch3\u003eAdded\u003c/h3\u003e\n\u003cul\u003e\n\u003cli\u003efs: add \u003ccode\u003eFile::set_times\u003c/code\u003e, \u003ccode\u003eFile::set_modified\u003c/code\u003e, and \u003ccode\u003efs::FileTimes\u003c/code\u003e (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8504\"\u003e#8504\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003enet: implement \u003ccode\u003eDefault\u003c/code\u003e for \u003ccode\u003enamed_pipe::{ClientOptions, ServerOptions}\u003c/code\u003e (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8378\"\u003e#8378\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eruntime: add \u003ccode\u003eBuilder::max_io_events_per_busy_tick\u003c/code\u003e (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8410\"\u003e#8410\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eruntime: stabilize \u003ccode\u003etokio::runtime::worker_index()\u003c/code\u003e (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8431\"\u003e#8431\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003esync: add \u003ccode\u003eSemaphorePermit::semaphore()\u003c/code\u003e (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8482\"\u003e#8482\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003esync: add \u003ccode\u003eblocking_acquire\u003c/code\u003e methods to \u003ccode\u003eSemaphore\u003c/code\u003e (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8269\"\u003e#8269\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003esync: add \u003ccode\u003ewatch::Receiver::is_closed\u003c/code\u003e (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8481\"\u003e#8481\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003ewasm: support JSPI for Emscripten target (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8285\"\u003e#8285\u003c/a\u003e, \u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8627\"\u003e#8627\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003ewasm: support the \u003ccode\u003ewasm32-unknown-emscripten\u003c/code\u003e target (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8281\"\u003e#8281\u003c/a\u003e, \u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8422\"\u003e#8422\u003c/a\u003e, \u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8528\"\u003e#8528\u003c/a\u003e, \u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8628\"\u003e#8628\u003c/a\u003e)\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch3\u003eChanged\u003c/h3\u003e\n\u003cul\u003e\n\u003cli\u003efs: reject zero in \u003ccode\u003eFile::set_max_buf_size\u003c/code\u003e (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8449\"\u003e#8449\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eio: make \u003ccode\u003ecopy_buf\u003c/code\u003e cooperative (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8459\"\u003e#8459\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eio: panic when \u003ccode\u003eBufReader\u003c/code\u003e or \u003ccode\u003eBufStream\u003c/code\u003e is created with zero capacity (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8494\"\u003e#8494\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eio: panic when in-memory pipe is created with zero capacity (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8397\"\u003e#8397\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eio: reject zero buffer size in \u003ccode\u003ecopy_bidirectional_with_sizes\u003c/code\u003e (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8454\"\u003e#8454\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eio: support vectored writes in \u003ccode\u003eSink\u003c/code\u003e (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8663\"\u003e#8663\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003enet: start \u003ccode\u003eUnixStream::pair\u003c/code\u003e and \u003ccode\u003eUnixDatagram::pair\u003c/code\u003e sockets as writable (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8444\"\u003e#8444\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003enet: start accepted sockets as readable and writable (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8435\"\u003e#8435\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003enet: support \u003ccode\u003eUCred\u003c/code\u003e on Fuchsia (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8257\"\u003e#8257\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eprocess: use \u003ccode\u003emio::unix::pipe\u003c/code\u003e for Unix child stdio (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8603\"\u003e#8603\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eruntime: cache-pad the shared injection queue and idle state (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8622\"\u003e#8622\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eruntime: decide auto-boxing of spawned futures at monomorphization time (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8414\"\u003e#8414\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eruntime: make \u003ccode\u003eenable_io\u003c/code\u003e available whenever \u003ccode\u003ert\u003c/code\u003e is enabled (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8387\"\u003e#8387\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eruntime: move \u003ccode\u003eblock_in_place\u003c/code\u003e setup into a non-generic function (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8315\"\u003e#8315\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eruntime: move the multi-thread inject queue to its own mutex (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8382\"\u003e#8382\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eruntime: panic when \u003ccode\u003emax_io_events_per_tick\u003c/code\u003e is zero (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8476\"\u003e#8476\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eruntime: suggest \u003ccode\u003e.await\u003c/code\u003e in nested \u003ccode\u003eblock_on\u003c/code\u003e panic message (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8501\"\u003e#8501\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003esync: avoid atomic reader count in broadcast slots (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8357\"\u003e#8357\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003esync: use acquire/release orderings in \u003ccode\u003eNotify\u003c/code\u003e (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8325\"\u003e#8325\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003etime: remove the runtime handle from timers (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8458\"\u003e#8458\u003c/a\u003e)\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/blockquote\u003e\n\u003cp\u003e... (truncated)\u003c/p\u003e\n\u003c/details\u003e\n\u003cdetails\u003e\n\u003csummary\u003eCommits\u003c/summary\u003e\n\u003cul\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/tokio-rs/tokio/commit/b53b2001ed2ccfd9b87e32a0c6da1a1eebdc80b8\"\u003e\u003ccode\u003eb53b200\u003c/code\u003e\u003c/a\u003e chore: prepare Tokio v1.54.1 (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8693\"\u003e#8693\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/tokio-rs/tokio/commit/5ed619bbaba87357ee5103d71b75e7c84cad99c9\"\u003e\u003ccode\u003e5ed619b\u003c/code\u003e\u003c/a\u003e time: revert \u0026quot;remove the runtime handle from timers\u0026quot; (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8692\"\u003e#8692\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/tokio-rs/tokio/commit/66d34e45c9c850ed72bc61f2ad107368217f040d\"\u003e\u003ccode\u003e66d34e4\u003c/code\u003e\u003c/a\u003e chore: prepare Tokio v1.54.0 (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8667\"\u003e#8667\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/tokio-rs/tokio/commit/0d322ce79aef3df11d2bfa87d379ac1d06168ade\"\u003e\u003ccode\u003e0d322ce\u003c/code\u003e\u003c/a\u003e io: support vectored writes in sink (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8663\"\u003e#8663\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/tokio-rs/tokio/commit/4167c5634ef9cd091cc5b37156f04cfdb0595414\"\u003e\u003ccode\u003e4167c56\u003c/code\u003e\u003c/a\u003e codec: expose I/O sources from codec errors (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8665\"\u003e#8665\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/tokio-rs/tokio/commit/cdd9dae3aa5056e23dc4d5bc4f852ab8fbf27aff\"\u003e\u003ccode\u003ecdd9dae\u003c/code\u003e\u003c/a\u003e fs: propagate iouring stress test failures (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8644\"\u003e#8644\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/tokio-rs/tokio/commit/52ac7816230efabcae6fcc09cbe7ae630a48eddf\"\u003e\u003ccode\u003e52ac781\u003c/code\u003e\u003c/a\u003e ci: run the emscripten JSPI tests as an auto-initialized ES module (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8640\"\u003e#8640\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/tokio-rs/tokio/commit/4890d027f8a566ccf081e7f019ed766f780df1af\"\u003e\u003ccode\u003e4890d02\u003c/code\u003e\u003c/a\u003e stream: implement \u003ccode\u003eFusedStream\u003c/code\u003e for \u003ccode\u003eJoinSetStream\u003c/code\u003e (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8629\"\u003e#8629\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/tokio-rs/tokio/commit/a6bd8e170c4e895fab26024bdfeea1c568567545\"\u003e\u003ccode\u003ea6bd8e1\u003c/code\u003e\u003c/a\u003e wasm: check Emscripten version when using JSPI (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8627\"\u003e#8627\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/tokio-rs/tokio/commit/9dd749d33408ffee1a4fd7c462546651416428a8\"\u003e\u003ccode\u003e9dd749d\u003c/code\u003e\u003c/a\u003e Merge \u0027tokio-1.53.3\u0027 into \u0027master\u0027 (\u003ca href\u003d\"https://redirect.github.com/tokio-rs/tokio/issues/8641\"\u003e#8641\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eAdditional commits viewable in \u003ca href\u003d\"https://github.com/tokio-rs/tokio/compare/tokio-1.53.3...tokio-1.54.1\"\u003ecompare view\u003c/a\u003e\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/details\u003e\n\u003cbr /\u003e\n\n[![Dependabot compatibility score](https://dependabot-badges.githubapp.com/badges/compatibility_score?dependency-name\u003dtokio\u0026package-manager\u003dcargo\u0026previous-version\u003d1.53.3\u0026new-version\u003d1.54.1)](https://docs.github.com/en/github/managing-security-vulnerabilities/about-dependabot-security-updates#about-compatibility-scores)\n\nDependabot will resolve any conflicts with this PR as long as you don\u0027t alter it yourself. You can also trigger a rebase manually by commenting `dependabot rebase`.\n\n[//]: # (dependabot-automerge-start)\n[//]: # (dependabot-automerge-end)\n\n---\n\n\u003cdetails\u003e\n\u003csummary\u003eDependabot commands and options\u003c/summary\u003e\n\u003cbr /\u003e\n\nYou can trigger Dependabot actions by commenting on this PR:\n- `dependabot rebase` will rebase this PR\n- `dependabot recreate` will recreate this PR, overwriting any edits that have been made to it\n- `dependabot show \u003cdependency name\u003e ignore conditions` will show all of the ignore conditions of the specified dependency\n- `dependabot ignore this major version` will close this PR and stop Dependabot creating any more for this major version (unless you reopen the PR or upgrade to it yourself)\n- `dependabot ignore this minor version` will close this PR and stop Dependabot creating any more for this minor version (unless you reopen the PR or upgrade to it yourself)\n- `dependabot ignore this dependency` will close this PR and stop Dependabot creating any more for this dependency (unless you reopen the PR or upgrade to it yourself)\n\n\u003c/details\u003e\n\nCloses #60 from dependabot[bot]/dependabot/cargo/tokio-1.54.1.\n\nAuthored-by: dependabot[bot] \u003c49699333+dependabot[bot]@users.noreply.github.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\n"
    },
    {
      "commit": "11e9cd9a9ce775174567c942ed74b8f7908e77f5",
      "tree": "ed31b0a194ad7f1de4b83a5dcd8eec95747c0cbb",
      "parents": [
        "027bc88c4ee8c507e97c4aabcc9e907f35ddee96"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Tue Nov 03 23:18:14 2026 -0800"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Tue Nov 03 23:18:14 2026 -0800"
      },
      "message": "[SPARK-60853] Build with --locked in CI and the Dockerfile so Cargo.lock is enforced\n\n### What changes were proposed in this pull request?\n\nAdd `--locked` to every cargo command in `.github/workflows/ci.yml` that\nresolves dependencies (clippy, build, test, the ignored Redis and tracing\nintegration tests, and the `drain_smoke` / `ha_smoke` examples), and to the\nrelease `cargo build` in the `Dockerfile`. The e2e jobs build their image from\nthe `Dockerfile`, so they are covered too. `cargo fmt` does not resolve\ndependencies and is unchanged. The `cargo install` in `audit.yml` already\nuses `--locked`.\n\n### Why are the changes needed?\n\nWithout `--locked`, cargo silently re-resolves on the runner when `Cargo.lock`\nis out of date with the manifests, for example a stale dependency list in an\nentry or a version that no longer satisfies a requirement. The job then passes\nwith a lockfile that differs from the one committed, so a broken `Cargo.lock`\ncan be merged without CI noticing. This matters most for lockfile-only changes\nsuch as dependabot PRs and RustSec updates, where the lockfile is the entire\nchange.\n\nFor the `Dockerfile` it also matters for releases: without `--locked`, the\ndependency versions in a published image are not guaranteed to match the\n`Cargo.lock` in the source release.\n\n### Does this PR introduce _any_ user-facing change?\n\nNo. CI and image build configuration only. Building the image now fails if\n`Cargo.lock` is out of date, instead of silently updating it.\n\n### How was this patch tested?\n\nThis PR\u0027s own CI run is the main test: every changed step runs with the new\nflag. Locally, on the pinned 1.99.0 toolchain against `main`\u0027s `Cargo.lock`:\n\n- `cargo clippy --workspace --all-targets --locked -- -D warnings`, `cargo\n  build --workspace --locked`: clean.\n- `cargo test --workspace --locked`: 204 passed, 0 failed.\n- `cargo test -p scg-proxy --test tracing_integration --locked -- --ignored`:\n  passed. `cargo run -p scg-proxy --example drain_smoke --locked`: all drain\n  invariants passed. The Redis test and `ha_smoke` were compiled with\n  `--locked`; running them needs the CI Redis service.\n- `Cargo.lock` was not modified by any of the above.\n- Negative check: with `memchr` removed from `pulldown-cmark`\u0027s dependency\n  list in `Cargo.lock`, `cargo build -p scg-genproto --locked` fails with\n  \"cannot update the lock file ... because --locked was passed\". Without\n  `--locked` the same build succeeds and quietly rewrites the lockfile, which\n  is the behavior CI has today.\n\n### Was this patch authored or co-authored using generative AI tooling?\n\nYes, co-authored with Claude Code.\n\nCloses #59 from viirya/SPARK-60853.\n\nAuthored-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\n"
    },
    {
      "commit": "027bc88c4ee8c507e97c4aabcc9e907f35ddee96",
      "tree": "f76211b32e7b19341d2950fc46108c9bfd186d7d",
      "parents": [
        "4d9af1fb0c16c4e859bc07818df36c19ea1644fa"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Tue Nov 03 23:17:36 2026 -0800"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Tue Nov 03 23:17:36 2026 -0800"
      },
      "message": "[SPARK-60852] Update pulldown-cmark in Cargo.lock to pick up recent RustSec advisories\n\n### What changes were proposed in this pull request?\n\nUpdate `pulldown-cmark` 0.13.4 -\u003e 0.13.6 in `Cargo.lock`. Lockfile-only; no\n`Cargo.toml` change, since `prost-build` requires `pulldown-cmark ^0.13`.\n\nOnly the entry\u0027s `version` and `checksum` lines change. The dependencies and\nfeatures of 0.13.6 are identical to 0.13.4 (bitflags, memchr, unicase; the same\noptional ones), so no other lockfile entry changes. The resolved set stays at\n431 crates.\n\n### Why are the changes needed?\n\nThe scheduled `cargo audit` run on `main` has failed every day since\n2026-10-29 because of two advisories against `pulldown-cmark` 0.13.4:\n\n| Advisory | Issue | Fixed in |\n|---|---|---|\n| RUSTSEC-2026-0414 | Exponential event growth with `ENABLE_WIKILINKS` | 0.13.5 |\n| RUSTSEC-2026-0433 | Quadratic parsing of nested image wikilinks with `ENABLE_WIKILINKS` | 0.13.6 |\n\n`pulldown-cmark` is a build-time dependency only:\n`scg-genproto` (build script) -\u003e `tonic-prost-build` -\u003e `prost-build`, which\ncleans up the markdown in `.proto` comments. `prost-build` parses with\n`Options::all()`, which includes `ENABLE_WIKILINKS`, so the affected code does\nrun, but only at build time and only on the vendored `.proto` files in this\nrepository. The crate is not linked into the gateway binary.\n\nThis is the last failure in the scheduled run: with it fixed, `cargo audit` on\n`main` exits 0.\n\n### Does this PR introduce _any_ user-facing change?\n\nNo. Build-time dependency only.\n\n### How was this patch tested?\n\n- The checksum was checked against three sources: the crates.io API, the\n  SHA-256 of `pulldown-cmark-0.13.6.crate` downloaded from static.crates.io,\n  and the new lockfile line, all `3db2be46...69bb7b`.\n- The dependency and feature sets were compared between 0.13.4 and 0.13.6\n  using both the crates.io API and the `Cargo.toml` inside the downloaded\n  `.crate`.\n- `cargo audit` exits 0, with only the two existing allowed warnings\n  (`unsafe-libyaml` unmaintained, `chacha20` yanked).\n- `dev/audit-new-advisories.sh` against `main`\u0027s `Cargo.lock` (what this PR\u0027s\n  `cargo audit` check runs) lists RUSTSEC-2026-0414 and RUSTSEC-2026-0433 as\n  resolved and nothing introduced.\n- The CI build exercises the new version: `scg-genproto`\u0027s build script runs\n  `prost-build` on every `.proto` file.\n- License unchanged: MIT.\n\n### Was this patch authored or co-authored using generative AI tooling?\n\nYes, co-authored with Claude Code.\n\nCloses #58 from viirya/SPARK-60852.\n\nAuthored-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\n"
    },
    {
      "commit": "4d9af1fb0c16c4e859bc07818df36c19ea1644fa",
      "tree": "dc3949ccbfb19f5287e0e855c4c17a4e312b7ac7",
      "parents": [
        "d96ea8f0ad07b4769789740b4691d8307cf8a020"
      ],
      "author": {
        "name": "dependabot[bot]",
        "email": "49699333+dependabot[bot]@users.noreply.github.com",
        "time": "Tue Nov 03 14:19:41 2026 -0800"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Tue Nov 03 14:19:41 2026 -0800"
      },
      "message": "Bump uuid from 1.24.0 to 1.29.0\n\nBumps [uuid](https://github.com/uuid-rs/uuid) from 1.24.0 to 1.29.0.\n\u003cdetails\u003e\n\u003csummary\u003eRelease notes\u003c/summary\u003e\n\u003cp\u003e\u003cem\u003eSourced from \u003ca href\u003d\"https://github.com/uuid-rs/uuid/releases\"\u003euuid\u0027s releases\u003c/a\u003e.\u003c/em\u003e\u003c/p\u003e\n\u003cblockquote\u003e\n\u003ch2\u003ev1.29.0\u003c/h2\u003e\n\u003ch2\u003eWhat\u0027s Changed\u003c/h2\u003e\n\u003cul\u003e\n\u003cli\u003eAdd versioned constructors to NonNilUuid by \u003ca href\u003d\"https://github.com/RanaPriyansh\"\u003e\u003ccode\u003e​RanaPriyansh\u003c/code\u003e\u003c/a\u003e in \u003ca href\u003d\"https://redirect.github.com/uuid-rs/uuid/pull/919\"\u003euuid-rs/uuid#919\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003ePrepare for 1.29.0 release by \u003ca href\u003d\"https://github.com/KodrAus\"\u003e\u003ccode\u003e​KodrAus\u003c/code\u003e\u003c/a\u003e in \u003ca href\u003d\"https://redirect.github.com/uuid-rs/uuid/pull/920\"\u003euuid-rs/uuid#920\u003c/a\u003e\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003e\u003cstrong\u003eFull Changelog\u003c/strong\u003e: \u003ca href\u003d\"https://github.com/uuid-rs/uuid/compare/v1.28.0...v1.29.0\"\u003ehttps://github.com/uuid-rs/uuid/compare/v1.28.0...v1.29.0\u003c/a\u003e\u003c/p\u003e\n\u003ch2\u003ev1.28.0\u003c/h2\u003e\n\u003ch2\u003eWhat\u0027s Changed\u003c/h2\u003e\n\u003cul\u003e\n\u003cli\u003eAdd optional serde format adapters by \u003ca href\u003d\"https://github.com/RanaPriyansh\"\u003e\u003ccode\u003e​RanaPriyansh\u003c/code\u003e\u003c/a\u003e in \u003ca href\u003d\"https://redirect.github.com/uuid-rs/uuid/pull/916\"\u003euuid-rs/uuid#916\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003ePrepare for 1.28.0 release by \u003ca href\u003d\"https://github.com/KodrAus\"\u003e\u003ccode\u003e​KodrAus\u003c/code\u003e\u003c/a\u003e in \u003ca href\u003d\"https://redirect.github.com/uuid-rs/uuid/pull/918\"\u003euuid-rs/uuid#918\u003c/a\u003e\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch2\u003eNew Contributors\u003c/h2\u003e\n\u003cul\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/RanaPriyansh\"\u003e\u003ccode\u003e​RanaPriyansh\u003c/code\u003e\u003c/a\u003e made their first contribution in \u003ca href\u003d\"https://redirect.github.com/uuid-rs/uuid/pull/916\"\u003euuid-rs/uuid#916\u003c/a\u003e\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003e\u003cstrong\u003eFull Changelog\u003c/strong\u003e: \u003ca href\u003d\"https://github.com/uuid-rs/uuid/compare/v1.27.0...v1.28.0\"\u003ehttps://github.com/uuid-rs/uuid/compare/v1.27.0...v1.28.0\u003c/a\u003e\u003c/p\u003e\n\u003ch2\u003ev1.27.0\u003c/h2\u003e\n\u003ch2\u003eWhat\u0027s Changed\u003c/h2\u003e\n\u003cul\u003e\n\u003cli\u003eAccept case-insensitive urn:uuid: prefixes when parsing by \u003ca href\u003d\"https://github.com/jaideeppyne\"\u003e\u003ccode\u003e​jaideeppyne\u003c/code\u003e\u003c/a\u003e in \u003ca href\u003d\"https://redirect.github.com/uuid-rs/uuid/pull/912\"\u003euuid-rs/uuid#912\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003eFix ordering of V7 UUIDs created by earlier seconds with later subsec nanos by \u003ca href\u003d\"https://github.com/KodrAus\"\u003e\u003ccode\u003e​KodrAus\u003c/code\u003e\u003c/a\u003e in \u003ca href\u003d\"https://redirect.github.com/uuid-rs/uuid/pull/915\"\u003euuid-rs/uuid#915\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003ePrepare for 1.27.0 release by \u003ca href\u003d\"https://github.com/KodrAus\"\u003e\u003ccode\u003e​KodrAus\u003c/code\u003e\u003c/a\u003e in \u003ca href\u003d\"https://redirect.github.com/uuid-rs/uuid/pull/917\"\u003euuid-rs/uuid#917\u003c/a\u003e\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch2\u003eNew Contributors\u003c/h2\u003e\n\u003cul\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/jaideeppyne\"\u003e\u003ccode\u003e​jaideeppyne\u003c/code\u003e\u003c/a\u003e made their first contribution in \u003ca href\u003d\"https://redirect.github.com/uuid-rs/uuid/pull/912\"\u003euuid-rs/uuid#912\u003c/a\u003e\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003e\u003cstrong\u003eFull Changelog\u003c/strong\u003e: \u003ca href\u003d\"https://github.com/uuid-rs/uuid/compare/v1.26.1...v1.27.0\"\u003ehttps://github.com/uuid-rs/uuid/compare/v1.26.1...v1.27.0\u003c/a\u003e\u003c/p\u003e\n\u003ch2\u003ev1.26.1\u003c/h2\u003e\n\u003ch2\u003eWhat\u0027s Changed\u003c/h2\u003e\n\u003cul\u003e\n\u003cli\u003eSeat the v7 counter below the version nibble by \u003ca href\u003d\"https://github.com/lenamonj\"\u003e\u003ccode\u003e​lenamonj\u003c/code\u003e\u003c/a\u003e in \u003ca href\u003d\"https://redirect.github.com/uuid-rs/uuid/pull/907\"\u003euuid-rs/uuid#907\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003eDon\u0027t panic in overflowing Timestamp to SystemTime conversion by \u003ca href\u003d\"https://github.com/KodrAus\"\u003e\u003ccode\u003e​KodrAus\u003c/code\u003e\u003c/a\u003e in \u003ca href\u003d\"https://redirect.github.com/uuid-rs/uuid/pull/909\"\u003euuid-rs/uuid#909\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003ePrepare for 1.26.1 release by \u003ca href\u003d\"https://github.com/KodrAus\"\u003e\u003ccode\u003e​KodrAus\u003c/code\u003e\u003c/a\u003e in \u003ca href\u003d\"https://redirect.github.com/uuid-rs/uuid/pull/910\"\u003euuid-rs/uuid#910\u003c/a\u003e\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch2\u003eNew Contributors\u003c/h2\u003e\n\u003cul\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/lenamonj\"\u003e\u003ccode\u003e​lenamonj\u003c/code\u003e\u003c/a\u003e made their first contribution in \u003ca href\u003d\"https://redirect.github.com/uuid-rs/uuid/pull/907\"\u003euuid-rs/uuid#907\u003c/a\u003e\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003e\u003cstrong\u003eFull Changelog\u003c/strong\u003e: \u003ca href\u003d\"https://github.com/uuid-rs/uuid/compare/v1.26.0...v1.26.1\"\u003ehttps://github.com/uuid-rs/uuid/compare/v1.26.0...v1.26.1\u003c/a\u003e\u003c/p\u003e\n\u003ch2\u003ev1.26.0\u003c/h2\u003e\n\u003ch2\u003eWhat\u0027s Changed\u003c/h2\u003e\n\u003cul\u003e\n\u003cli\u003eAdd ContextV7::with_additional_precision_bits by \u003ca href\u003d\"https://github.com/ChrisJr404\"\u003e\u003ccode\u003e​ChrisJr404\u003c/code\u003e\u003c/a\u003e in \u003ca href\u003d\"https://redirect.github.com/uuid-rs/uuid/pull/904\"\u003euuid-rs/uuid#904\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003ePrepare for 1.26.0 release by \u003ca href\u003d\"https://github.com/KodrAus\"\u003e\u003ccode\u003e​KodrAus\u003c/code\u003e\u003c/a\u003e in \u003ca href\u003d\"https://redirect.github.com/uuid-rs/uuid/pull/905\"\u003euuid-rs/uuid#905\u003c/a\u003e\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003e\u003cstrong\u003eFull Changelog\u003c/strong\u003e: \u003ca href\u003d\"https://github.com/uuid-rs/uuid/compare/1.25.0...v1.26.0\"\u003ehttps://github.com/uuid-rs/uuid/compare/1.25.0...v1.26.0\u003c/a\u003e\u003c/p\u003e\n\u003ch2\u003e1.25.0\u003c/h2\u003e\n\u003ch2\u003eWhat\u0027s Changed\u003c/h2\u003e\n\u003c/blockquote\u003e\n\u003cp\u003e... (truncated)\u003c/p\u003e\n\u003c/details\u003e\n\u003cdetails\u003e\n\u003csummary\u003eCommits\u003c/summary\u003e\n\u003cul\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/uuid-rs/uuid/commit/4643e607d0ded1f2f95c1c427456e17bd5edf708\"\u003e\u003ccode\u003e4643e60\u003c/code\u003e\u003c/a\u003e Merge pull request \u003ca href\u003d\"https://redirect.github.com/uuid-rs/uuid/issues/920\"\u003e#920\u003c/a\u003e from uuid-rs/cargo/v1.29.0\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/uuid-rs/uuid/commit/a3631fd73dd228a50f6bc7a544b2f5e05134bbc8\"\u003e\u003ccode\u003ea3631fd\u003c/code\u003e\u003c/a\u003e prepare for 1.29.0 release\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/uuid-rs/uuid/commit/304913efacbf3ed0023b9d8af3bd59cbfa236b60\"\u003e\u003ccode\u003e304913e\u003c/code\u003e\u003c/a\u003e Merge pull request \u003ca href\u003d\"https://redirect.github.com/uuid-rs/uuid/issues/919\"\u003e#919\u003c/a\u003e from RanaPriyansh/fix/non-nil-versioned-constructors-901\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/uuid-rs/uuid/commit/d49f963255bde95edd0a36369263520b1dd67308\"\u003e\u003ccode\u003ed49f963\u003c/code\u003e\u003c/a\u003e Reuse existing version tests for NonNilUuid constructors\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/uuid-rs/uuid/commit/2f024047699f8ad721dcc44fce9171b88aa5aab1\"\u003e\u003ccode\u003e2f02404\u003c/code\u003e\u003c/a\u003e Add versioned constructors to NonNilUuid\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/uuid-rs/uuid/commit/276c23628a669ca9fca637f3326c21b114e88049\"\u003e\u003ccode\u003e276c236\u003c/code\u003e\u003c/a\u003e Merge pull request \u003ca href\u003d\"https://redirect.github.com/uuid-rs/uuid/issues/918\"\u003e#918\u003c/a\u003e from uuid-rs/cargo/v1.28.0\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/uuid-rs/uuid/commit/4b3b9f55260d402ed3184577bc35a054ba9a28f4\"\u003e\u003ccode\u003e4b3b9f5\u003c/code\u003e\u003c/a\u003e prepare for 1.28.0 release\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/uuid-rs/uuid/commit/cbf9c818b89754a83d4ca54f2a9bf809e2476f67\"\u003e\u003ccode\u003ecbf9c81\u003c/code\u003e\u003c/a\u003e Merge pull request \u003ca href\u003d\"https://redirect.github.com/uuid-rs/uuid/issues/916\"\u003e#916\u003c/a\u003e from RanaPriyansh/fix/optional-serde-adapters-840\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/uuid-rs/uuid/commit/784a0229388c3df8d729eb212f0580d2617191b2\"\u003e\u003ccode\u003e784a022\u003c/code\u003e\u003c/a\u003e Inline expected tokens in optional compact serde test\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/uuid-rs/uuid/commit/a5678c3156831969fbd4573b7afdd0585d88e3a1\"\u003e\u003ccode\u003ea5678c3\u003c/code\u003e\u003c/a\u003e Simplify optional serde adapter tests to option tokens\u003c/li\u003e\n\u003cli\u003eAdditional commits viewable in \u003ca href\u003d\"https://github.com/uuid-rs/uuid/compare/v1.24.0...v1.29.0\"\u003ecompare view\u003c/a\u003e\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/details\u003e\n\u003cbr /\u003e\n\n[![Dependabot compatibility score](https://dependabot-badges.githubapp.com/badges/compatibility_score?dependency-name\u003duuid\u0026package-manager\u003dcargo\u0026previous-version\u003d1.24.0\u0026new-version\u003d1.29.0)](https://docs.github.com/en/github/managing-security-vulnerabilities/about-dependabot-security-updates#about-compatibility-scores)\n\nDependabot will resolve any conflicts with this PR as long as you don\u0027t alter it yourself. You can also trigger a rebase manually by commenting `dependabot rebase`.\n\n[//]: # (dependabot-automerge-start)\n[//]: # (dependabot-automerge-end)\n\n---\n\n\u003cdetails\u003e\n\u003csummary\u003eDependabot commands and options\u003c/summary\u003e\n\u003cbr /\u003e\n\nYou can trigger Dependabot actions by commenting on this PR:\n- `dependabot rebase` will rebase this PR\n- `dependabot recreate` will recreate this PR, overwriting any edits that have been made to it\n- `dependabot show \u003cdependency name\u003e ignore conditions` will show all of the ignore conditions of the specified dependency\n- `dependabot ignore this major version` will close this PR and stop Dependabot creating any more for this major version (unless you reopen the PR or upgrade to it yourself)\n- `dependabot ignore this minor version` will close this PR and stop Dependabot creating any more for this minor version (unless you reopen the PR or upgrade to it yourself)\n- `dependabot ignore this dependency` will close this PR and stop Dependabot creating any more for this dependency (unless you reopen the PR or upgrade to it yourself)\n\n\u003c/details\u003e\n\nCloses #57 from dependabot[bot]/dependabot/cargo/uuid-1.29.0.\n\nAuthored-by: dependabot[bot] \u003c49699333+dependabot[bot]@users.noreply.github.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\n"
    },
    {
      "commit": "d96ea8f0ad07b4769789740b4691d8307cf8a020",
      "tree": "7a50c2307aef466bad5f5bd259214ede2a4b2ec2",
      "parents": [
        "d54a1804f81de3033e1398fb5b6921238e29d20f"
      ],
      "author": {
        "name": "dependabot[bot]",
        "email": "49699333+dependabot[bot]@users.noreply.github.com",
        "time": "Wed Oct 28 10:56:30 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Wed Oct 28 10:56:30 2026 -0700"
      },
      "message": "Bump hyper from 1.11.0 to 1.12.0\n\nBumps [hyper](https://github.com/hyperium/hyper) from 1.11.0 to 1.12.0.\n\u003cdetails\u003e\n\u003csummary\u003eRelease notes\u003c/summary\u003e\n\u003cp\u003e\u003cem\u003eSourced from \u003ca href\u003d\"https://github.com/hyperium/hyper/releases\"\u003ehyper\u0027s releases\u003c/a\u003e.\u003c/em\u003e\u003c/p\u003e\n\u003cblockquote\u003e\n\u003ch2\u003ev1.12.0\u003c/h2\u003e\n\u003ch2\u003eFeatures\u003c/h2\u003e\n\u003cul\u003e\n\u003cli\u003e\u003cstrong\u003ehttp1:\u003c/strong\u003e add max_header_size limit for server and client (\u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/issues/4183\"\u003e#4183\u003c/a\u003e) (\u003ca href\u003d\"https://github.com/hyperium/hyper/commit/d1cd14eff392be63a2318d690cc9380723ec61a1\"\u003ed1cd14ef\u003c/a\u003e, closes \u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/issues/3832\"\u003e#3832\u003c/a\u003e)\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch2\u003eBug Fixes\u003c/h2\u003e\n\u003cul\u003e\n\u003cli\u003e\u003cstrong\u003ehttp1:\u003c/strong\u003e\n\u003cul\u003e\n\u003cli\u003ewithdraw a stale want when the client dispatcher takes a request (\u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/issues/4208\"\u003e#4208\u003c/a\u003e) (\u003ca href\u003d\"https://github.com/hyperium/hyper/commit/c44f1cc51c2bc45b502031a646d3d54315bd7a16\"\u003ec44f1cc5\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eremove content-length if transfer-encoding is auto-added (\u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/issues/4224\"\u003e#4224\u003c/a\u003e) (\u003ca href\u003d\"https://github.com/hyperium/hyper/commit/69396fdd960629d69b09cc6c9b367ef8650c001d\"\u003e69396fdd\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003ealways remove content-length if transfer-encoding is sent (\u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/issues/4223\"\u003e#4223\u003c/a\u003e) (\u003ca href\u003d\"https://github.com/hyperium/hyper/commit/a5d7fd7861598c0b8c857cb75240a75bcfbd96a3\"\u003ea5d7fd78\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003elet a Connection close token win over a later keep-alive (\u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/issues/4201\"\u003e#4201\u003c/a\u003e) (\u003ca href\u003d\"https://github.com/hyperium/hyper/commit/5dadff76d1cb5d21046218df3568569fbba8ed21\"\u003e5dadff76\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003epreserve hop-by-hop when setting close or keep-alive (\u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/issues/4196\"\u003e#4196\u003c/a\u003e) (\u003ca href\u003d\"https://github.com/hyperium/hyper/commit/74a39081ff1496970fb75da75c5282b23a9b69e5\"\u003e74a39081\u003c/a\u003e)\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003ehttp2:\u003c/strong\u003e do not reserve capacity for idle Upgraded streams (\u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/issues/4198\"\u003e#4198\u003c/a\u003e) (\u003ca href\u003d\"https://github.com/hyperium/hyper/commit/e89d0d93d9566210ee95aea234a00952ebc4a202\"\u003ee89d0d93\u003c/a\u003e)\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch2\u003eEverything\u003c/h2\u003e\n\u003cul\u003e\n\u003cli\u003edocs: fix typo responsibile -\u0026gt; responsible by \u003ca href\u003d\"https://github.com/vaibhav8a\"\u003e\u003ccode\u003e​vaibhav8a\u003c/code\u003e\u003c/a\u003e in \u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/pull/4177\"\u003ehyperium/hyper#4177\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003estyle: replace single-character lifetime names with descriptive names by \u003ca href\u003d\"https://github.com/dox4\"\u003e\u003ccode\u003e​dox4\u003c/code\u003e\u003c/a\u003e in \u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/pull/4173\"\u003ehyperium/hyper#4173\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003erefactor(http1): resolve collapsible_match lint by \u003ca href\u003d\"https://github.com/nakaryo716\"\u003e\u003ccode\u003e​nakaryo716\u003c/code\u003e\u003c/a\u003e in \u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/pull/4178\"\u003ehyperium/hyper#4178\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003ehyper Improvement Proposals by \u003ca href\u003d\"https://github.com/seanmonstar\"\u003e\u003ccode\u003e​seanmonstar\u003c/code\u003e\u003c/a\u003e in \u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/pull/4131\"\u003ehyperium/hyper#4131\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003eperf(body): simpler custom Incoming channel by \u003ca href\u003d\"https://github.com/seanmonstar\"\u003e\u003ccode\u003e​seanmonstar\u003c/code\u003e\u003c/a\u003e in \u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/pull/4168\"\u003ehyperium/hyper#4168\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003edocs(maintainers): add 0x676e67 as a collaborator by \u003ca href\u003d\"https://github.com/seanmonstar\"\u003e\u003ccode\u003e​seanmonstar\u003c/code\u003e\u003c/a\u003e in \u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/pull/4184\"\u003ehyperium/hyper#4184\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003efix(ext): remove broken \u003ccode\u003eOriginalHeaderOrder\u003c/code\u003e doctest by \u003ca href\u003d\"https://github.com/cratelyn\"\u003e\u003ccode\u003e​cratelyn\u003c/code\u003e\u003c/a\u003e in \u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/pull/4188\"\u003ehyperium/hyper#4188\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003ebuild(deps): remove unused \u003ccode\u003efutures-channel\u003c/code\u003e from \u003ccode\u003ehttp1\u003c/code\u003e feature by \u003ca href\u003d\"https://github.com/0x676e67\"\u003e\u003ccode\u003e​0x676e67\u003c/code\u003e\u003c/a\u003e in \u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/pull/4185\"\u003ehyperium/hyper#4185\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003echore(ci): update to cargo-check-external-types 0.5.0 by \u003ca href\u003d\"https://github.com/tottoto\"\u003e\u003ccode\u003e​tottoto\u003c/code\u003e\u003c/a\u003e in \u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/pull/4189\"\u003ehyperium/hyper#4189\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003estyle(lib): address ignored_unit_patterns lint by \u003ca href\u003d\"https://github.com/nakaryo716\"\u003e\u003ccode\u003e​nakaryo716\u003c/code\u003e\u003c/a\u003e in \u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/pull/4193\"\u003ehyperium/hyper#4193\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003ePreserve hop-by-hop when setting close or keep-alive by \u003ca href\u003d\"https://github.com/dtolnay\"\u003e\u003ccode\u003e​dtolnay\u003c/code\u003e\u003c/a\u003e in \u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/pull/4196\"\u003ehyperium/hyper#4196\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003efix(http2): do not reserve capacity for idle Upgraded streams by \u003ca href\u003d\"https://github.com/Dev-next-gen\"\u003e\u003ccode\u003e​Dev-next-gen\u003c/code\u003e\u003c/a\u003e in \u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/pull/4198\"\u003ehyperium/hyper#4198\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003estyle(lib): Address clippy panics and missing_panics_doc lints by \u003ca href\u003d\"https://github.com/Siech0\"\u003e\u003ccode\u003e​Siech0\u003c/code\u003e\u003c/a\u003e in \u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/pull/4191\"\u003ehyperium/hyper#4191\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003edocs: fix typos in capi README, MSRV and governance docs by \u003ca href\u003d\"https://github.com/toyeshhm\"\u003e\u003ccode\u003e​toyeshhm\u003c/code\u003e\u003c/a\u003e in \u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/pull/4204\"\u003ehyperium/hyper#4204\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003estyle(lib): Address trivial, low quantity clippy lints. by \u003ca href\u003d\"https://github.com/Siech0\"\u003e\u003ccode\u003e​Siech0\u003c/code\u003e\u003c/a\u003e in \u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/pull/4190\"\u003ehyperium/hyper#4190\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003echore(clippy): allow clippy lints following 1.99.0 release by \u003ca href\u003d\"https://github.com/dswij\"\u003e\u003ccode\u003e​dswij\u003c/code\u003e\u003c/a\u003e in \u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/pull/4219\"\u003ehyperium/hyper#4219\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003estyle: remove useless_borrows_in_formatting from the clippy allowlist by \u003ca href\u003d\"https://github.com/ump45nose\"\u003e\u003ccode\u003e​ump45nose\u003c/code\u003e\u003c/a\u003e in \u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/pull/4220\"\u003ehyperium/hyper#4220\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003estyle(h1): mark partial Debug impls as non-exhaustive by \u003ca href\u003d\"https://github.com/badimalex\"\u003e\u003ccode\u003e​badimalex\u003c/code\u003e\u003c/a\u003e in \u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/pull/4218\"\u003ehyperium/hyper#4218\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003estyle(lib): Address cast-possible-truncation clippy findings. by \u003ca href\u003d\"https://github.com/Siech0\"\u003e\u003ccode\u003e​Siech0\u003c/code\u003e\u003c/a\u003e in \u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/pull/4192\"\u003ehyperium/hyper#4192\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003efeat(http1): add max_header_size limit for server and client by \u003ca href\u003d\"https://github.com/Aditya-9-6\"\u003e\u003ccode\u003e​Aditya-9-6\u003c/code\u003e\u003c/a\u003e in \u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/pull/4183\"\u003ehyperium/hyper#4183\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003efix(http1): let a Connection close token win over a later keep-alive by \u003ca href\u003d\"https://github.com/shoemoney\"\u003e\u003ccode\u003e​shoemoney\u003c/code\u003e\u003c/a\u003e in \u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/pull/4201\"\u003ehyperium/hyper#4201\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003efix(http1): always remove content-length if transfer-encoding is sent by \u003ca href\u003d\"https://github.com/seanmonstar\"\u003e\u003ccode\u003e​seanmonstar\u003c/code\u003e\u003c/a\u003e in \u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/pull/4223\"\u003ehyperium/hyper#4223\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003efix(http1): remove invalid content-length if transfer-encoding is auto-added by \u003ca href\u003d\"https://github.com/seanmonstar\"\u003e\u003ccode\u003e​seanmonstar\u003c/code\u003e\u003c/a\u003e in \u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/pull/4224\"\u003ehyperium/hyper#4224\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003efix(http1): withdraw a stale want when the client dispatcher takes a request by \u003ca href\u003d\"https://github.com/shodoco\"\u003e\u003ccode\u003e​shodoco\u003c/code\u003e\u003c/a\u003e in \u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/pull/4208\"\u003ehyperium/hyper#4208\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003eperf(body): skip waking a closed peer when a channel end drops by \u003ca href\u003d\"https://github.com/0x676e67\"\u003e\u003ccode\u003e​0x676e67\u003c/code\u003e\u003c/a\u003e in \u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/pull/4225\"\u003ehyperium/hyper#4225\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003ePrepare v1.12.0 by \u003ca href\u003d\"https://github.com/seanmonstar\"\u003e\u003ccode\u003e​seanmonstar\u003c/code\u003e\u003c/a\u003e in \u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/pull/4226\"\u003ehyperium/hyper#4226\u003c/a\u003e\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch2\u003eNew Contributors\u003c/h2\u003e\n\u003cul\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/vaibhav8a\"\u003e\u003ccode\u003e​vaibhav8a\u003c/code\u003e\u003c/a\u003e made their first contribution in \u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/pull/4177\"\u003ehyperium/hyper#4177\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/dox4\"\u003e\u003ccode\u003e​dox4\u003c/code\u003e\u003c/a\u003e made their first contribution in \u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/pull/4173\"\u003ehyperium/hyper#4173\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/dtolnay\"\u003e\u003ccode\u003e​dtolnay\u003c/code\u003e\u003c/a\u003e made their first contribution in \u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/pull/4196\"\u003ehyperium/hyper#4196\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/Dev-next-gen\"\u003e\u003ccode\u003e​Dev-next-gen\u003c/code\u003e\u003c/a\u003e made their first contribution in \u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/pull/4198\"\u003ehyperium/hyper#4198\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/Siech0\"\u003e\u003ccode\u003e​Siech0\u003c/code\u003e\u003c/a\u003e made their first contribution in \u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/pull/4191\"\u003ehyperium/hyper#4191\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/toyeshhm\"\u003e\u003ccode\u003e​toyeshhm\u003c/code\u003e\u003c/a\u003e made their first contribution in \u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/pull/4204\"\u003ehyperium/hyper#4204\u003c/a\u003e\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/blockquote\u003e\n\u003cp\u003e... (truncated)\u003c/p\u003e\n\u003c/details\u003e\n\u003cdetails\u003e\n\u003csummary\u003eChangelog\u003c/summary\u003e\n\u003cp\u003e\u003cem\u003eSourced from \u003ca href\u003d\"https://github.com/hyperium/hyper/blob/master/CHANGELOG.md\"\u003ehyper\u0027s changelog\u003c/a\u003e.\u003c/em\u003e\u003c/p\u003e\n\u003cblockquote\u003e\n\u003ch2\u003ev1.12.0 (2026-10-06)\u003c/h2\u003e\n\u003ch4\u003eBug Fixes\u003c/h4\u003e\n\u003cul\u003e\n\u003cli\u003e\u003cstrong\u003ehttp1:\u003c/strong\u003e\n\u003cul\u003e\n\u003cli\u003ewithdraw a stale want when the client dispatcher takes a request (\u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/issues/4208\"\u003e#4208\u003c/a\u003e) (\u003ca href\u003d\"https://github.com/hyperium/hyper/commit/c44f1cc51c2bc45b502031a646d3d54315bd7a16\"\u003ec44f1cc5\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eremove content-length if transfer-encoding is auto-added (\u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/issues/4224\"\u003e#4224\u003c/a\u003e) (\u003ca href\u003d\"https://github.com/hyperium/hyper/commit/69396fdd960629d69b09cc6c9b367ef8650c001d\"\u003e69396fdd\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003ealways remove content-length if transfer-encoding is sent (\u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/issues/4223\"\u003e#4223\u003c/a\u003e) (\u003ca href\u003d\"https://github.com/hyperium/hyper/commit/a5d7fd7861598c0b8c857cb75240a75bcfbd96a3\"\u003ea5d7fd78\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003elet a Connection close token win over a later keep-alive (\u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/issues/4201\"\u003e#4201\u003c/a\u003e) (\u003ca href\u003d\"https://github.com/hyperium/hyper/commit/5dadff76d1cb5d21046218df3568569fbba8ed21\"\u003e5dadff76\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003epreserve hop-by-hop when setting close or keep-alive (\u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/issues/4196\"\u003e#4196\u003c/a\u003e) (\u003ca href\u003d\"https://github.com/hyperium/hyper/commit/74a39081ff1496970fb75da75c5282b23a9b69e5\"\u003e74a39081\u003c/a\u003e)\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003ehttp2:\u003c/strong\u003e do not reserve capacity for idle Upgraded streams (\u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/issues/4198\"\u003e#4198\u003c/a\u003e) (\u003ca href\u003d\"https://github.com/hyperium/hyper/commit/e89d0d93d9566210ee95aea234a00952ebc4a202\"\u003ee89d0d93\u003c/a\u003e)\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch4\u003eFeatures\u003c/h4\u003e\n\u003cul\u003e\n\u003cli\u003e\u003cstrong\u003ehttp1:\u003c/strong\u003e add max_header_size limit for server and client (\u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/issues/4183\"\u003e#4183\u003c/a\u003e) (\u003ca href\u003d\"https://github.com/hyperium/hyper/commit/d1cd14eff392be63a2318d690cc9380723ec61a1\"\u003ed1cd14ef\u003c/a\u003e, closes \u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/issues/3832\"\u003e#3832\u003c/a\u003e)\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch3\u003ev1.11.1 (2026-08-27)\u003c/h3\u003e\n\u003ch4\u003eBug Fixes\u003c/h4\u003e\n\u003cul\u003e\n\u003cli\u003e\u003cstrong\u003ehttp1:\u003c/strong\u003e\n\u003cul\u003e\n\u003cli\u003edetect TE: trailers caselessly and with other values (\u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/issues/4152\"\u003e#4152\u003c/a\u003e) (\u003ca href\u003d\"https://github.com/hyperium/hyper/commit/72fcf545c9b4a261f31d89bafd3019a2d4e9e1d7\"\u003e72fcf545\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003erecognize \u003ccode\u003e\\n\\r\\n\u003c/code\u003e as a head terminator in the partial-read fast path (\u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/issues/4147\"\u003e#4147\u003c/a\u003e) (\u003ca href\u003d\"https://github.com/hyperium/hyper/commit/3534d75c898da17b023e6bff9b4cf71afed123ad\"\u003e3534d75c\u003c/a\u003e, closes \u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/issues/4145\"\u003e#4145\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eflush bytes buffered by the write re-check before yielding (\u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/issues/4143\"\u003e#4143\u003c/a\u003e) (\u003ca href\u003d\"https://github.com/hyperium/hyper/commit/f660f5bf6eed3fe793f899507ff5bb9e266d4b0d\"\u003ef660f5bf\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003eevict pooled conn on request-side Connection: close (\u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/issues/4110\"\u003e#4110\u003c/a\u003e) (\u003ca href\u003d\"https://github.com/hyperium/hyper/commit/da19be08f34d3da01c4c0f9cfe2837e7d377f73d\"\u003eda19be08\u003c/a\u003e)\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/blockquote\u003e\n\u003c/details\u003e\n\u003cdetails\u003e\n\u003csummary\u003eCommits\u003c/summary\u003e\n\u003cul\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/hyperium/hyper/commit/7f67699ce32d717081deb7018b4da7d679277ac0\"\u003e\u003ccode\u003e7f67699\u003c/code\u003e\u003c/a\u003e v1.12.0\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/hyperium/hyper/commit/254bf1271bf4385edd743316ca7c508b3c77d6cb\"\u003e\u003ccode\u003e254bf12\u003c/code\u003e\u003c/a\u003e perf(body): skip waking a closed peer when a channel end drops (\u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/issues/4225\"\u003e#4225\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/hyperium/hyper/commit/c44f1cc51c2bc45b502031a646d3d54315bd7a16\"\u003e\u003ccode\u003ec44f1cc\u003c/code\u003e\u003c/a\u003e fix(http1): withdraw a stale want when the client dispatcher takes a request ...\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/hyperium/hyper/commit/69396fdd960629d69b09cc6c9b367ef8650c001d\"\u003e\u003ccode\u003e69396fd\u003c/code\u003e\u003c/a\u003e fix(http1): remove content-length if transfer-encoding is auto-added (\u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/issues/4224\"\u003e#4224\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/hyperium/hyper/commit/a5d7fd7861598c0b8c857cb75240a75bcfbd96a3\"\u003e\u003ccode\u003ea5d7fd7\u003c/code\u003e\u003c/a\u003e fix(http1): always remove content-length if transfer-encoding is sent (\u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/issues/4223\"\u003e#4223\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/hyperium/hyper/commit/5dadff76d1cb5d21046218df3568569fbba8ed21\"\u003e\u003ccode\u003e5dadff7\u003c/code\u003e\u003c/a\u003e fix(http1): let a Connection close token win over a later keep-alive (\u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/issues/4201\"\u003e#4201\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/hyperium/hyper/commit/d1cd14eff392be63a2318d690cc9380723ec61a1\"\u003e\u003ccode\u003ed1cd14e\u003c/code\u003e\u003c/a\u003e feat(http1): add max_header_size limit for server and client (\u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/issues/4183\"\u003e#4183\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/hyperium/hyper/commit/714c7cf46fd405c88eb34c2f243ae5b146ba8f07\"\u003e\u003ccode\u003e714c7cf\u003c/code\u003e\u003c/a\u003e style(lib): address cast-possible-truncation clippy findings (\u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/issues/4192\"\u003e#4192\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/hyperium/hyper/commit/0d474402c20c17939c30aaecde56d1a3e21b2b0a\"\u003e\u003ccode\u003e0d47440\u003c/code\u003e\u003c/a\u003e style(lib): mark partial Debug impls as non-exhaustive (\u003ca href\u003d\"https://redirect.github.com/hyperium/hyper/issues/4218\"\u003e#4218\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/hyperium/hyper/commit/da3b1b197c954ac5454a92562fe3d729a87cba56\"\u003e\u003ccode\u003eda3b1b1\u003c/code\u003e\u003c/a\u003e style(lib): remove useless_borrows_in_formatting from the clippy allowlist (#...\u003c/li\u003e\n\u003cli\u003eAdditional commits viewable in \u003ca href\u003d\"https://github.com/hyperium/hyper/compare/v1.11.0...v1.12.0\"\u003ecompare view\u003c/a\u003e\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/details\u003e\n\u003cbr /\u003e\n\n[![Dependabot compatibility score](https://dependabot-badges.githubapp.com/badges/compatibility_score?dependency-name\u003dhyper\u0026package-manager\u003dcargo\u0026previous-version\u003d1.11.0\u0026new-version\u003d1.12.0)](https://docs.github.com/en/github/managing-security-vulnerabilities/about-dependabot-security-updates#about-compatibility-scores)\n\nDependabot will resolve any conflicts with this PR as long as you don\u0027t alter it yourself. You can also trigger a rebase manually by commenting `dependabot rebase`.\n\n[//]: # (dependabot-automerge-start)\n[//]: # (dependabot-automerge-end)\n\n---\n\n\u003cdetails\u003e\n\u003csummary\u003eDependabot commands and options\u003c/summary\u003e\n\u003cbr /\u003e\n\nYou can trigger Dependabot actions by commenting on this PR:\n- `dependabot rebase` will rebase this PR\n- `dependabot recreate` will recreate this PR, overwriting any edits that have been made to it\n- `dependabot show \u003cdependency name\u003e ignore conditions` will show all of the ignore conditions of the specified dependency\n- `dependabot ignore this major version` will close this PR and stop Dependabot creating any more for this major version (unless you reopen the PR or upgrade to it yourself)\n- `dependabot ignore this minor version` will close this PR and stop Dependabot creating any more for this minor version (unless you reopen the PR or upgrade to it yourself)\n- `dependabot ignore this dependency` will close this PR and stop Dependabot creating any more for this dependency (unless you reopen the PR or upgrade to it yourself)\n\n\u003c/details\u003e\n\nCloses #49 from dependabot[bot]/dependabot/cargo/hyper-1.12.0.\n\nAuthored-by: dependabot[bot] \u003c49699333+dependabot[bot]@users.noreply.github.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\n"
    },
    {
      "commit": "d54a1804f81de3033e1398fb5b6921238e29d20f",
      "tree": "076604f743ca312fdf11dd4afe15b4f532a6f7ce",
      "parents": [
        "6fdd890b570fc72ca219285281fc774909e7b785"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Tue Oct 27 23:26:51 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Tue Oct 27 23:26:51 2026 -0700"
      },
      "message": "[SPARK-60591] Ignore RUSTSEC-2026-0387 (crypto-bigint) in cargo audit\n\n### What changes were proposed in this pull request?\n\nAdd RUSTSEC-2026-0387 (`crypto-bigint`) to the ignore list in\n`.cargo/audit.toml`, next to RUSTSEC-2023-0071 (`rsa`), with the reasoning and\na pointer to SPARK-60591.\n\nSPARK-60591 also serves as the tracking issue for this ignore, in the same role\nSPARK-60369 has for `rsa`, so it should **stay open when this is merged**. It\nlists the conditions for removing the entry.\n\n### Why are the changes needed?\n\n`cargo audit` reports RUSTSEC-2026-0387 against `crypto-bigint` 0.5.5: timing\nvariability in operations intended to be constant-time (published 2026-10-23).\nOnce SPARK-60590 (#54) lands, it is the scheduled audit run\u0027s only remaining\nfailure.\n\n**No fix is reachable.** The fix is in `crypto-bigint` 0.7.6. The crate\narrives through jsonwebtoken\u0027s `rust_crypto` provider: `p256` / `p384` (ES256 /\nES384) -\u003e `ecdsa` 0.16 -\u003e `elliptic-curve` 0.13 -\u003e `crypto-bigint` 0.5.\njsonwebtoken 11 (#39) resolves the same versions, so upgrading it does not\nhelp.\n\n**The exposure is low**, for the same reason as `rsa`. Timing variability in\nconstant-time arithmetic matters when the operands are secret, i.e. in\nprivate-key operations. Production code only verifies signatures with public\nkeys: it uses `DecodingKey` (`from_ec_pem`, `from_ec_components`, and the RSA,\nEd25519 and HMAC equivalents), and every `EncodingKey` or signing use is inside\n`#[cfg(test)]`.\n\nA longer-term option that would remove both this advisory and the `rsa` one is\njsonwebtoken\u0027s `aws_lc_rs` provider, which drops the RustCrypto `rsa` and\n`elliptic-curve` stacks. It brings a C build dependency (aws-lc needs cmake)\nand new licenses to review, which SPARK-59552 deliberately avoided, so it is not\nproposed here.\n\n### Does this PR introduce _any_ user-facing change?\n\nNo. CI configuration only.\n\n### How was this patch tested?\n\nThe ignore is by advisory ID, so it covers nothing else. Checked with\n`cargo audit` and the new config:\n\n| Lockfile | Result |\n|---|---|\n| `main`\u0027s | exit 1: `crypto-bigint` and `rsa` ignored; the two webpki advisories fixed by #54 still reported |\n| #54\u0027s (i.e. `main` once both land) | **exit 0**: no vulnerabilities, only the unmaintained (`unsafe-libyaml`) and yanked (`chacha20`) warnings, which do not fail the run |\n\nSo with this and #54 merged, the scheduled run on `main` turns green.\n\n### Was this patch authored or co-authored using generative AI tooling?\n\nYes, co-authored with Claude Code.\n\nCloses #55 from viirya/ci-13-ignore-crypto-bigint.\n\nAuthored-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\n"
    },
    {
      "commit": "6fdd890b570fc72ca219285281fc774909e7b785",
      "tree": "a35b05a8caf336b5798f492a2d153d8dad4c84c5",
      "parents": [
        "3c61a474ea79971fa6c15c01252c11f423995666"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Tue Oct 27 23:26:16 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Tue Oct 27 23:26:16 2026 -0700"
      },
      "message": "[SPARK-60590] Update webpki-roots, rustls-webpki and tokio in Cargo.lock to pick up recent RustSec advisories\n\n### What changes were proposed in this pull request?\n\nUpdate three crates in `Cargo.lock`. Lockfile-only; no `Cargo.toml` change.\n\n```\ncargo update -p webpki-roots --precise 1.0.10\ncargo update -p rustls-webpki --precise 0.103.16\ncargo update -p tokio --precise 1.53.3\n```\n\nThe resolved set is **431 crates before and after**, and these three are the\nonly versions that change. The diff also re-points four dependents from\n`windows-sys` 0.61.2 to 0.52.0. That is not a downgrade: both versions are in\nthe graph before and after, and they are Windows-only.\n\n### Why are the changes needed?\n\nThe scheduled `cargo audit` run on `main` has failed daily since it was\nenabled in SPARK-60370 (#52). Two of its findings, plus one warning, have\nfixed releases:\n\n| Advisory | Crate | Issue |\n|---|---|---|\n| RUSTSEC-2026-0381 | `webpki-roots` 1.0.9 -\u003e 1.0.10 | Bundled root certificates allow trust in distrusted emSign intermediates |\n| RUSTSEC-2026-0382 | `rustls-webpki` 0.103.14 -\u003e 0.103.16 | Certificate verification may accept distrusted intermediates |\n| RUSTSEC-2026-0386 (warning: unsound) | `tokio` 1.53.1 -\u003e 1.53.3 | `AsyncFd::new` and related constructors violate I/O safety |\n\nThe two webpki advisories affect the gateway: both crates are compiled into it\nand sit on the outbound TLS certificate-verification path (`kube-client` for the\nK8s API, `reqwest` for OIDC JWKS, tonic / `opentelemetry-otlp`).\n\n**Not addressed here:** RUSTSEC-2026-0387 (`crypto-bigint` 0.5.5, timing\nvariability in operations intended to be constant-time), the scheduled run\u0027s\nremaining failure. The fix is in `crypto-bigint` 0.7.6, which nothing in the\ncurrent dependency graph can use; jsonwebtoken 11 (#39) still resolves 0.5.5.\nIt will be handled separately.\n\n### Does this PR introduce _any_ user-facing change?\n\nNo. No configuration or API change.\n\n### How was this patch tested?\n\nOn the pinned 1.99.0 toolchain:\n\n- `cargo audit` reports **1 vulnerability (crypto-bigint) instead of 3**, and\n  no longer warns about `tokio`.\n- `cargo clippy --workspace --all-targets -- -D warnings` -- clean.\n- `cargo fmt --all -- --check` -- clean.\n- `cargo test --workspace` -- 204 passed, 0 failed, 10 ignored. `tokio` is the\n  gateway\u0027s async runtime, so this exercises the new version.\n- `dev/audit-new-advisories.sh` against `main`\u0027s `Cargo.lock` (what this PR\u0027s\n  `cargo audit` check runs) exits 0, listing the two webpki advisories as\n  resolved and `crypto-bigint` as already present.\n- Licenses are unchanged and all Category A: `rustls-webpki` ISC, `tokio` MIT,\n  `webpki-roots` CDLA-Permissive-2.0 (Category A per LEGAL-732).\n- The lockfile claim was checked by comparing the full resolved package set\n  (name + version for all 431 entries) against `main`.\n\n### Was this patch authored or co-authored using generative AI tooling?\n\nYes, co-authored with Claude Code.\n\nCloses #54 from viirya/dep-07-webpki-tokio.\n\nAuthored-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\n"
    },
    {
      "commit": "3c61a474ea79971fa6c15c01252c11f423995666",
      "tree": "40e99782c5b8d22123135e7a4c4bae55687ce90e",
      "parents": [
        "4e6676dbd51f1b375f5a4c5f065d8c9533d43f0f"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Tue Oct 27 13:42:55 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Tue Oct 27 13:42:55 2026 -0700"
      },
      "message": "[SPARK-60526] Use the Apache product name, fix docs that point outside the repository, and tidy the Dockerfile\n\n### What changes were proposed in this pull request?\n\nCleanups found while preparing the first release. None changes runtime\nbehaviour.\n\n**Product name.** Use \"Apache Spark Connect Gateway\" in the two user-visible\nplaces that said \"Open-source Spark Connect Gateway\" -- the clap `about` text\nshown by `gateway --help` (`crates/gateway/src/main.rs`) and the Helm chart\ndescription (`deploy/helm/scg/Chart.yaml`) -- and as the README title, which\nwas the bare repository name.\n\n**Links outside the repository.** `README.md` and `docs/architecture.md`\nlinked to documents under `../plans/`, which is not part of this repository.\nBoth now point at the SPIP that proposed the project,\n[SPARK-58455](https://issues.apache.org/jira/browse/SPARK-58455), which states\nthe problem and links to the design document. In the README the two link\ndefinitions for those files are replaced by one for the SPIP, and the\nnumbered links are renumbered in order of first use, so the text reads [1] to\n[4] from top to bottom (previously [5] appeared before [4]).\n\nThe README also called the feature set \"Production-ready\"; it now just lists\nthe features.\n\n**Dockerfile.**\n\n- The builder was `rust:1.82-slim` while `rust-toolchain.toml` pins 1.99.0,\n  so rustup installed a second toolchain on every build. It is now\n  `rust:1.99.0-slim-bookworm`; rustup only adds the `rustfmt` and `clippy`\n  components `rust-toolchain.toml` lists.\n- The builder now names its Debian release, with a comment explaining why it\n  must match the runtime image (see below).\n- `libssl-dev` and `pkg-config` are no longer installed. Nothing in the\n  dependency graph uses `openssl-sys` -- TLS is rustls with ring -- and\n  `pkg-config` was only there to locate libssl.\n\n### Why are the changes needed?\n\nThe ASF name for the product is \"Apache Spark Connect Gateway\", and it is what\nusers see in `--help` and in the chart.\n\nThe `../plans/` links are broken for every reader of the Apache repository and\nof a source release.\n\nThe Debian pin matters because of how glibc versions work. The runtime stage\nis `distroless/cc-debian12` (bookworm), and the binary requires the glibc\nsymbol versions of the glibc it was *linked* against in the builder. glibc is\nonly backward compatible, so if the unpinned `rust:*-slim` tag moved to a newer\nDebian, the build could produce a binary that fails to start on the runtime\nimage with ``version `GLIBC_x.yy\u0027 not found``. Pinning the builder to bookworm\nkeeps both stages on the same release; the comment says to move them together.\n\n### Does this PR introduce _any_ user-facing change?\n\nYes, cosmetic only: `gateway --help` prints \"Apache Spark Connect Gateway\"\ninstead of \"Open-source Spark Connect Gateway\", and the Helm chart description\nchanges the same way. No configuration, API, or behaviour change.\n\n### How was this patch tested?\n\n- `cargo fmt --all -- --check` and `cargo clippy --workspace --all-targets --\n  -D warnings` -- clean.\n- `cargo test --workspace` -- 204 passed, 0 failed.\n- `gateway --help` prints the new name; `helm lint deploy/helm/scg` passes.\n- The image was built for linux/amd64 from this Dockerfile without\n  `libssl-dev` or `pkg-config`, and the resulting image runs `gateway\n  --version` and `gateway --help`.\n\nHow the Dockerfile facts were established: `gateway` was built for\nlinux/amd64 with the pinned toolchain on bookworm. `readelf -d` shows it needs\nonly `libc.so.6`, `libm.so.6`, `libgcc_s.so.1` and `ld-linux-x86-64.so.2` (no\nlibssl, no libstdc++), with `GLIBC_2.34` as the highest required symbol\nversion against 2.36 in `distroless/cc-debian12`. The `rust:1.99.0-slim-bookworm`\nimage ships `cargo`, `rustc` and `rust-std` only, which is why rustup still\nadds `rustfmt` and `clippy`.\n\nThe e2e jobs build the image from this Dockerfile as well, so CI exercises it\nindependently.\n\n### Was this patch authored or co-authored using generative AI tooling?\n\nYes, co-authored with Claude Code.\n\nCloses #53 from viirya/docs-01-release-cleanup.\n\nAuthored-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\n"
    },
    {
      "commit": "4e6676dbd51f1b375f5a4c5f065d8c9533d43f0f",
      "tree": "a6a0df55f2e829eafef10999b82202cd2e8448b2",
      "parents": [
        "2de43b973fa5ce6783d55617b43ea2a5675a8e11"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Fri Oct 23 09:29:17 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Fri Oct 23 09:29:17 2026 -0700"
      },
      "message": "[SPARK-60370] Run cargo audit in CI on a daily schedule and on dependency changes\n\n### What changes were proposed in this pull request?\n\nRun `cargo audit` in CI, checking `Cargo.lock` against the RustSec advisory\ndatabase.\n\n| File | Purpose |\n|---|---|\n| `.github/workflows/audit.yml` | The workflow (triggers below) |\n| `dev/audit-new-advisories.sh` | Pull-request check: report only advisories the change introduces |\n| `.cargo/audit.toml` | Ignores RUSTSEC-2023-0071 (`rsa`), pointing at SPARK-60369 |\n\nThe workflow keeps two questions apart:\n\n- **Is `main` affected by a known advisory?** A daily schedule (and\n  `workflow_dispatch`) runs a full `cargo audit` on `main`.\n- **Does this change introduce one?** Pull requests that touch `Cargo.lock`, a\n  `Cargo.toml`, `.cargo/audit.toml`, or the audit tooling run\n  `dev/audit-new-advisories.sh`. It audits the base revision\u0027s `Cargo.lock` and\n  the pull request\u0027s against the same snapshot of the database, and fails only\n  on advisory/crate pairs that the base does not have. Advisories already on\n  `main` are listed but not counted against the pull request.\n\nThere is **no push trigger**. A merge has already been checked as a pull\nrequest, and a full audit on push would mark every dependency merge red while\n`main` carries any unfixed advisory, including merges that fix one.\n\n`dev/audit-new-advisories.sh` exits 0 when nothing is introduced, 1 when\nsomething is, and 2 when `cargo audit` cannot produce a report. `cargo audit`\nitself exits 1 on findings, so its status cannot tell a finding from a failure\n(for example, the database could not be fetched); the script requires a JSON\nreport and treats anything else as an error rather than a pass. Pairs are\ncompared without versions, so moving a crate from one affected version to\nanother does not count as new.\n\n`cargo-audit` is installed with `cargo install --version 0.22.1 --locked`\ninstead of a third-party action, to stay within the Apache GitHub Actions\nallowlist. Checkout and toolchain setup are the same pinned steps `ci.yml` uses.\n\n### Why are the changes needed?\n\nNothing in CI checked dependencies for known vulnerabilities. Three advisories\n(`h2`, `rustls`, `rsa`) sat on `main` with CI green and were found only by\nrunning `cargo audit` by hand. The `h2` fix from SPARK-60115 is a `--precise`\npin on a transitive dependency, which a later `cargo update` or a `hyper` /\n`tonic` bump could silently undo.\n\nThe reason for not simply failing on any advisory: the database changes daily,\nso such a check would turn pull requests red for vulnerabilities they did not\nintroduce -- the failure mode SPARK-59989 fixed for the floating Rust toolchain.\nWhile this change was being prepared, eleven new advisories affecting the\nlockfile were published, so this is not hypothetical.\n\n**The `rsa` ignore.** RUSTSEC-2023-0071 (Marvin Attack) has no fixed release:\n0.9.10 is the latest stable and 0.10 exists only as a release candidate. The\nattack recovers a private key from the timing of private-key operations. The\ngateway only verifies RS256 signatures with public keys through `jsonwebtoken`;\nevery `EncodingKey` / `RsaPrivateKey` use is inside `#[cfg(test)]`, and the\ndirect `rsa` dependency is a dev-dependency. SPARK-60369 tracks removing the\nignore and lists when to revisit it.\n\n**What to expect after merge.** The scheduled run will be red until the\nadvisories currently on `main` are fixed: #51 (SPARK-60371) addresses nine of\nthem, and RUSTSEC-2026-0381 / -0382 (`webpki-roots`, `rustls-webpki`) will be\nhandled separately. That is the intended signal. Pull requests are not\naffected by it.\n\n### Does this PR introduce _any_ user-facing change?\n\nNo. CI only.\n\n### How was this patch tested?\n\n`dev/audit-new-advisories.sh` was run locally against real lockfiles:\n\n| Base | Head | Result |\n|---|---|---|\n| `main` | `main` | exit 0; existing advisories listed as already present |\n| `main` | #51\u0027s lockfile | exit 0; nine advisories listed as resolved |\n| `main` | `main` before SPARK-60115 (`h2` 0.4.15) | **exit 1**; RUSTSEC-2026-0258 reported as introduced |\n| -- | no arguments / missing file | exit 2 |\n| `main` | unparseable lockfile (also tried as base) | exit 2, not a pass |\n\nRUSTSEC-2023-0071 is absent from every report, confirming `.cargo/audit.toml`\nis honoured when a lockfile is given with `--file`.\n\nThe base-lockfile step (`git fetch --no-tags --depth\u003d1 origin \u003cbase sha\u003e` with\ncredentials not persisted, then `git show \u003csha\u003e:Cargo.lock`) was exercised in a\nfresh shallow clone of this repository that did not contain the commit.\n\nThis pull request does not change `Cargo.lock`, so its own pull-request check\nshould report no introduced advisories.\n\n### Was this patch authored or co-authored using generative AI tooling?\n\nYes, co-authored with Claude Code.\n\nCloses #52 from viirya/ci-12-cargo-audit.\n\nAuthored-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\n"
    },
    {
      "commit": "2de43b973fa5ce6783d55617b43ea2a5675a8e11",
      "tree": "e9f628bba5b1bdbee2f9c7eef76bc068ff79cc38",
      "parents": [
        "af4f0802169e4987b3a864ad4c32c9f8022ab38b"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Wed Oct 21 23:04:30 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Wed Oct 21 23:04:30 2026 -0700"
      },
      "message": "[SPARK-60371] Update quinn-proto and rustls in Cargo.lock to pick up recent RustSec advisories\n\n### What changes were proposed in this pull request?\n\nUpdate two crates in `Cargo.lock` to pick up fixes for nine RustSec advisories.\nLockfile-only; no `Cargo.toml` change is needed.\n\n```\ncargo update -p quinn-proto --precise 0.11.19\ncargo update -p rustls --precise 0.23.46\n```\n\nThe resolved set is **431 crates before and after**; `quinn-proto`\n(0.11.16 -\u003e 0.11.19) and `rustls` (0.23.45 -\u003e 0.23.46) are the only two whose\nversion changes.\n\n### Why are the changes needed?\n\n`cargo audit` against `main` (af4f080) reports these, all with fixed releases.\n\n**`rustls` 0.23.45 -\u003e 0.23.46.** These apply: the gateway uses rustls as a TLS\nclient (`kube-client` for the K8s API, `reqwest` for OIDC JWKS, tonic /\n`opentelemetry-otlp`).\n\n| Advisory | Issue |\n|---|---|\n| RUSTSEC-2026-0346 | `Debug` impls for client session values expose TLS secrets |\n| RUSTSEC-2026-0347 | buffered application data sent to outer server after ECH rejection |\n| RUSTSEC-2026-0348 | brotli certificate decompression can allocate up to 1 GiB before authentication |\n| RUSTSEC-2026-0349 | unauthenticated ServerHello can downgrade later connections to classical key exchange |\n\n**`quinn-proto` 0.11.16 -\u003e 0.11.19.** The gateway is not exposed to these.\n`quinn` is enabled only by reqwest\u0027s `http3` feature, which the workspace does\nnot turn on (`reqwest` is built with `rustls-tls` and `json`), so `quinn-proto`\nappears in `Cargo.lock` but is never compiled. `cargo audit` reads the lockfile\nrather than the build graph, so it still reports them, and updating is\npreferable to ignoring advisories that have a fix.\n\n| Advisory | Issue |\n|---|---|\n| RUSTSEC-2026-0338 | zero-length STREAM frames can panic during unordered reads |\n| RUSTSEC-2026-0339 | unbounded growth of pending connection ID retirements |\n| RUSTSEC-2026-0340 | peer minimum ACK delay can panic with ACK_FREQUENCY |\n| RUSTSEC-2026-0341 | stream reassembly memory-exhaustion guard can be bypassed |\n| RUSTSEC-2026-0342 | zero-length DATAGRAM frames bypass the receive buffer limit |\n\nThis is also a step toward SPARK-60370 (running `cargo audit` in CI).\n\n**Not covered here:** RUSTSEC-2026-0381 (`webpki-roots`) and\nRUSTSEC-2026-0382 (`rustls-webpki`) were published after this change was\nprepared and will be handled separately.\n\n### Does this PR introduce _any_ user-facing change?\n\nNo. No configuration or API change. The behavioural effect is confined to the\noutbound TLS client fixes listed above.\n\n### How was this patch tested?\n\nOn the pinned 1.99.0 toolchain:\n\n- `cargo audit` no longer reports any `quinn-proto` or `rustls` advisory.\n- `cargo clippy --workspace --all-targets -- -D warnings` -- clean.\n- `cargo fmt --all -- --check` -- clean.\n- `cargo test --workspace` -- 204 passed, 0 failed, 10 ignored, matching the\n  baseline on `main`. `rustls` is compiled into the gateway, so this exercises\n  the new version.\n- The build produces no `quinn` artifact, confirming `quinn-proto` is not\n  compiled.\n- The lockfile claim was checked by comparing the full resolved package set\n  (name + version for all 431 entries) against `main`, not by reading the diff.\n\n### Was this patch authored or co-authored using generative AI tooling?\n\nYes, co-authored with Claude Code.\n\nCloses #51 from viirya/dep-06-quinn-rustls.\n\nAuthored-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\n"
    },
    {
      "commit": "af4f0802169e4987b3a864ad4c32c9f8022ab38b",
      "tree": "6dd7a1b3e56b17dfae36fb19c542e082270c1a89",
      "parents": [
        "5db2a73a0464ca042b5d52c4bc39c67a15042ac8"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "liangchi.hsieh@databricks.com",
        "time": "Mon Oct 12 22:18:59 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "liangchi.hsieh@databricks.com",
        "time": "Mon Oct 12 22:18:59 2026 -0700"
      },
      "message": "[SPARK-60115] Pin the transitive h2 dependency to 0.4.16 to pick up RUSTSEC-2026-0258\n\n### What changes were proposed in this pull request?\n\nPin the transitive `h2` dependency to 0.4.16 to pick up the fix for\nRUSTSEC-2026-0258.\n\n`h2` is not a direct dependency -- it appears nowhere in `Cargo.toml`, arriving\nvia `hyper` and `tonic` 0.14.6 -- so this is a lockfile-only change:\n\n```\ncargo update -p h2 --precise 0.4.16\n```\n\nThe resolved package set is **431 crates before and after**, and `h2 0.4.15 -\u003e\n0.4.16` is the only version that moves.\n\nThe diff also shows four `windows-sys 0.61.2 -\u003e 0.52.0` lines inside other\npackages\u0027 dependency lists. **These are not a downgrade.** Both 0.52.0 and\n0.61.2 are present in the graph before and after this change; cargo only\nre-pointed some dependents at the copy that was already there. They are\nWindows-only crates and CI is `ubuntu-latest` throughout, so they affect no\nbuild here.\n\n### Why are the changes needed?\n\n`cargo audit` reported three advisories on `main` at 131c830. One has since\nbeen fixed by #47; this addresses the remaining one that is both server-side and\nhas a fix available:\n\n```\nRUSTSEC-2026-0258 -- h2 unbounded empty DATA frames\nhttps://github.com/hyperium/hyper/security/advisories/GHSA-q83h-524g-xf6h\naffected: h2 0.4.15, patched: \u003e\u003d 0.4.16\n```\n\n`h2` reaches the gateway through `hyper` -\u003e `tonic` 0.14.6, and the tonic server\n*is* the gateway\u0027s own client-facing listener on `:15003`. A peer that can open\na connection and send unbounded empty DATA frames is the threat model of a\ncluster- or internet-facing gRPC proxy, not a hypothetical.\n\n#### Scope: what this does not fix\n\nThe other two advisories are deliberately left out of this change.\n\n- **RUSTSEC-2026-0285** -- `rustls` 0.23.43, \"TLS 1.3 handshake messages\n  incorrectly accepted across encryption level boundaries\", 5.3 medium.\n  **Already fixed on `main` by #47** (rustls 0.23.45), which this branch is now\n  rebased onto, so it is out of scope here. For the record the gateway\u0027s\n  exposure was limited: it does not terminate TLS (no `ServerConfig` / `TlsAcceptor` / `ServerTlsConfig` anywhere\n  under `crates/`, and `:15003` is plaintext), and pulls `rustls` only on client\n  paths -- `kube-client` for the K8s API, `reqwest` for OIDC JWKS, and\n  `opentelemetry-otlp`\u0027s `tls-ring`.\n\n- **RUSTSEC-2023-0071** -- `rsa` 0.9.10, Marvin Attack timing sidechannel, 5.9\n  medium. No fixed version exists, so there is nothing to bump to. It arrives\n  via `jsonwebtoken`\u0027s RSA backend under the `rust_crypto` provider selected in\n  SPARK-59552, and needs a decision -- accept the risk, or change backend --\n  rather than a version bump. Possibly worth resolving alongside #39.\n\n#### A limitation of this fix\n\nA `--precise` pin on a transitive dependency is not durable: the next\n`cargo update`, or a dependabot bump of `hyper` or `tonic`, can re-resolve `h2`\nand silently drop back below 0.4.16.\n\nThe durable fix is running `cargo audit` in CI, which is **not wired up today**\n-- all three advisories reached `main` unnoticed and were found only by running\nit by hand. That is worth its own change, and this PR does not attempt it.\n\n### Does this PR introduce _any_ user-facing change?\n\nNo. No configuration, API, or behavioural change; the only effect is that the\nHTTP/2 implementation underneath the gRPC server no longer accepts an unbounded\nstream of empty DATA frames.\n\n### How was this patch tested?\n\nOn the pinned 1.99.0 toolchain:\n\n- `cargo audit` now reports **1 vulnerability, down from 2 on `main`**. The\n  only remaining finding is `rsa` / RUSTSEC-2023-0071, which has no fixed\n  version. (This PR was rebased onto `main` after #47 landed, so the `rustls`\n  finding is already gone from the baseline; an earlier revision of this\n  description said \"2 instead of 3\", measured before that merge.)\n- `cargo clippy --workspace --all-targets -- -D warnings` -- clean.\n- `cargo fmt --all -- --check` -- clean.\n- `cargo test --workspace` -- **204 passed, 0 failed, 10 ignored**, matching the\n  baseline on `main`.\n\nThe lockfile claim was verified by comparing the full resolved package set\n(name + version for all 431 entries) before and after, rather than by reading\nthe diff -- `cargo update --dry-run` had reported only the `h2` line and did not\nmention the `windows-sys` re-pointing.\n\n### Was this patch authored or co-authored using generative AI tooling?\n\nYes, co-authored with Claude Code.\n\nCloses #48 from viirya/dep-05-h2-security.\n\nAuthored-by: Liang-Chi Hsieh \u003cliangchi.hsieh@databricks.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cliangchi.hsieh@databricks.com\u003e\n"
    },
    {
      "commit": "5db2a73a0464ca042b5d52c4bc39c67a15042ac8",
      "tree": "5ef0638c5e4824dd133ee8c006aaff199d916a4d",
      "parents": [
        "131c830324acdcf2979d1d09d2992c78cea1ec1b"
      ],
      "author": {
        "name": "dependabot[bot]",
        "email": "49699333+dependabot[bot]@users.noreply.github.com",
        "time": "Sat Oct 10 16:00:56 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "liangchi.hsieh@databricks.com",
        "time": "Sat Oct 10 16:00:56 2026 -0700"
      },
      "message": "Bump rustls from 0.23.43 to 0.23.45\n\nBumps [rustls](https://github.com/rustls/rustls) from 0.23.43 to 0.23.45.\n\u003cdetails\u003e\n\u003csummary\u003eCommits\u003c/summary\u003e\n\u003cul\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/rustls/rustls/commit/2976d90fd1c2db6b518700dd101b714069cfcb17\"\u003e\u003ccode\u003e2976d90\u003c/code\u003e\u003c/a\u003e Prepare 0.23.45\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/rustls/rustls/commit/f9d4ee92b20da535ef5982ffcaf6f62385f79641\"\u003e\u003ccode\u003ef9d4ee9\u003c/code\u003e\u003c/a\u003e Handshake \u0026quot;alignment\u0026quot; check covers previously-received messages\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/rustls/rustls/commit/c4e92e8c9ee29ca3f552d4938efa73680c179179\"\u003e\u003ccode\u003ec4e92e8\u003c/code\u003e\u003c/a\u003e Keep data needed for HRR processing together\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/rustls/rustls/commit/e553a7a70f99cad4722f06b32cc37857008a7d3f\"\u003e\u003ccode\u003ee553a7a\u003c/code\u003e\u003c/a\u003e server: TLS1.2 is not available after a HRR\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/rustls/rustls/commit/5dff9da73798253d8d35909dd8548db522cd973c\"\u003e\u003ccode\u003e5dff9da\u003c/code\u003e\u003c/a\u003e Test whether server negotiates TLS1.2 after HRR\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/rustls/rustls/commit/185a063b5fe1e9698c589d8be6a83d67224b95a1\"\u003e\u003ccode\u003e185a063\u003c/code\u003e\u003c/a\u003e reject a second ClientHello that changes the cipher suite\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/rustls/rustls/commit/9fafe6dd0c15ea6e47399e94b2c6b647861f874b\"\u003e\u003ccode\u003e9fafe6d\u003c/code\u003e\u003c/a\u003e reject a second ClientHello that drops pre_shared_key\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/rustls/rustls/commit/1b42c5f78e8e826d6afa72ce0ead3e3a4e624d79\"\u003e\u003ccode\u003e1b42c5f\u003c/code\u003e\u003c/a\u003e providers: zeroize private key DER\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/rustls/rustls/commit/64ad386785c718c74262b79e6893813db381adfe\"\u003e\u003ccode\u003e64ad386\u003c/code\u003e\u003c/a\u003e Bump version to 0.23.44\u003c/li\u003e\n\u003cli\u003e\u003ca href\u003d\"https://github.com/rustls/rustls/commit/1efbf662ed79d1392c542c0822805a818d232ac0\"\u003e\u003ccode\u003e1efbf66\u003c/code\u003e\u003c/a\u003e bogo: remove PostQuantum setup\u003c/li\u003e\n\u003cli\u003eAdditional commits viewable in \u003ca href\u003d\"https://github.com/rustls/rustls/compare/v/0.23.43...v/0.23.45\"\u003ecompare view\u003c/a\u003e\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/details\u003e\n\u003cbr /\u003e\n\n[![Dependabot compatibility score](https://dependabot-badges.githubapp.com/badges/compatibility_score?dependency-name\u003drustls\u0026package-manager\u003dcargo\u0026previous-version\u003d0.23.43\u0026new-version\u003d0.23.45)](https://docs.github.com/en/github/managing-security-vulnerabilities/about-dependabot-security-updates#about-compatibility-scores)\n\nDependabot will resolve any conflicts with this PR as long as you don\u0027t alter it yourself. You can also trigger a rebase manually by commenting `dependabot rebase`.\n\n[//]: # (dependabot-automerge-start)\n[//]: # (dependabot-automerge-end)\n\n---\n\n\u003cdetails\u003e\n\u003csummary\u003eDependabot commands and options\u003c/summary\u003e\n\u003cbr /\u003e\n\nYou can trigger Dependabot actions by commenting on this PR:\n- `dependabot rebase` will rebase this PR\n- `dependabot recreate` will recreate this PR, overwriting any edits that have been made to it\n- `dependabot show \u003cdependency name\u003e ignore conditions` will show all of the ignore conditions of the specified dependency\n- `dependabot ignore this major version` will close this PR and stop Dependabot creating any more for this major version (unless you reopen the PR or upgrade to it yourself)\n- `dependabot ignore this minor version` will close this PR and stop Dependabot creating any more for this minor version (unless you reopen the PR or upgrade to it yourself)\n- `dependabot ignore this dependency` will close this PR and stop Dependabot creating any more for this dependency (unless you reopen the PR or upgrade to it yourself)\nYou can disable automated security fix PRs for this repo from the [Security Alerts page](https://github.com/apache/spark-connect-gateway/network/alerts).\n\n\u003c/details\u003e\n\nCloses #47 from dependabot[bot]/dependabot/cargo/rustls-0.23.45.\n\nAuthored-by: dependabot[bot] \u003c49699333+dependabot[bot]@users.noreply.github.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cliangchi.hsieh@databricks.com\u003e\n"
    },
    {
      "commit": "131c830324acdcf2979d1d09d2992c78cea1ec1b",
      "tree": "8ac7f8feab12a5470914a6973b2cd286051eb126",
      "parents": [
        "f022ad9bfa05cfd5f8b811965339b9c81a244d6b"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Fri Oct 09 00:49:07 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "liangchi.hsieh@databricks.com",
        "time": "Fri Oct 09 00:49:07 2026 -0700"
      },
      "message": "[SPARK-59973] Upgrade the OpenTelemetry crates to 0.33 and tracing-opentelemetry to 0.34\n\n### What changes were proposed in this pull request?\n\nMove the six `opentelemetry*` crates from 0.32 to 0.33, and\n`tracing-opentelemetry` from 0.33 to 0.34:\n\n```\nopentelemetry                       0.32 -\u003e 0.33\nopentelemetry_sdk                   0.32 -\u003e 0.33\nopentelemetry-otlp                  0.32 -\u003e 0.33\nopentelemetry-semantic-conventions  0.32 -\u003e 0.33\nopentelemetry-stdout                0.32 -\u003e 0.33\nopentelemetry-proto                 0.32 -\u003e 0.33\ntracing-opentelemetry               0.33 -\u003e 0.34\n```\n\n`Cargo.toml` and `Cargo.lock` only — no source changes, as the APIs this repo uses\nare unchanged between 0.32 and 0.33.\n\nThis supersedes #41, which is closed.\n\n### Why are the changes needed?\n\nRoutine dependency maintenance, but the shape is worth explaining since the title\nnames two things and the diff moves eight crates.\n\nThese crates share types across their public APIs, so raising one on its own\nleaves two majors of `opentelemetry` in the dependency graph and the traits stop\nlining up. `tracing-opentelemetry`\u0027s version is offset by one minor — 0.34 is the\nrelease that tracks `opentelemetry` 0.33 — so the whole family moves together.\nThis is the same shape as SPARK-59548, which took the group from 0.31 to 0.32.\n\nWorth noting that dependabot raised this *correctly*, as a single grouped PR under\nthe grouping added in SPARK-59550, rather than as the seven individually\nunmergeable single-crate PRs it would have produced before. The grouping also\nbehaved as intended at the edges: `tracing-opentelemetry` was included, while\n`tracing` and `tracing-subscriber` were left out — they carry no version coupling\nto OpenTelemetry, and sweeping them in would have held their updates behind this\nfamily for no reason.\n\n### Does this PR introduce _any_ user-facing change?\n\nNo. Dependency versions only; no change to configuration, metrics, span names, or\nany emitted telemetry.\n\n### How was this patch tested?\n\n**The check that shows the version skew is actually gone** is the lockfile\nresolving a single `opentelemetry`. Building is not sufficient proof — a skewed\ngraph still compiles until the mismatched trait is named, which is exactly how\n#8 looked before it failed:\n\n```\n$ grep -c \u0027^name \u003d \"opentelemetry\"$\u0027 Cargo.lock\n1\n```\n\n`opentelemetry` and `opentelemetry_sdk` resolve to 0.33.0, `tracing-opentelemetry`\nto 0.34.0, and `cargo metadata --locked` confirms the lock agrees with the\nmanifest.\n\n**The tracing integration test matters more here than the unit tests**, and\npasses:\n\n```\ntest tracing_emits_spans_and_propagates_traceparent ... ok\n```\n\nFeature unification in this dependency group has previously caused application\nspans to be dropped silently — enabling `opentelemetry`\u0027s\n`spec_unstable_logs_enabled` through an unrelated path made the linked\n`tracing-opentelemetry` discard them — and neither a build nor the unit tests\nwould catch that. This test asserts spans are emitted and `traceparent` is\npropagated, so it is the one that would notice.\n\n**The nine Redis integration tests** were run against a real Redis as well and are\nunaffected, as expected.\n\n**Full workspace:**\n\n| check | result |\n|---|---|\n| `cargo test --workspace` | 204 passed, 0 failed, 10 ignored — same as the baseline on `main` |\n| `cargo clippy --workspace --all-targets -- -D warnings` | clean |\n| `cargo fmt --all --check` | clean |\n\nI also confirmed this PR leaves the proto work from #40 and #42 untouched\n(`dev/sync-protos.sh` reports `identical: 11`), since a sibling dependabot PR had\nquietly reverted it and that is worth checking rather than assuming.\n\n### Was this patch authored or co-authored using generative AI tooling?\n\nYes, co-authored with Claude Code.\n\nCloses #43 from viirya/dep-04-opentelemetry-0.33.\n\nAuthored-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cliangchi.hsieh@databricks.com\u003e\n"
    },
    {
      "commit": "f022ad9bfa05cfd5f8b811965339b9c81a244d6b",
      "tree": "90948a8f227b86f8601f62ca8c66a1fa33192f5d",
      "parents": [
        "f154d512a04224a205808c8aef31087fecb3ff68"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Thu Oct 08 23:56:15 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "liangchi.hsieh@databricks.com",
        "time": "Thu Oct 08 23:56:15 2026 -0700"
      },
      "message": "[SPARK-59989] Pin the Rust toolchain and silence double_must_use on async_trait traits\n\n### What changes were proposed in this pull request?\n\nFix `main`\u0027s failing CI, in two parts:\n\n- **`rust-toolchain.toml`** — pin `channel` from `\"stable\"` to `\"1.99.0\"`, with the\n  rationale and the steps for moving the pin recorded in the file.\n- **`crates/routing/src/lib.rs`, `crates/auth/src/lib.rs`** — allow\n  `clippy::double_must_use` on the two `#[async_trait]` traits it fires on, with a\n  comment explaining why.\n\n### Why are the changes needed?\n\n`main`\u0027s CI has been red since 2026-10-04 on a commit that changed nothing\nrelevant:\n\n```\n2026-10-04  CI:      failure     \u003c- main\n2026-10-04  E2E:     success\n```\n\nBecause `rust-toolchain.toml` floated on `stable`, CI picked up the newly released\nRust 1.99.0, whose clippy added the `double_must_use` lint.\n\nThe lint fires on every `#[async_trait]` trait here. `async_trait` expands each\nmethod into one returning a `#[must_use]` `Pin\u003cBox\u003cdyn Future\u003e\u003e`, and the lint then\nobjects that the inner type — a `Result`, `Option` or `String` — is already\nmust-use. **The attribute belongs to the macro expansion, not to our code**, so\nthere is nothing in these traits to restructure; allowing it is the right answer\nrather than a workaround.\n\n**On pinning.** A floating channel means CI can turn red on a commit that changed\nnothing, at an unpredictable time, whenever upstream Rust ships a new lint or\nchanges behaviour — which is exactly what happened here. Pinning makes the build\nreproducible between contributors and CI, and turns a toolchain upgrade into a\nreviewable change of its own rather than an event arriving on its own schedule.\n\nThe cost is real and worth stating: Rust upgrades become manual. That seems clearly\nbetter than the alternative, but it is a trade-off rather than a free win.\n\nMSRV is unaffected — `rust-version` stays at `1.82`, which is the minimum this\ncompiles with, a separate thing from the version used to develop and gate.\n\n### Does this PR introduce _any_ user-facing change?\n\nNo. Build configuration and two lint attributes; no behaviour change.\n\n### How was this patch tested?\n\n**Everything was verified on 1.99.0**, because that is what CI resolves to — and\n1.97.1, which a contributor may well have locally, does not reproduce the failure\nat all. That gap is the whole problem, so checking on the old version would have\nproved nothing.\n\n- The failure **reproduces before the change**: 6 errors locally on 1.99.0,\n  matching CI.\n- `cargo +1.99.0 clippy --workspace --all-targets -- -D warnings` — clean after.\n- `cargo +1.99.0 fmt --all --check` — clean.\n- `cargo +1.99.0 test --workspace` — 204 passed, 0 failed, 10 ignored.\n- A **bare `cargo`** now resolves to 1.99.0 through `rust-toolchain.toml` and\n  passes, which confirms the pin takes effect for contributors rather than only\n  when the version is named explicitly.\n- **The pin takes effect in CI too**, which is worth showing explicitly because\n  `.github/actions/setup-builder` still runs `rustup toolchain install stable` and\n  the log\u0027s `Installing stable` line reads as though the pin were bypassed. It is\n  not — rustup resolves the file ahead of the default, and says so:\n\n  ```\n  info: note that the toolchain \u00271.99.0-x86_64-unknown-linux-gnu\u0027 is currently in use\n        (overridden by \u0027.../rust-toolchain.toml\u0027)\n  ```\n\n  So `setup-builder`\u0027s `rust-version: stable` input is now inert for version\n  selection; it still installs the components. Simplifying that input is worth\n  doing, but as a separate change rather than widening this one.\n\n**One detail about scope worth raising for review.** The CI log named only\n`crates/routing/src/lib.rs`, because cargo stopped once that crate failed.\nReproducing locally showed `crates/auth/src/lib.rs:60` is affected too:\n\n```\ncrates/auth/src/lib.rs:60\ncrates/routing/src/lib.rs:199\n```\n\nSo fixing what the log reported would have left CI red on the next run with a\ndifferent crate named. Both are fixed here — which is the argument for reproducing\na toolchain failure locally rather than reading it off the log.\n\nThis unblocks #43 (SPARK-59973), which is red for this reason and not for anything\nin its own diff.\n\n### Was this patch authored or co-authored using generative AI tooling?\n\nYes, co-authored with Claude Code.\n\nCloses #44 from viirya/ci-11-pin-toolchain.\n\nAuthored-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cliangchi.hsieh@databricks.com\u003e\n"
    },
    {
      "commit": "f154d512a04224a205808c8aef31087fecb3ff68",
      "tree": "62865c676254a4b306ab51e14c47a0cbcdbaa967",
      "parents": [
        "6657bf24c6e4b41aaed824ce479e14802b2684bf"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Sat Oct 03 22:38:05 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "liangchi.hsieh@databricks.com",
        "time": "Sat Oct 03 22:38:05 2026 -0700"
      },
      "message": "[SPARK-59908] Sync the vendored Spark Connect protos to v4.2.0\n\n### What changes were proposed in this pull request?\n\nSync the vendored Spark Connect protocol definitions to upstream `v4.2.0`:\n\n```bash\ndev/sync-protos.sh --sync --ref v4.2.0\n```\n\n`+31` lines in `relations.proto` and `+42` in `pipelines.proto`, **no deletions** —\nthe drift was one-directional, so closing it is purely additive:\n\n| File | Added |\n|---|---|\n| `relations.proto` | the `NearestByJoin` message, and its entry in the `Relation.rel_type` oneof (tag 47) |\n| `pipelines.proto` | the `AutoCdcFlowDetails` message, and the `import \"spark/connect/expressions.proto\"` it needs |\n\n`proto/PROVENANCE.md` and the `UPSTREAM_REF` comment in `dev/sync-protos.sh` are\nupdated to describe the new state.\n\n### Why are the changes needed?\n\n#40 (SPARK-59856) recorded that the vendored protos were behind upstream and added\n`dev/sync-protos.sh`, but deliberately left the drift in place — re-syncing moves\nthe protocol the gateway speaks, so it wanted its own review rather than riding\nalong with a documentation change. This is that review.\n\n**Why `v4.2.0`.** It is the newest *released* Spark tag. `v4.3.0` exists only as\nrelease candidates, and syncing to an rc would pin the gateway to a protocol that\ncan still change before release. For reference, the tree is 4 files away from\n`v4.3.0-rc1`, so there is another sync to make once 4.3.0 ships — this is a step,\nnot an end state.\n\n### Does this PR introduce _any_ user-facing change?\n\nNo. The added messages extend the protocol the gateway can describe; they do not\nchange how it routes, authenticates, or forwards anything. No configuration,\nmetric, or audit-event change.\n\n### How was this patch tested?\n\nA compiling binding is weak evidence for a protocol change, so the verification\ngoes past the build.\n\n**Byte-identical to upstream.** Checked file by file rather than trusting the\nscript that wrote them — all 11 of 11 match `v4.2.0` exactly, and\n`dev/sync-protos.sh` reports `identical: 11   differing: 0`.\n\n**The new messages really are generated**, not just compiled past.\n`cargo build -p scg-genproto` regenerates the bindings, and both appear in the\ngenerated Rust as real structs:\n\n```rust\npub struct NearestByJoin {\n    #[prost(message, optional, boxed, tag \u003d \"1\")]\n    pub left: ::core::option::Option\u003c::prost::alloc::boxed::Box\u003cRelation\u003e\u003e,\n```\n\n**The new `rel_type` variant cannot break an exhaustive match.** A new oneof\nvariant is the one way this change could have broken compilation somewhere\nunexpected, so I checked rather than assumed: grepping the workspace outside\n`crates/genproto` for `rel_type` / `RelType` returns nothing. The single place that\ndoes inspect message contents, `crates/proxy/src/config_filter.rs`, matches on\n`ConfigRequest.operation.op_type` — untouched here.\n\n**End to end against a real Spark backend**, on kind, deployed through the Helm\nchart: `AnalyzePlan` returns the correct schema for a `range` plan, and\n`ExecutePlan` streams 15 response messages back, with the gateway\u0027s audit log\nconfirming both RPCs passed through it:\n\n```\n   1 \"rpc\":\"AnalyzePlan\"\n   1 \"rpc\":\"ExecutePlan\"\n```\n\nThe client encoded its requests **from the newly synced `.proto` files**, which is\nwhat makes this a test of the sync rather than of the gateway alone. It also\nconfirms the `v4.2.0` protocol stays compatible with a Spark 4.0 backend, which is\nwhat the e2e jobs run.\n\n**Nothing regressed:** `cargo test --workspace` is 204 passed, 0 failed, 10\nignored — matching `main` — with `clippy --workspace --all-targets -D warnings` and\n`cargo fmt --check` clean.\n\n### Was this patch authored or co-authored using generative AI tooling?\n\nYes, co-authored with Claude Code.\n\nCloses #42 from viirya/proto-02-sync-4.2.0.\n\nAuthored-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cliangchi.hsieh@databricks.com\u003e\n"
    },
    {
      "commit": "6657bf24c6e4b41aaed824ce479e14802b2684bf",
      "tree": "c78b82eab70f6a95bf97747ff25a79ee2ea3ce6a",
      "parents": [
        "0148e8d86515e6dbefa5721018e0da3ded54cf82"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Tue Sep 29 11:47:41 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "liangchi.hsieh@databricks.com",
        "time": "Tue Sep 29 11:47:41 2026 -0700"
      },
      "message": "[SPARK-59856] Record the provenance of the vendored Spark Connect protos and add a sync script\n\n### What changes were proposed in this pull request?\n\nRecord where the vendored Spark Connect protocol definitions came from, and add a\nscript to check or re-sync them.\n\n- **`proto/PROVENANCE.md`** — upstream project and path, the baseline revision,\n  the licensing position, the two known deltas from upstream, and the re-sync\n  procedure.\n- **`dev/sync-protos.sh`** — compares `proto/spark/connect/` against\n  `apache/spark` and reports drift; overwrites only with `--sync`.\n- **`README.md`** — points at both, and separates *regenerating the bindings*\n  (already documented) from *updating the `.proto` files themselves* (was not).\n\n`proto/spark/connect/` itself is untouched by this PR.\n\n### Why are the changes needed?\n\nThe eleven `.proto` files under `proto/spark/connect/` are copied verbatim from\n`apache/spark` and are the gateway\u0027s source of protocol truth — `crates/genproto`\ngenerates the Rust bindings from them on every build. Nothing in the repo recorded\nwhere they came from. README called them a \"vendored read-only mirror of upstream\"\nbut named no revision, no upstream path, and no way to re-sync.\n\nThree concrete costs:\n\n1. **A release manager has no answer** to \"where did these come from, and under\n   what license\" — a question that gets asked during an ASF release vote.\n2. **Nobody can tell how far behind upstream the copy is**, or which revision to\n   sync from, when Spark Connect\u0027s protocol moves.\n3. **\"Have these been edited locally since they were copied?\"** could only be\n   answered by archaeology.\n\nEstablishing the answer meant comparing all eleven files against upstream tags, so\nthis records the result rather than leaving it to be redone.\n\n**On the revision.** The copy is from the Spark 4.2 development line, not a\nrelease: nine of the eleven files are byte-identical to `v4.2.0`, and the files\nthemselves reference fields \"deprecated since Spark 4.2+\", but no release tag\nmatches — `v4.2.0-rc1` through `rc5` were checked too. The doc records `v4.2.0` as\nthe baseline to diff against and states explicitly that this is *not* a claim the\nfiles were taken from that tag. A version number that looks precise but is wrong\nwould be worse than recording the uncertainty.\n\n**The comparison surfaced drift that was not previously known.** The vendored\nprotos are *behind* upstream:\n\n| File | Missing relative to `v4.2.0` |\n|---|---|\n| `relations.proto` | the `NearestByJoin` message and its `Relation.rel_type` entry |\n| `pipelines.proto` | the `AutoCdcFlowDetails` message and its `expressions.proto` import |\n\nBoth are pure upstream additions — nothing was edited locally. This PR records\nthat and deliberately **does not** fix it: re-syncing moves the wire protocol and\ndeserves review on its own, not folded into a documentation change.\n\nThe drift does not affect the gateway today. `crates/proxy` forwards each request\nonward whole and reads only session and identity metadata — nothing there inspects\na plan\u0027s `rel_type` or walks the relation tree — so a request carrying one of the\nnewer messages still routes and forwards correctly. I checked that in the code\nrather than assuming it, since it is the reason the drift is not urgent.\n\n### Does this PR introduce _any_ user-facing change?\n\nNo. Documentation and a developer script; no protocol, code, or behaviour change.\n\n### How was this patch tested?\n\n**The script was exercised, not just written.** Check mode agrees with the manual\ncomparison that established the provenance in the first place:\n\n```\n$ dev/sync-protos.sh\nvendored: proto/spark/connect\nupstream: apache/spark v4.2.0 sql/connect/common/src/main/protobuf/spark/connect\n\n  DIFFERS  pipelines.proto  (+42/-0 lines vs upstream)\n  DIFFERS  relations.proto  (+31/-0 lines vs upstream)\n\nidentical: 9   differing: 2   only-upstream: 0   only-local: 0\n```\n\nThe `+N/-0` shape is itself the evidence that the drift is one-directional — we\nare behind upstream, not forked from it.\n\n`--ref v4.1.3` reports 6 identical / 5 differing, also matching the manual\ncomparison, which shows the `--ref` path works rather than the default being\nhardcoded somewhere.\n\n**`--sync` was tested on a throwaway copy of the tree** rather than in place,\nsince it rewrites protocol files: it took that copy from 9/2 to 11/0, and the\nrepo\u0027s own `proto/` was verified unchanged afterwards.\n\n**Exit codes**: 1 for an unknown ref (with a message naming the ref), 2 for an\nunknown argument, 1 for `--ref` with no value, 0 for a clean check. Drift itself\ndoes not fail the check — the point of the default mode is to be runnable any time\nto see where things stand.\n\n**Nothing regressed**: `proto/spark/` is unchanged, `cargo build -p scg-genproto`\nstill compiles the bindings from those files, and the workspace is 204 passed, 0\nfailed, 10 ignored — the same as `main`.\n\nOne implementation note for review: the script fetches the upstream *directory\nlisting* rather than iterating over the local files, so a `.proto` added upstream\nis reported as drift instead of being invisible. The reverse case — vendored here,\nabsent upstream — is reported too.\n\n### Was this patch authored or co-authored using generative AI tooling?\n\nYes, co-authored with Claude Code.\n\nCloses #40 from viirya/proto-01-provenance.\n\nAuthored-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cliangchi.hsieh@databricks.com\u003e\n"
    },
    {
      "commit": "0148e8d86515e6dbefa5721018e0da3ded54cf82",
      "tree": "a2fecb098b476ef2faed6da19dda0dc73d3c4d61",
      "parents": [
        "cbe7252c98953e02a7acd3db619b704cac6fd4d6"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Mon Sep 21 23:39:20 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "liangchi.hsieh@databricks.com",
        "time": "Mon Sep 21 23:39:20 2026 -0700"
      },
      "message": "[SPARK-59657] Update the redis and base64 dependencies\n\n### What changes were proposed in this pull request?\n\nTwo small dependency updates:\n\n```\nredis   1.6.0 -\u003e 1.7.0  (Cargo.lock only)\nbase64  0.22  -\u003e 0.23   (crates/auth dev-dependency)\n```\n\nThe lock also picks up `combine` 4.6.7 -\u003e 4.6.8, which is redis\u0027s parser\ndependency coming along with it.\n\nThis supersedes #34 and #33, both closed.\n\n### Why are the changes needed?\n\nRoutine dependency maintenance. The reason it is a hand-written PR rather than\nmerging dependabot\u0027s two is worth stating, because it is not obvious from their\ndiffs.\n\nBoth branches were cut before SPARK-59552 landed, so each one *also* reverts\n`jsonwebtoken` from\n\n```toml\n{ version \u003d \"10\", default-features \u003d false, features \u003d [\"use_pem\", \"rust_crypto\"] }\n```\n\nback to `\"9\"`, dropping around 270 lines from `Cargo.lock` with it. That would\nsilently undo the crypto provider selection — and because `jsonwebtoken \u003d \"9\"`\ncompiles cleanly and passes every test, **CI would not catch it**. Five of the\ncurrent dependabot PRs share this property, so it is worth knowing about when\ntriaging the rest of the batch. Applying the two updates directly avoids waiting\nfor dependabot to rebase.\n\nOne deliberate difference from #34: it also narrowed the manifest requirement from\n`redis \u003d \"1\"` to `\"1.7\"`. That narrowing is unnecessary — the existing `\"1\"`\nrequirement already admits 1.7.0 — so the root `Cargo.toml` is untouched here and\nonly `Cargo.lock` moves.\n\n### Does this PR introduce _any_ user-facing change?\n\nNo. Dependency versions only.\n\n### How was this patch tested?\n\nThe two updates carry different risk, so they were checked differently rather\nthan uniformly.\n\n**`redis` is a production dependency and backs the affinity store**, so it was\nexercised against a real Redis instead of being taken as verified because it\ncompiles. All nine Redis integration tests pass:\n\n```\nscg-store-redis (5)   session_bind_resolve_forget_roundtrip\n                      op_reverse_index_roundtrip\n                      session_ttl_expires\n                      bind_session_if_absent_is_atomic\n                      two_stores_against_same_redis_share_state\nscg-ratelimit (4)     burst_admits_then_rejects\n                      refill_admits_after_wait\n                      tenants_are_independent\n                      two_replicas_share_the_bucket\n```\n\nThese are the `#[ignore]`d tests that CI\u0027s `integration` job runs; they cover the\natomicity and TTL behaviour the affinity store depends on, which a build cannot\nspeak to.\n\n**`base64` is a dev-dependency of `crates/auth` only** — its 30 tests pass.\n\n**Full workspace:**\n\n| check | result |\n|---|---|\n| `cargo test --workspace` | 204 passed, 0 failed, 10 ignored — same as the baseline on `main` |\n| `cargo clippy --workspace --all-targets -- -D warnings` | clean |\n| `cargo fmt --all --check` | clean |\n\nI also confirmed the provider selection survived, since that is the specific thing\nthe dependabot versions would have broken: `Cargo.lock` still resolves\n`jsonwebtoken` 10.4.0, and the only `scg-auth` dependency line that changed is\n`base64 0.22.1` -\u003e `0.23.1`.\n\n### Was this patch authored or co-authored using generative AI tooling?\n\nYes, co-authored with Claude Code.\n\nCloses #35 from viirya/dep-03-redis-base64.\n\nAuthored-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cliangchi.hsieh@databricks.com\u003e\n"
    },
    {
      "commit": "cbe7252c98953e02a7acd3db619b704cac6fd4d6",
      "tree": "6d8b54c0b42aa9791ab400248e8a5b5c7791431c",
      "parents": [
        "4015aacbfcd02ec7568746da3dc3e69fb798f13a"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Thu Sep 17 17:30:13 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "liangchi.hsieh@databricks.com",
        "time": "Thu Sep 17 17:30:13 2026 -0700"
      },
      "message": "[SPARK-59552] Upgrade jsonwebtoken to 10 and select the rust_crypto provider\n\n### What changes were proposed in this pull request?\n\nMove `jsonwebtoken` from 9 to 10 and select the `rust_crypto` backend explicitly:\n\n```toml\njsonwebtoken \u003d { version \u003d \"10\", default-features \u003d false,\n                 features \u003d [\"use_pem\", \"rust_crypto\"] }\n```\n\n`Cargo.toml` and `Cargo.lock` only; no source changes.\n\nThis supersedes #6, which is closed.\n\n### Why are the changes needed?\n\nVersion 10 makes the crypto backend pluggable and refuses to guess one. What\nmakes this more than a version bump is that **it fails at runtime, not at compile\ntime** — `jsonwebtoken \u003d \"10\"` on its own builds cleanly, so the gateway would\nstart normally and panic on the first JWT it verified:\n\n```\nCould not automatically determine the process-level CryptoProvider from\njsonwebtoken crate features. Call CryptoProvider::install_default() before this\npoint to select a provider manually, or make sure exactly one of the\n\u0027rust_crypto\u0027 and \u0027aws_lc_rs\u0027 features is enabled.\n```\n\nThe 14 test failures dependabot reported on #6 were a symptom of this rather than\nthe problem itself.\n\n**Why `rust_crypto`.** It is the pure-Rust provider, consistent with this repo\u0027s\nexisting choices (`reqwest` with `rustls-tls`, `opentelemetry-otlp` with\n`tls-ring`), and it adds no C toolchain dependency — which matters for the\ndistroless runtime image. The alternative, `aws_lc_rs`, is the FIPS-oriented route\nbut brings a C dependency.\n\n**Why `use_pem` is listed explicitly.** It is a *default* feature, so\n`default-features \u003d false` with only a provider fails to compile:\n`EncodingKey::from_rsa_pem`, used by both the JWT and OIDC verifiers, lives behind\nit. This is an easy trap for anyone revisiting these features later.\n\n### Does this PR introduce _any_ user-facing change?\n\nNo. The set of accepted JWT algorithms, the configuration surface, and the\nrejection behaviour are all unchanged — see the algorithm round-trip check below.\n\n### How was this patch tested?\n\n**Every algorithm the gateway\u0027s config accepts was checked to sign *and* verify\nunder this provider**, not merely to compile. `rust_crypto` listing `p256` and\n`p384` among its features does not by itself prove `ES384` works, and a provider\nmissing one algorithm would fail only for the users who had configured it —\ninvisible to a build and to the unit tests. Real keys generated with `openssl`,\nfull round-trip per algorithm:\n\n```\nHS256 OK  HS384 OK  HS512 OK\nRS256 OK  RS384 OK  RS512 OK\nPS256 OK  PS384 OK  PS512 OK\nES256 OK  ES384 OK\n```\n\nThat is all 11 algorithms accepted by `crates/config`.\n\n**No new C dependency**, which is the main reason for choosing this provider, so\nit is worth showing rather than asserting. `jsonwebtoken`\u0027s dependencies are pure\nRust (`rsa`, `p256`, `p384`, `sha2`, `hmac`, `ed25519-dalek`). `ring` does appear\nin the tree, but `cargo tree -i ring` shows it arriving via `rustls` /\n`rustls-webpki`, and it is already present on `main` — unchanged by this patch.\n\n**The standard quartet:**\n\n| check | result |\n|---|---|\n| `cargo test --workspace` | 204 passed, 0 failed, 10 ignored — same as the baseline on `main` |\n| `cargo clippy --workspace --all-targets -- -D warnings` | clean |\n| `cargo fmt --all --check` | clean |\n| `cargo build --workspace` | clean |\n\nThe 17 `jwt::` and `oidc::` tests all pass, which includes the 14 that #6 reported\nas failing.\n\nOne note for review: the `Cargo.lock` diff is large (269 lines) compared with a\ntypical bump. That is expected — selecting a provider replaces the crypto\nimplementation stack rather than moving a single crate.\n\n### Was this patch authored or co-authored using generative AI tooling?\n\nYes, co-authored with Claude Code.\n\nCloses #29 from viirya/dep-02-jsonwebtoken-10.\n\nAuthored-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cliangchi.hsieh@databricks.com\u003e\n"
    },
    {
      "commit": "4015aacbfcd02ec7568746da3dc3e69fb798f13a",
      "tree": "45033205ad0846014ee3a57b6c5fd95e16ada4a9",
      "parents": [
        "ef2d2ba63da1c2b0988425199d19c6f1affec134"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Thu Sep 17 17:29:26 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "liangchi.hsieh@databricks.com",
        "time": "Thu Sep 17 17:29:26 2026 -0700"
      },
      "message": "[SPARK-59550] Group coupled crate families in dependabot updates\n\n### What changes were proposed in this pull request?\n\nAdd `.github/dependabot.yml` so dependabot raises coupled crate families in one\nPR instead of one crate at a time, and skips patch-level updates.\n\nThree groups:\n\n| group | crates | why they are coupled |\n|---|---|---|\n| `opentelemetry` | 7 | the `opentelemetry*` family plus `tracing-opentelemetry`, whose version is offset by one minor (0.33 tracks opentelemetry 0.32) |\n| `tonic-prost` | 6 | `tonic-prost` depends on `tonic` and `prost`, `tonic-health` on `tonic` and `tonic-prost`, and `opentelemetry-proto` on `tonic` and `prost` as well |\n| `kube` | 2 | `kube` depends on `k8s-openapi`, whose version selects the Kubernetes API version its generated types target |\n\nPatch-level updates are ignored for every dependency.\n\n### Why are the changes needed?\n\nCrates in these families share types across their public APIs, so bumping one\nmember on its own leaves two major versions of the family in the dependency graph\nand the traits stop lining up. A PR in that state cannot be made green by\nrerunning it — it is unmergeable by construction.\n\nThat is not hypothetical. Dependabot opened #8 to bump `opentelemetry_sdk` alone;\nit was closed as unmergeable, and the upgrade that actually worked\n(SPARK-59548, #27) had to move all seven `opentelemetry*` crates **and**\n`tracing-opentelemetry` together. Without grouping, dependabot will keep opening\nthat same PR every time any one of these crates releases.\n\nPatch updates are ignored because dependencies here are pinned by minor and\n`Cargo.lock` is committed, so patch PRs are noise — a real patch pickup rides\nalong with the next minor bump or a deliberate `cargo update`. Happy to drop that\npart if you would rather see them.\n\n### Does this PR introduce _any_ user-facing change?\n\nNo. Dependabot configuration only; no dependency versions change here and no code\nis touched.\n\n### How was this patch tested?\n\nA config file is only useful if it takes effect, and a bad `dependabot.yml` is\nignored silently rather than reported, so I checked it three ways instead of\neyeballing it.\n\n**The patterns match the intended crates and nothing else.** Matched with\n`fnmatch` against the real names in `[workspace.dependencies]`:\n\n```\nopentelemetry: opentelemetry, opentelemetry-otlp, opentelemetry-proto,\n               opentelemetry-semantic-conventions, opentelemetry-stdout,\n               opentelemetry_sdk, tracing-opentelemetry          (7)\ntonic-prost:   prost, prost-types, tonic, tonic-health,\n               tonic-prost, tonic-prost-build                    (6)\nkube:          k8s-openapi, kube                                 (2)\n```\n\nNote `opentelemetry_sdk` needs the `opentelemetry_*` pattern separately from\n`opentelemetry-*` — the underscore would otherwise be missed.\n\n**Nothing is over-matched.** `tracing` and `tracing-subscriber` stay ungrouped, as\nthey should: they carry no version coupling to OpenTelemetry, and a `tracing-*`\npattern would have swept them in and needlessly held their updates behind the\nOpenTelemetry family.\n\n**The coupling itself was verified, not assumed** — each relationship above was\nread out of `Cargo.lock`\u0027s dependency lists rather than inferred from the crate\nnames. That is how the `kube` / `k8s-openapi` pair was found; it was not in the\noriginal scope.\n\nThe schema was also checked against dependabot v2\u0027s permitted keys for `updates`,\n`groups` and `ignore` entries: no unknown fields.\n\n### Was this patch authored or co-authored using generative AI tooling?\n\nYes, co-authored with Claude Code.\n\nCloses #28 from viirya/ci-10-dependabot-groups.\n\nAuthored-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cliangchi.hsieh@databricks.com\u003e\n"
    },
    {
      "commit": "ef2d2ba63da1c2b0988425199d19c6f1affec134",
      "tree": "728956ad8cd2b01da99b30ea7cb69c7b68cfe1d3",
      "parents": [
        "ab637ce20c1056836fb596232d0724586720b791"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Thu Sep 17 17:28:47 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "liangchi.hsieh@databricks.com",
        "time": "Thu Sep 17 17:28:47 2026 -0700"
      },
      "message": "[SPARK-59548] Upgrade the OpenTelemetry crates to 0.32 and tracing-opentelemetry to 0.33\n\n### What changes were proposed in this pull request?\n\nMove the seven `opentelemetry*` crates from 0.31 to 0.32, and\n`tracing-opentelemetry` from 0.32 to 0.33:\n\n```\nopentelemetry                       0.31 -\u003e 0.32\nopentelemetry_sdk                   0.31 -\u003e 0.32\nopentelemetry-otlp                  0.31 -\u003e 0.32\nopentelemetry-semantic-conventions  0.31 -\u003e 0.32\nopentelemetry-stdout                0.31 -\u003e 0.32\nopentelemetry-proto                 0.31 -\u003e 0.32\ntracing-opentelemetry               0.32 -\u003e 0.33\n```\n\n`Cargo.toml` and `Cargo.lock` only. No source changes were needed — the APIs this\nrepo uses are unchanged between 0.31 and 0.32.\n\nThis supersedes #8, which is closed.\n\n### Why are the changes needed?\n\nDependabot opened #8 to bump `opentelemetry_sdk` on its own. That cannot compile:\nthe other six `opentelemetry*` crates stay at 0.31, so both 0.31 and 0.32 of\n`opentelemetry` land in the dependency graph and the traits stop lining up.\nReproduced locally on `main`:\n\n```\nerror[E0308]: mismatched types\n   --\u003e crates/observability/src/tracing.rs:167:25\nerror[E0277]: the trait bound `opentelemetry_otlp::SpanExporter:\n              opentelemetry_sdk::trace::SpanExporter` is not satisfied\n   --\u003e crates/observability/src/tracing.rs:199:30\nerror[E0277]: the trait bound `TraceContextPropagator: TextMapPropagator`\n              is not satisfied\n   --\u003e crates/observability/src/tracing.rs:206:37\n```\n\nMoving the whole group to 0.32 is still not sufficient on its own, because\n`tracing-opentelemetry` 0.32 itself depends on `opentelemetry` 0.31 and pulls it\nback in:\n\n```\nerror[E0277]: the trait bound `SdkTracer: opentelemetry::trace::tracer::Tracer`\n              is not satisfied\n   --\u003e crates/observability/src/tracing.rs:209:65\n```\n\n0.33 is the `tracing-opentelemetry` release that tracks `opentelemetry` 0.32, so\nall eight crates move at once.\n\n### Does this PR introduce _any_ user-facing change?\n\nNo. Dependency versions only; no change to configuration, metrics, span names, or\nany emitted telemetry.\n\n### How was this patch tested?\n\nThe full CI quartet on this branch:\n\n| check | result |\n|---|---|\n| `cargo test --workspace` | 204 passed, 0 failed, 10 ignored — same as the baseline on `main` |\n| `cargo clippy --workspace --all-targets -- -D warnings` | clean |\n| `cargo fmt --all --check` | clean |\n| `cargo build --workspace` | clean |\n\nThe check that actually shows the version skew is gone is the lockfile resolving a\nsingle `opentelemetry`:\n\n```\n$ grep -c \u0027^name \u003d \"opentelemetry\"\u0027 Cargo.lock\n1\n```\n\nBeyond the quartet I ran the `#[ignore]`d integration tests, because the unit\ntests are the weakest signal for this particular dependency group:\n\n- `tracing_emits_spans_and_propagates_traceparent` passes. This one matters most:\n  feature unification in this group has previously caused application spans to be\n  dropped silently, and neither a build nor the unit tests would catch that.\n- The Redis integration tests (`scg-store-redis`, `scg-ratelimit`) pass —\n  5 and 4 tests respectively — unaffected, as expected.\n\n### Was this patch authored or co-authored using generative AI tooling?\n\nYes, co-authored with Claude Code.\n\nCloses #27 from viirya/dep-01-opentelemetry-0.32.\n\nAuthored-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cliangchi.hsieh@databricks.com\u003e\n"
    },
    {
      "commit": "ab637ce20c1056836fb596232d0724586720b791",
      "tree": "86ead0a63b6ba657de073096447ff13175fcfccc",
      "parents": [
        "c06c50955f78b59d4654ff7605f6f5342eb6f454"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Mon Sep 14 23:43:11 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "liangchi.hsieh@databricks.com",
        "time": "Mon Sep 14 23:43:11 2026 -0700"
      },
      "message": "[SPARK-59482] Fix the kubectl wait race in the e2e jobs and add a .dockerignore\n\n### What changes were proposed in this pull request?\n\nTwo independent CI fixes.\n\n**1. Replace `kubectl wait` with `kubectl rollout status` in all five e2e jobs.**\nEvery job waited for its pods with\n\n```\nkubectl wait --for\u003dcondition\u003dready pod -l \u003cselector\u003e\n```\n\nThat command fails immediately with `error: no matching resources found` when\nthe selector matches zero pods, and `--timeout` governs only how long it waits\nfor the *condition*, not for the *resource to appear*. A Deployment created\nmoments earlier, whose ReplicaSet has not yet produced a pod, loses the race.\n\nAll eight call sites now use `kubectl rollout status deployment/\u003cname\u003e`, which\nwaits on the controller instead of on a label match. Where one selector\npreviously covered two Deployments (`e2e-multitenant`\u0027s\n`app in (spark-connect-team-a,spark-connect-team-b)`), it becomes one call per\nDeployment in a loop.\n\n**2. Stop the log assertions failing on SIGPIPE.** Five steps asserted on\ngateway logs with `kubectl logs ... | grep -q PATTERN`. Under `set -o pipefail`,\n`grep -q` closes the pipe on its first match, `kubectl logs` takes SIGPIPE, and\nthe step reports 128+13 \u003d 141 even though the assertion held. Each now captures\nthe log into a variable and greps a herestring.\n\n**3. Add a `.dockerignore`.** The repo had none, so the `COPY . .` in the\nDockerfile sent the entire working tree to the daemon as build context.\n\nThe exclusions deliberately keep everything the build reads: the crates,\n`proto/` (genproto\u0027s `build.rs` reads it at build time and panics if it is\nempty), `Cargo.toml`, `Cargo.lock` and `rust-toolchain.toml`.\n\n### Why are the changes needed?\n\nThe `kubectl wait` race makes the e2e jobs randomly red. It turned\n`e2e-auth-jwt` red on PR #24, which had not touched that job at all — a rerun\npassed, confirming flakiness rather than a regression. Random failures cost\nreviewers time deciding whether a red check is their fault, and erode trust in\nthe signal.\n\n`rollout status` is not a weaker check: a Deployment rollout only completes once\nits pods pass readiness, and every Spark backend Deployment under\n`deploy/examples/` declares a `readinessProbe`. So the guarantee is the same,\nwithout the race. Two jobs already used `rollout status` for the gateway and\nRedis; this makes all eight sites consistent.\n\nThe SIGPIPE bug is the same class of problem: a step reporting failure for a\nreason unrelated to what it asserts. It surfaced on this PR\u0027s own first CI run —\n`e2e-smoke` went red with exit 141 immediately after printing the two metrics it\nhad just verified. The other four call sites passed only because their output fit\nin the pipe buffer or the match came last, so they were latent rather than\ncorrect. Capturing into a variable also means a failing assertion can print what\nwas actually in the log.\n\nWithout a `.dockerignore`, `docker build .` ships `target/` — tens of gigabytes\nonce anyone has built locally. CI never noticed because a fresh checkout has no\n`target/`, but every contributor building the image locally pays it, and at that\nsize the transfer looks like a hang rather than progress.\n\n### Does this PR introduce _any_ user-facing change?\n\nNo. CI configuration and build context only; no change to the gateway, the Helm\nchart, or any published artifact.\n\n### How was this patch tested?\n\n**The wait fix** is exercised by this PR\u0027s own CI run: all five e2e jobs take\nthe new code path on real kind clusters.\n\nI verified the semantics are equivalent rather than assuming it — every Spark\nbackend Deployment under `deploy/examples/` declares a `readinessProbe`, and\n`kubectl rollout status` only reports success once the rollout completes, which\nrequires those probes to pass.\n\n**The SIGPIPE fix** was reproduced and verified locally, since the failure is a\nshell-level behaviour rather than anything cluster-specific:\n\n```\n$ bash -c \u0027set -euo pipefail; seq 1 200000 | grep -q \"^5$\"\u0027   # the old shape\nexit 141\n$ bash -c \u0027set -euo pipefail; big\u003d$(seq 1 200000); grep -q \"^5$\" \u003c\u003c\u003c\"$big\"\u0027\nexit 0\n```\n\n**The `.dockerignore`** was verified locally. Build context, measured before and\nafter:\n\n| | build context |\n|---|---|\n| before | 15 GB+ |\n| after | 999 kB |\n\nSize alone would not prove the exclusions are safe, so I listed the resulting\ncontext with a throwaway probe image: all 15 crates, `proto/spark/connect/*.proto`,\n`Cargo.lock`, `rust-toolchain.toml`, `LICENSE` and `NOTICE` are present, while\n`target/`, `.git/`, `docs/` and `deploy/` are gone. A full image build then\nsucceeded from that context and the resulting binary runs:\n\n```\n$ docker run --rm --entrypoint /gateway scg:verify --help\nOpen-source Spark Connect Gateway\n\nUsage: gateway [OPTIONS]\n```\n\nNeither `.github/.licenserc.yaml` nor `dev/.rat-excludes` exempts\n`.dockerignore`, so it carries a full ASF header in the same style as\n`rust-toolchain.toml`; the `Check license header` job passes.\n\n### Was this patch authored or co-authored using generative AI tooling?\n\nYes, co-authored with Claude Code.\n\nCloses #26 from viirya/ci-09-fix-wait-flakiness-and-dockerignore.\n\nAuthored-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cliangchi.hsieh@databricks.com\u003e\n"
    },
    {
      "commit": "c06c50955f78b59d4654ff7605f6f5342eb6f454",
      "tree": "04126f813eae8079a189456cadc226290fa42b7c",
      "parents": [
        "a131f41f96fe890ad4677f4531545a6deef3bc52"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Sun Sep 13 17:53:52 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "liangchi.hsieh@databricks.com",
        "time": "Sun Sep 13 17:53:52 2026 -0700"
      },
      "message": "[SPARK-59445] Run the multi-replica Redis affinity end-to-end walkthrough in CI\n\nThe e2e workflow now covers baseline routing, JWT identity, the backend\ntrust boundary, and per-tenant pool isolation, but nothing in CI stands\nthe gateway up as more than one replica. The multi-replica walkthrough\nunder deploy/examples/e2e-multi-replica-redis has never run\nautomatically, so the chart\u0027s Redis topology and the gateway\u0027s behaviour\nwhen a replica dies are only verified by hand.\n\nThe cross-replica semantics themselves are already covered in-process by\nthe ha_smoke example, which runs in the integration job against a real\nRedis. What is missing is the Kubernetes shape around it: the bundled\nRedis StatefulSet, the synthesized redis:// URL, two gateway replicas\nbehind one Service, and a pod actually being deleted.\n\nThis adds an e2e-multi-replica-redis job that asserts:\n\n(1) Both gateway replicas report affinity_store\u003dredis on startup. The\npod count is checked first, so a selector that matched nothing could not\nlet the check pass silently.\n\n(2) A session driven through the Service writes its binding to Redis,\nunder the documented key shape, with a value that is one of the real\nbackend pod IPs and a TTL within the configured hour. Redis is asserted\nempty beforehand so a stale key cannot satisfy this.\n\n(3) After deleting one replica, the binding is unchanged, and a client\narriving at the replacement pod -- port-forwarded directly, so the\nsurviving replica cannot serve it -- reaches the same backend. The\nreplacement\u0027s own log must name that backend, an address it could only\nhave read from Redis.\n\n(4) Fresh sessions still spread across both backends. Without this,\nassertion 3 would also pass on a gateway that always picked the same\nbackend regardless of affinity.\n\nVerified locally against a kind cluster first. The replacement replica\nreturned the same serverSideSessionId as the original, confirming the\nsame Spark driver answered, and eight fresh sessions split 4/3 across the\ntwo backends while the pinned session stayed put.\n\n---\n\nVerified locally against a kind cluster (`scg-redis`) before writing the workflow.\nThe decisive evidence for assertion 3: after deleting the replica that had served\nthe session, a client port-forwarded **directly to the replacement pod** got back\n\n```\nserverSideSessionId: 28870e4e-f886-42b9-b686-04a095e0abb9\n```\n\nbyte-identical to what the original replica returned. That id is minted by the\nSpark driver, so the same driver answered — and the replacement pod, which had\nnever seen the session, could only have found that backend in Redis.\n\nThe negative control (assertion 4) split 8 fresh sessions 4/3 across the two\nbackends while the pinned session stayed on its own, so the pinning is a property\nof the binding rather than of the pool.\n\nCross-replica *semantics* are already covered by `ha_smoke` in the integration\njob; this job covers the Kubernetes topology around it.\n\nCloses #24 from viirya/ci-08-e2e-multi-replica-redis.\n\nAuthored-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cliangchi.hsieh@databricks.com\u003e\n"
    },
    {
      "commit": "a131f41f96fe890ad4677f4531545a6deef3bc52",
      "tree": "4d756ee3d4a6f83a836a7da0723a447a6fe36f1f",
      "parents": [
        "33226eea90beec0946f9f250e491d36dc7344141"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Thu Sep 10 13:52:07 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "liangchi.hsieh@databricks.com",
        "time": "Thu Sep 10 13:52:07 2026 -0700"
      },
      "message": "[SPARK-59360] Run the multi-tenant routing end-to-end walkthrough in CI\n\nThe e2e workflow covers baseline routing, JWT identity, and the backend\ntrust boundary, but nothing in CI exercises per-tenant backend pools. The\nmulti-tenant walkthrough under deploy/examples/e2e-multitenant has never\nrun automatically, so the chart\u0027s tenantPools.overrides templating and\nthe gateway\u0027s pool-isolation behaviour are only verified by hand.\n\nThis adds an e2e-multitenant job that stands up two independent Spark\nConnect deployments, one per tenant, and asserts:\n\n(1) The chart renders both tenant pool overrides plus the strict\nonUnknownTenant\u003dreject policy, and the gateway starts one Kubernetes\nEndpoints watcher per override pool with no default pool at all.\n\n(2) Two clients with JWTs claiming different tenants both succeed, and\nthe session.create audit events show each tenant reached only its own\nbackend Endpoint. This is the isolation assertion: the backend IPs are\ncaptured from the two Services beforehand and compared exactly, so a\nrouting regression that sent both tenants to one pool would fail.\n\n(3) A JWT claiming a tenant with no configured pool is rejected with\nPermissionDenied, while a JWT carrying no tenant claim is rejected with\nUnauthenticated. The two rejects must stay distinct: one identified\nitself and was refused, the other never identified a tenant.\n\n(4) Both rejects are observable -- the unknown tenant as an rpc.error\naudit event, the missing claim as an auth.failure with\nreason\u003dmissing_tenant plus its metric -- and only the two successful\nsessions bound a backend.\n\nVerified locally against a kind cluster before wiring it up: each tenant\nstayed pinned to its own backend across repeated sessions, and both\nreject paths returned the documented status codes and messages.\n\nCo-authored-by: Claude Code\n\n---\n\nVerified locally against a kind cluster (`scg-mt`) before writing the workflow.\nEach tenant stayed pinned to its own backend across repeated sessions:\n\n```\nteam-a -\u003e [\u002710.244.0.5:15002\u0027]\nteam-b -\u003e [\u002710.244.0.6:15002\u0027]\n```\n\nmatching the two Services\u0027 Endpoints exactly. Both reject paths returned the\ndocumented status codes and messages, and audit/metrics agreed (2\n`session.create`, 1 `rpc.error`, 1 `auth.failure{reason\u003dmissing_tenant}`,\n`scg_backend_pool_size 0` as documented for strict mode).\n\nCloses #23 from viirya/ci-07-e2e-multitenant.\n\nAuthored-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cliangchi.hsieh@databricks.com\u003e\n"
    },
    {
      "commit": "33226eea90beec0946f9f250e491d36dc7344141",
      "tree": "13e7c2eb86f3be754b6fe38fbcab578eef50bebc",
      "parents": [
        "7cb8aa6c7bf8d0ce3c82b28f812ece979827d903"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Tue Sep 08 17:42:45 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "liangchi.hsieh@databricks.com",
        "time": "Tue Sep 08 17:42:45 2026 -0700"
      },
      "message": "[SPARK-59324] Run the trust-boundary end-to-end walkthrough in CI\n\n### What changes were proposed in this pull request?\n\nThe deploy/examples/e2e-trust-boundary walkthrough is what verifies the\ngateway-to-backend trust boundary: the backend enforces\nspark.connect.authenticate.token, the gateway presents it, and clients must not be\nable to read it back. None of that was covered by CI. The in-process tests cover the\noutbound-credential plumbing, but nothing verified the deployed arrangement or the\nConfig-response withholding against a real Spark backend.\n\nThis adds an e2e-trust-boundary job to the E2E workflow, on its own kind cluster,\nrunning the walkthrough\u0027s four assertions:\n\n(1) A direct connection to the backend, bypassing the gateway, is refused with\nUNAUTHENTICATED \"No authentication token provided\". This runs before the gateway is\ninstalled, so nothing could be presenting a token.\n\n(2) The same client succeeds through the gateway, whose startup log confirms\n\"outbound backend authentication enabled\".\n\n(3) Negative control: with backendToken.enabled\u003dfalse, that same client through that\nsame gateway is refused again with the identical error. Without this, (2) would prove\nnothing; it is what shows the token is the discriminator rather than the network path\nor some other setting.\n\n(4) A client asking the Config RPC for spark.connect.authenticate.token sees it\nunset, and the gateway records a config.redacted audit event naming the withheld key.\n\nAssertion (4) is why the token layer holds at all: Spark\u0027s Config RPC will otherwise\nhand back any config key the server holds, which would let any client read the\ncredential and then dial the backend directly.\n\nThe walkthrough\u0027s NetworkPolicy half is deliberately out of scope: kind\u0027s default CNI\ndoes not enforce NetworkPolicy, which is itself why the token layer matters, since it\ndoes not depend on the CNI.\n\nVerified by running all four assertions locally first: direct connection refused,\nthrough-gateway query returned 10 rows, tokenless gateway refused with the identical\nerror, and the token read back as unset with one config.redacted audit event naming\nspark.connect.authenticate.token.\n\n### Why are the changes needed?\n\nThe trust boundary is a security property: without it, any pod that can reach\nthe backend Service can run arbitrary Spark jobs as anyone. The deployed\narrangement that enforces it — Secret, backend enforcement, gateway outbound\ncredential, and Config withholding — had no CI coverage.\n\n### Does this PR introduce _any_ user-facing change?\n\nNo.\n\n### How was this patch tested?\n\nAll four assertions were run locally before writing the job (see above). This\nPR\u0027s own CI run exercises them on a hosted runner.\n\n### Was this patch authored or co-authored using generative AI tooling?\n\nYes, co-authored with Claude Code.\n\nCloses #22 from viirya/ci-06-e2e-trust-boundary.\n\nAuthored-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cliangchi.hsieh@databricks.com\u003e\n"
    },
    {
      "commit": "7cb8aa6c7bf8d0ce3c82b28f812ece979827d903",
      "tree": "9e5c85501c8150f9853c94c173223f32274844a9",
      "parents": [
        "f17fbaa6c0544eb1c7b2069d6a8ef39480a9effa"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Mon Sep 07 19:58:11 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "liangchi.hsieh@databricks.com",
        "time": "Mon Sep 07 19:58:11 2026 -0700"
      },
      "message": "[SPARK-59270] Run the JWT auth end-to-end walkthrough in CI\n\n### What changes were proposed in this pull request?\n\nThe `deploy/examples/e2e-auth-jwt` walkthrough is the only thing that exercises the\nHelm chart\u0027s auth block end to end. The in-process tests cover `JwtAuthenticator`\nitself, but nothing rendered the chart\u0027s `auth.type: jwt` ConfigMap or drove a real\nPySpark client through the bearer-token transport — so a break in the chart\u0027s auth\ntemplating, or in claim propagation, would go unnoticed.\n\nThis adds an `e2e-auth-jwt` job to the E2E workflow, running in parallel with\n`e2e-smoke` on its own kind cluster. It asserts:\n\n- The chart renders `auth.type: jwt` with the expected issuer and claim mapping\n  (`tenant_claim`, `groups_claim`).\n- A valid HS256 token is accepted, and the `sub` / `tenant` / `groups` claims reach\n  the audit pipeline: the `session.create` event must report **`user_id\u003dalice`, not\n  `anonymous`**, plus `tenant\u003dteam-a` and `groups\u003ddevs,admins`.\n- Three rejection paths fail with `UNAUTHENTICATED`: an expired token, a token\n  signed with the wrong secret, and no token at all.\n- **Signature-validation failures all report the same generic `\"invalid JWT\"`.** A\n  verifier that distinguishes \"expired\" from \"bad signature\" tells an attacker which\n  half to work on, so this is asserted rather than merely observed.\n- The `auth.failure` audit events use only the fixed reasons `invalid_token` and\n  `missing_token`, and `scg_auth_failures_total` counts both.\n\nTokens are signed with the 8-line `openssl` helper from the walkthrough — no JWT\nlibrary needed. PySpark is pinned to `4.0.0` to match the backend image, as in the\n`e2e-smoke` job.\n\n### Why are the changes needed?\n\nAuth is the gateway\u0027s security boundary, and the chart\u0027s auth templating had no\ncoverage at all: the unit tests never render the chart, and `e2e-smoke` runs with\n`auth: none`. This also pins down the deliberate non-disclosure property of the\nerror messages, which is easy to regress by \"improving\" an error string.\n\n### Does this PR introduce _any_ user-facing change?\n\nNo.\n\n### How was this patch tested?\n\nRan the whole walkthrough locally before writing the job: the chart rendered the\nexpected auth block, the valid token produced `user_id\u003dalice` / `tenant\u003dteam-a` /\n`groups\u003ddevs,admins` in the audit log, all three rejection paths returned\n`UNAUTHENTICATED` with the expected messages (the two signature failures with the\nidentical generic message), and the audit reasons and metrics matched\n(`invalid_token` x4, `missing_token` x2). This PR\u0027s own run exercises the job on a\nhosted runner.\n\n### Was this patch authored or co-authored using generative AI tooling?\n\nYes, co-authored with Claude Code.\n\nCloses #21 from viirya/ci-05-e2e-more.\n\nAuthored-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cliangchi.hsieh@databricks.com\u003e\n"
    },
    {
      "commit": "f17fbaa6c0544eb1c7b2069d6a8ef39480a9effa",
      "tree": "34634b630fda16437edd62574339c73f48cd93be",
      "parents": [
        "595d9f3b0657705528e06fa8c64776aaabce4afe"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Fri Sep 04 00:33:40 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "liangchi.hsieh@databricks.com",
        "time": "Fri Sep 04 00:33:40 2026 -0700"
      },
      "message": "[SPARK-59224] Add an end-to-end workflow running the gateway on kind with a real PySpark client\n\n### What changes were proposed in this pull request?\n\nThe `deploy/examples/e2e-smoke` walkthrough exercises the whole deployment path —\ngateway image, Helm chart, Kubernetes Endpoints discovery, session affinity, audit\nand metrics — but only by hand. Nothing in CI covered it, so a break in the chart,\nthe Dockerfile or the K8s pool would only surface when someone next ran the\nwalkthrough manually.\n\nThis adds an `E2E` workflow that automates it:\n\n1. build the gateway image\n2. create a kind cluster, load the image into it\n3. deploy two `apache/spark:4.0.0` Spark Connect servers\n4. install the gateway with the Helm chart + the example values\n5. assert the gateway discovered **both** backends (K8s Endpoints watch)\n6. run the repo\u0027s own `test/integration/client_smoke.py` through a port-forward\n7. assert `scg_backend_pool_size`, the `ExecutePlan` counters, and that the audit\n   log recorded `ExecutePlan`\n\n`kind` and `helm` are installed with plain `curl` at pinned versions rather than\nthird-party actions — both to stay clear of the ASF GitHub Actions allowlist and to\nkeep the versions explicit. The only action used is `actions/checkout`. On failure\nthe job dumps pod state and gateway/Spark logs; the kind cluster is always deleted.\n\n**Cost on a hosted runner — measured, not estimated:** **6m22s**, broken down as\n`docker build` 208s (55%), PySpark install + client 57s, `kind create cluster` 41s,\nSpark backends ready (~700 MiB image pull) 43s, Helm install + gateway ready 11s,\ntool install and assertions ~14s. It runs in parallel with the existing `build`\njob (2m19s), so it adds roughly four minutes to a PR\u0027s wall clock in exchange for\ncovering the whole deployment path. Most of that is the release build, so there is\nheadroom later if it needs trimming (caching the Rust build layers).\n\nThe PySpark client is pinned to `4.0.0` to match the `apache/spark:4.0.0` backend:\npyspark 4.2\u0027s `createDataFrame` reads the SQL config\n`spark.sql.session.localRelationSizeLimit`, which Spark 4.0 does not have, and the\nserver answers `SQL_CONF_NOT_FOUND`.\n\n### Why are the changes needed?\n\nThis is the only coverage of the real deployment path: the Helm chart, the\nKubernetes Endpoints-watch pool, and session affinity as observed by an actual\nSpark client. The in-process integration tests cover the proxy logic, but nothing\nelse verifies that the chart deploys a working gateway.\n\n### Does this PR introduce _any_ user-facing change?\n\nNo.\n\n### How was this patch tested?\n\nRan the entire walkthrough locally before writing the workflow. The PySpark client\nreturned correct results including a TempView query — the meaningful check that\nsession affinity held, since a TempView lives in one driver\u0027s memory and a\nmisrouted follow-up RPC would fail it. `scg_backend_pool_size 2`, the RPC counters\nand the audit records were all as expected. This PR\u0027s own run exercises the\nworkflow on a hosted runner.\n\n### Was this patch authored or co-authored using generative AI tooling?\n\nYes, co-authored with Claude Code.\n\nCloses #20 from viirya/ci-04-e2e-nightly.\n\nAuthored-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cliangchi.hsieh@databricks.com\u003e\n"
    },
    {
      "commit": "595d9f3b0657705528e06fa8c64776aaabce4afe",
      "tree": "788a97a2c54a5bc2cb9b268053e8612a7404a94b",
      "parents": [
        "cdea6eb6ebb2d7b01aabdb6c9aba135e96fb0940"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Thu Sep 03 10:00:19 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "liangchi.hsieh@databricks.com",
        "time": "Thu Sep 03 10:00:19 2026 -0700"
      },
      "message": "[SPARK-59196] Run the graceful-drain and multi-replica HA smoke examples in CI\n\n### What changes were proposed in this pull request?\n\nTwo binaries under `crates/proxy/examples/` assert end-to-end invariants, but were\nonly ever *compiled* by `cargo clippy --all-targets` and never executed — so the\nbehaviour they cover had no CI verification at all:\n\n- **`drain_smoke`** — exercises the two-step drain the gateway performs on SIGTERM\n  (in-process, triggered through the same channel `shutdown_signal()` uses).\n  Asserts that `/readyz` flips to not-ready when the drain starts, that in-flight\n  streams keep receiving messages through the drain, and that the gateway shuts\n  down cleanly afterwards. **No external dependencies.**\n- **`ha_smoke`** — spawns two real `SparkConnectProxy` gateways sharing one Redis\n  affinity store and one static pool of two fake backends, then verifies three\n  multi-replica HA properties: shared affinity state across replicas, that\n  failover preserves an existing session binding, and that the op-id reverse index\n  resolves across replicas. Needs a Redis (`REDIS_URL`, default\n  `redis://127.0.0.1:6399`).\n\nBoth signal failure by panicking on an assert, so a plain `cargo run` is a\nsufficient CI check. This adds them as two steps in the `integration` job, plus a\nRedis **service container** on `:6399` for `ha_smoke`.\n\nThe Redis *tests* in that job are unaffected — they start their own container\nthrough `testcontainers` and do not use the service container.\n\nThe kind/Helm end-to-end walkthroughs under `deploy/examples/` remain out of\nscope; they need a much heavier harness and are better suited to a separate\nscheduled workflow.\n\n### Why are the changes needed?\n\nGraceful drain and multi-replica HA are both production behaviours — the first is\nwhat makes a rolling restart not drop in-flight queries, the second is what makes\nrunning more than one replica correct. Neither had any automated coverage, even\nthough the code to verify them was already in the tree.\n\n### Does this PR introduce _any_ user-facing change?\n\nNo.\n\n### How was this patch tested?\n\nLocally, with a Redis on `:6399` standing in for the service container, running\nall four steps of the job in sequence: the testcontainers Redis tests still pass\n(4 + 5), the tracing test passes, and both smoke examples exit 0 — \"all drain\ninvariants passed\" and \"all HA invariants passed\". This PR\u0027s own CI run exercises\nthe new steps and the service container.\n\n### Was this patch authored or co-authored using generative AI tooling?\n\nYes, co-authored with Claude Code.\n\nCloses #19 from viirya/ci-03-smoke-examples.\n\nAuthored-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cliangchi.hsieh@databricks.com\u003e\n"
    },
    {
      "commit": "cdea6eb6ebb2d7b01aabdb6c9aba135e96fb0940",
      "tree": "9979e00010a8a1ce642c1b3ec9cb5bc37f2dfbb9",
      "parents": [
        "28a6b19c9c9faf9330919ed7c9450c9701aab571"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Wed Sep 02 22:40:31 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "liangchi.hsieh@databricks.com",
        "time": "Wed Sep 02 22:40:31 2026 -0700"
      },
      "message": "[SPARK-59189] Run the proxy tracing integration test in CI\n\n### What changes were proposed in this pull request?\n\n`crates/proxy/tests/tracing_integration.rs` verifies the gateway\u0027s OpenTelemetry\nbehaviour end to end — that a forwarded RPC emits application spans, and that the\nW3C `traceparent` is propagated to the backend. It is `#[ignore]` by default, so\nit has never run in CI.\n\nThe reason it opts out of workspace-wide runs is a Cargo feature-unification\neffect, **not** a flaky test: `cargo test --workspace` puts `kube` in the\ndependency graph, which transitively enables `opentelemetry`\u0027s\n`spec_unstable_logs_enabled`. The `tracing-opentelemetry` that then links silently\ndrops application spans, so only h2-internal spans reach the in-memory exporter\nand the assertions fail. Built on its own, the test passes.\n\nThis adds a step to the `integration` job (added in SPARK-59161) that runs the\ntest with its own package-scoped cargo invocation:\n\n```\ncargo test -p scg-proxy --test tracing_integration -- --ignored\n```\n\nThe `#[ignore]` marker stays, so workspace-wide runs remain reliable.\n\nThe job is also renamed from \"integration tests (Docker)\" to **\"integration\ntests\"** — this test needs no Docker, and the job now covers integration tests\nthat simply need their own cargo invocation. Each step still says what it needs.\n\n### Why are the changes needed?\n\nTracing is the gateway\u0027s main debugging surface in production, and span emission\nplus `traceparent` propagation had no CI coverage at all. A regression there is\ninvisible until someone tries to debug a live request.\n\n### Does this PR introduce _any_ user-facing change?\n\nNo.\n\n### How was this patch tested?\n\nVerified both directions locally, confirming the documented behaviour:\n\n```\ncargo test -p scg-proxy --test tracing_integration -- --ignored   # 1 passed\ncargo test --workspace  --test tracing_integration -- --ignored   # FAILED\n```\n\nThe first is what CI now runs. This PR\u0027s own CI run exercises the new step.\n\n### Was this patch authored or co-authored using generative AI tooling?\n\nYes, co-authored with Claude Code.\n\nCloses #18 from viirya/ci-02-tracing-integration.\n\nAuthored-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cliangchi.hsieh@databricks.com\u003e\n"
    },
    {
      "commit": "28a6b19c9c9faf9330919ed7c9450c9701aab571",
      "tree": "6a3893a2989f2008503b648a9519b9b8dafaed9d",
      "parents": [
        "f6a702451ff05364246322f0f402246313479e7f"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Wed Sep 02 13:12:33 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "liangchi.hsieh@databricks.com",
        "time": "Wed Sep 02 13:12:33 2026 -0700"
      },
      "message": "[SPARK-59161] Run the Redis-backed integration tests in CI\n\n### What changes were proposed in this pull request?\n\nThe Redis-backed integration tests in `scg-store-redis` and `scg-ratelimit` are\nmarked `#[ignore \u003d \"requires Docker\"]`, so `cargo test --workspace` skips them and\nthey have never run in CI. That leaves 9 integration tests uncovered:\n\n- **`scg-store-redis`** — the Redis `AffinityStore` round-trips:\n  bind/resolve/forget, session TTL expiry, atomic bind-if-absent, the op reverse\n  index, and two stores sharing one Redis.\n- **`scg-ratelimit`** — the distributed token bucket: burst admit-then-reject,\n  refill after wait, per-tenant isolation, and two replicas sharing a bucket.\n\nThese tests bring up their own Redis container through `testcontainers`, and the\nGitHub-hosted runners already provide a Docker daemon, so no `services:` block or\nextra setup is needed. This adds one CI step that runs them explicitly:\n\n```\ncargo test -p scg-store-redis -p scg-ratelimit -- --ignored\n```\n\nThe `#[ignore]` markers stay, so a local `cargo test` still needs no Docker — CI\nopts in instead.\n\nThis lands as a **separate `integration` job** rather than another step in the\nexisting one, so that:\n\n- a Docker-related failure reads as its own red check instead of being reported as\n  an \"fmt / clippy / build / test\" failure,\n- it can be re-run on its own (useful for anything container-backed), and\n- later integration suites have a natural home.\n\nIt runs in parallel with the existing job.\n\n### Why are the changes needed?\n\nThe Redis backends are what make multi-replica HA work (cross-replica session\naffinity and a shared rate-limit bucket), and none of that behaviour was verified\nby CI. A regression in either would have gone unnoticed.\n\n### Does this PR introduce _any_ user-facing change?\n\nNo.\n\n### How was this patch tested?\n\nRan the tests locally with a Docker daemon available:\n`cargo test -p scg-store-redis -p scg-ratelimit -- --ignored` — 9 passed\n(store-redis 5, ratelimit 4) in about 10 seconds including container startup, so\nthe added CI time is small. The rest of the quartet is unchanged. This PR\u0027s own\nCI run exercises the new step on a GitHub-hosted runner.\n\n### Was this patch authored or co-authored using generative AI tooling?\n\nYes, co-authored with Claude Code.\n\nCloses #17 from viirya/ci-01-redis-integration.\n\nAuthored-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cliangchi.hsieh@databricks.com\u003e\n"
    },
    {
      "commit": "f6a702451ff05364246322f0f402246313479e7f",
      "tree": "0fb581a07a470582698fb465752e83bf6e2dabc7",
      "parents": [
        "75ae4e7f093d46cf3e470f591281108691804507"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Tue Sep 01 15:00:16 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "liangchi.hsieh@databricks.com",
        "time": "Tue Sep 01 15:00:16 2026 -0700"
      },
      "message": "[SPARK-59158] Add docs, Helm chart, deployment examples, and Dockerfile\n\n### What changes were proposed in this pull request?\n\nThe **final PR** of the Spark Connect Gateway reference implementation import\n(tracked in the do-not-merge reference PR #1). All Rust code has landed (15\ncrates, SPARK-58753 through SPARK-59145); this brings in everything that isn\u0027t\ncode:\n\n- `Dockerfile` and `config.example.yaml`\n- The full `README.md` (replacing the minimal bootstrap stub) and `ROADMAP.md`\n- `docs/` — architecture, deployment, multitenancy, observability, perf-baseline,\n  runbook\n- `deploy/helm/scg/` — the Helm chart (templates, values, chart README)\n- `deploy/examples/` — end-to-end walkthroughs: smoke, multitenant,\n  multi-replica-redis, scale-test, kubeflow-spark-operator, trust-boundary,\n  auth-jwt, plus plain Spark Connect server manifests\n- `test/integration/` — the Python client smoke test\n\nNo Rust code changes, so the workspace `members` list is untouched.\n\nWith this merged, the import is complete and the reference PR #1 can be closed.\n\n### Why are the changes needed?\n\nImplements the reference gateway from the approved SPIP (SPARK-58455), imported\none reviewable piece at a time. This completes the import.\n\n### Does this PR introduce _any_ user-facing change?\n\nNo code change. It replaces the placeholder README with the real project README\nand adds user-facing documentation, a Helm chart, and deployment examples.\n\n### How was this patch tested?\n\n`cargo fmt --all --check`, `cargo clippy --workspace --all-targets -- -D warnings`,\n`cargo build --workspace`, and `cargo test --workspace` all still pass locally\n(204 tests pass, 10 `#[ignore]`) — unaffected, since there are no Rust changes.\n`helm lint deploy/helm/scg` passes. All relative links in README.md and docs/\nwere verified to resolve. Every file requiring an Apache license header has one;\n`.md`/`.json`, Helm `NOTES.txt` and `.helmignore` are covered by the existing\nlicenserc ignore globs.\n\n### Was this patch authored or co-authored using generative AI tooling?\n\nYes, co-authored with Claude Code.\n\nCloses #16 from viirya/import-13-docs-deploy.\n\nAuthored-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cliangchi.hsieh@databricks.com\u003e\n"
    },
    {
      "commit": "75ae4e7f093d46cf3e470f591281108691804507",
      "tree": "c73d19120364c8852fa146a85afd8d03e3256459",
      "parents": [
        "52f1dc0d7c7cf73db9c5d1fd602931459437d3fe"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Tue Sep 01 10:09:21 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "liangchi.hsieh@databricks.com",
        "time": "Tue Sep 01 10:09:21 2026 -0700"
      },
      "message": "[SPARK-59145] Add gateway binary crate\n\n### What changes were proposed in this pull request?\n\nPart of importing the Spark Connect Gateway reference implementation into\n`apache/spark-connect-gateway` as a series of small, bottom-up PRs (tracked in\nthe do-not-merge reference PR #1).\n\nThis adds the `gateway` crate — the binary entry point that wires all the\npreviously imported layers into a runnable server. It reads the YAML config\n(`--config`, default `config.yaml`), builds the chosen backend pool (static or\nKubernetes service-watch) and affinity store (in-memory or Redis), installs the\nconfigured authenticator, tenant resolver, rate limiter, and audit sink, wraps\nthe pool in active health probing, starts the admin HTTP server for metrics and\nhealth endpoints, and serves the Spark Connect gRPC surface through the proxy.\n\n`gateway` depends on all 14 other crates in the workspace, so it lands last of\nthe code PRs and closes the dependency graph. With this PR the workspace\n`members` list is complete (15 crates) and `cargo build --workspace` produces a\nrunnable `gateway` binary.\n\nOnly the docs/deploy PR (Dockerfile, Helm chart, examples, docs, full README)\nremains after this one.\n\n### Why are the changes needed?\n\nImplements the reference gateway from the approved SPIP (SPARK-58455), imported\none reviewable piece at a time.\n\n### Does this PR introduce _any_ user-facing change?\n\nNo — the binary is new, and nothing shipped previously.\n\n### How was this patch tested?\n\n`cargo fmt --all --check`, `cargo clippy --workspace --all-targets -- -D warnings`,\n`cargo build --workspace`, and `cargo test --workspace` all pass locally (204\ntests pass, 10 Redis/testcontainer tests `#[ignore]`). The built binary runs:\n`gateway --help` prints its CLI.\n\n### Was this patch authored or co-authored using generative AI tooling?\n\nYes, co-authored with Claude Code.\n\nCloses #15 from viirya/import-12-gateway.\n\nAuthored-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cliangchi.hsieh@databricks.com\u003e\n"
    },
    {
      "commit": "52f1dc0d7c7cf73db9c5d1fd602931459437d3fe",
      "tree": "dc352aef2a6f39bdb9a396ab06a26fe51a54f039",
      "parents": [
        "aa16a0406cd5dd81759dabd45d361d6893e911f4"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Tue Sep 01 00:15:11 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "liangchi.hsieh@databricks.com",
        "time": "Tue Sep 01 00:15:11 2026 -0700"
      },
      "message": "[SPARK-59103] Add scg-proxy crate\n\n### What changes were proposed in this pull request?\n\nPart of importing the Spark Connect Gateway reference implementation into\n`apache/spark-connect-gateway` as a series of small, bottom-up PRs (tracked in\nthe do-not-merge reference PR #1).\n\nThis adds the `scg-proxy` crate — the gRPC proxy core of the gateway, and the\ncrate that ties the previously imported layers together. It contains:\n\n- **handler** — a `SparkConnectService` implementation that forwards every RPC to\n  a backend chosen by routing + session affinity, stamping the authenticated\n  identity onto the forwarded request\n- **dial** — a per-backend tonic `Channel` cache, so the dial cost isn\u0027t paid per\n  RPC\n- **outbound** — the credentials the gateway presents on the gateway→backend hop\n- **config_filter** — withholds backend-only secrets from `Config` RPC responses\n\n`scg-proxy` depends on `scg-routing`, `scg-auth`, `scg-genproto`,\n`scg-observability`, `scg-tenant`, `scg-ratelimit`, and `scg-audit`. Its examples\nand integration tests additionally use `scg-store-memory`, `scg-store-redis`,\n`scg-pool-static`, and `scg-healthcheck` as dev-dependencies — all of which are\nalready in the workspace, which is why this PR lands after them.\n\nIt is imported complete: src, all integration tests (forwarding, auth, audit,\ntenant, ratelimit, observability/tracing, multi-tenant e2e, clone-session,\noutbound-token), and all examples (ha_smoke, drain_smoke, load, e2e_client,\nfake_collector).\n\n### Why are the changes needed?\n\nImplements the reference gateway from the approved SPIP (SPARK-58455), imported\none reviewable piece at a time.\n\n### Does this PR introduce _any_ user-facing change?\n\nNo.\n\n### How was this patch tested?\n\n`cargo fmt --all --check`, `cargo clippy --workspace --all-targets -- -D warnings`,\n`cargo build --workspace`, and `cargo test --workspace` all pass locally: 201\ntests pass, 10 Redis/testcontainer tests are `#[ignore]` so no Docker is needed\nin CI. Note `--all-targets` compiles all of this crate\u0027s examples and integration\ntest binaries, which is why its four dev-dependency crates had to land first.\n\n### Was this patch authored or co-authored using generative AI tooling?\n\nYes, co-authored with Claude Code.\n\nCloses #14 from viirya/import-11-proxy.\n\nAuthored-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cliangchi.hsieh@databricks.com\u003e\n"
    },
    {
      "commit": "aa16a0406cd5dd81759dabd45d361d6893e911f4",
      "tree": "0468e6cd4dfadbdc1181dd6fa6ed49abaa8a6824",
      "parents": [
        "cf716c0b8d05d40270997501cbb4e54f0b341cbe"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Sat Aug 29 00:44:23 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "liangchi.hsieh@databricks.com",
        "time": "Sat Aug 29 00:44:23 2026 -0700"
      },
      "message": "[SPARK-59074] Add scg-pool-k8s crate\n\n### What changes were proposed in this pull request?\n\nPart of importing the Spark Connect Gateway reference implementation into\n`apache/spark-connect-gateway` as a series of small, bottom-up PRs (tracked in\nthe do-not-merge reference PR #1).\n\nThis adds the `scg-pool-k8s` crate — a Kubernetes service-watch backend `Pool`.\nIt watches a `Service`\u0027s `Endpoints` object in a given namespace and exposes the\nready endpoints as the routing backend set, so the gateway\u0027s backend membership\ntracks the Spark Connect pods dynamically. This is the dynamic counterpart to\nthe fixed-membership `scg-pool-static`.\n\n`scg-pool-k8s` depends on the `scg-routing` crate whose `Pool` trait it\nimplements, and pulls in the kube / k8s-openapi client stack. It is imported as\nits own PR (rather than grouped) because of that larger dependency surface. It\ncompiles and tests on its own (unit tests plus a router integration test) and is\nappended to the workspace `members` list. It is consumed later by the `gateway`\nbinary.\n\n### Why are the changes needed?\n\nImplements the reference gateway from the approved SPIP (SPARK-58455), imported\none reviewable piece at a time.\n\n### Does this PR introduce _any_ user-facing change?\n\nNo.\n\n### How was this patch tested?\n\n`cargo fmt --all --check`, `cargo clippy --workspace --all-targets -- -D warnings`,\n`cargo build --workspace`, and `cargo test --workspace` all pass locally.\n\n### Was this patch authored or co-authored using generative AI tooling?\n\nYes, co-authored with Claude Code.\n\nCloses #13 from viirya/import-10-pool-k8s.\n\nAuthored-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cliangchi.hsieh@databricks.com\u003e\n"
    },
    {
      "commit": "cf716c0b8d05d40270997501cbb4e54f0b341cbe",
      "tree": "82f34b06edcc9e98acc7d190216772efd2ecb6d0",
      "parents": [
        "2d08fc0249a99f93cd2e36869f18f7188f064707"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Thu Aug 27 23:02:29 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "liangchi.hsieh@databricks.com",
        "time": "Thu Aug 27 23:02:29 2026 -0700"
      },
      "message": "[SPARK-59046] Add scg-store-redis and scg-healthcheck crates\n\n### What changes were proposed in this pull request?\n\nPart of importing the Spark Connect Gateway reference implementation into\n`apache/spark-connect-gateway` as a series of small, bottom-up PRs (tracked in\nthe do-not-merge reference PR #1).\n\nThis adds two more crates that implement/wrap the `scg-routing` traits, grouped\ninto one PR:\n\n- `scg-store-redis` — a Redis-backed `AffinityStore` for cross-replica HA (the\n  multi-replica counterpart to the in-process `scg-store-memory`).\n- `scg-healthcheck` — active gRPC health probing wrapped around a `Pool`,\n  evicting unhealthy backends from routing.\n\nBoth depend on the `scg-routing` crate. They compile and test on their own;\nstore-redis\u0027s Redis integration tests are `#[ignore]` so they compile but don\u0027t\nrequire Docker in CI. Both are appended to the workspace `members` list and are\nconsumed later by the `proxy` crate (both are proxy dev-dependencies used by its\nexamples/tests). With this PR, all four proxy dev-dependency crates\n(store-memory, pool-static, store-redis, healthcheck) are in place, so the proxy\nPR can follow.\n\n### Why are the changes needed?\n\nImplements the reference gateway from the approved SPIP (SPARK-58455), imported\none reviewable piece at a time.\n\n### Does this PR introduce _any_ user-facing change?\n\nNo.\n\n### How was this patch tested?\n\n`cargo fmt --all --check`, `cargo clippy --workspace --all-targets -- -D warnings`,\n`cargo build --workspace`, and `cargo test --workspace` all pass locally\n(Redis integration tests are `#[ignore]`, so no Docker is required in CI).\n\n### Was this patch authored or co-authored using generative AI tooling?\n\nYes, co-authored with Claude Code.\n\nCloses #12 from viirya/import-09-store-redis-healthcheck.\n\nAuthored-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cliangchi.hsieh@databricks.com\u003e\n"
    },
    {
      "commit": "2d08fc0249a99f93cd2e36869f18f7188f064707",
      "tree": "f472cad514c3419cc6026486921f158adb7deb07",
      "parents": [
        "5b52380754ad4aaab33c46f24f99b73a3834baaa"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Wed Aug 26 23:50:19 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "liangchi.hsieh@databricks.com",
        "time": "Wed Aug 26 23:50:19 2026 -0700"
      },
      "message": "[SPARK-59021] Add scg-store-memory and scg-pool-static crates\n\n### What changes were proposed in this pull request?\n\nPart of importing the Spark Connect Gateway reference implementation into\n`apache/spark-connect-gateway` as a series of small, bottom-up PRs (tracked in\nthe do-not-merge reference PR #1).\n\nThis adds two small crates that implement the `scg-routing` traits, grouped into\none PR:\n\n- `scg-store-memory` — an in-process `AffinityStore` for single-replica\n  deployments and tests (for multi-replica HA, `scg-store-redis`, added in a\n  later PR, is used instead).\n- `scg-pool-static` — a backend `Pool` whose membership is fixed at startup (the\n  dynamic K8s-Endpoints-watch variant `scg-pool-k8s` comes in a later PR).\n\nBoth depend on the `scg-routing` crate whose traits they implement. They compile\nand test on their own (2 unit tests each) and are appended to the workspace\n`members` list. They are consumed later by the `proxy` crate (both are proxy\ndev-dependencies used by its tests).\n\n### Why are the changes needed?\n\nImplements the reference gateway from the approved SPIP (SPARK-58455), imported\none reviewable piece at a time.\n\n### Does this PR introduce _any_ user-facing change?\n\nNo.\n\n### How was this patch tested?\n\n`cargo fmt --all --check`, `cargo clippy --workspace --all-targets -- -D warnings`,\n`cargo build --workspace`, and `cargo test --workspace` all pass locally\n(store-memory 2 tests, pool-static 2 tests).\n\n### Was this patch authored or co-authored using generative AI tooling?\n\nYes, co-authored with Claude Code.\n\nCloses #11 from viirya/import-08-store-memory-pool-static.\n\nAuthored-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cliangchi.hsieh@databricks.com\u003e\n"
    },
    {
      "commit": "5b52380754ad4aaab33c46f24f99b73a3834baaa",
      "tree": "e45b8da21b66286d37e28e9f5adf98153f59dd4e",
      "parents": [
        "8415202509f8dfef30d7f83c8ab29aebb44175a6"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Wed Aug 26 01:30:24 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "liangchi.hsieh@databricks.com",
        "time": "Wed Aug 26 01:30:24 2026 -0700"
      },
      "message": "[SPARK-58984] Add scg-tenant crate\n\n### What changes were proposed in this pull request?\n\nPart of importing the Spark Connect Gateway reference implementation into\n`apache/spark-connect-gateway` as a series of small, bottom-up PRs (tracked in\nthe do-not-merge reference PR #1).\n\nThis adds the `scg-tenant` crate — tenant resolution for multi-tenant routing.\nThe gateway\u0027s routing key is `(tenant, user_id, session_id)`; this crate\nresolves the tenant for a request so the proxy can route and isolate per tenant.\n\n`scg-tenant` depends on the `scg-auth` crate for the request identity it\nresolves the tenant from. It compiles and tests on its own (9 unit tests) and is\nappended to the workspace `members` list. It is consumed later by the `proxy`\ncrate.\n\n### Why are the changes needed?\n\nImplements the reference gateway from the approved SPIP (SPARK-58455), imported\none reviewable piece at a time.\n\n### Does this PR introduce _any_ user-facing change?\n\nNo.\n\n### How was this patch tested?\n\n`cargo fmt --all --check`, `cargo clippy --workspace --all-targets -- -D warnings`,\n`cargo build --workspace`, and `cargo test --workspace` all pass locally\n(scg-tenant has 9 unit tests).\n\n### Was this patch authored or co-authored using generative AI tooling?\n\nYes, co-authored with Claude Code.\n\nCloses #10 from viirya/import-07-tenant.\n\nAuthored-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cliangchi.hsieh@databricks.com\u003e\n"
    },
    {
      "commit": "8415202509f8dfef30d7f83c8ab29aebb44175a6",
      "tree": "537160ad30d916906813f62886803d72149f5d52",
      "parents": [
        "e6a8cba030414c90e5734fb837cea4c0f1c14738"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Mon Aug 24 19:10:22 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "liangchi.hsieh@databricks.com",
        "time": "Mon Aug 24 19:10:22 2026 -0700"
      },
      "message": "[SPARK-58961] Add scg-audit and scg-ratelimit crates\n\n### What changes were proposed in this pull request?\n\nPart of importing the Spark Connect Gateway reference implementation into\n`apache/spark-connect-gateway` as a series of small, bottom-up PRs (tracked in\nthe do-not-merge reference PR #1).\n\nThis adds two small leaf crates, grouped into one PR:\n\n- `scg-audit` — structured audit logging for the gateway.\n- `scg-ratelimit` — per-tenant (and optional per-user) token-bucket rate\n  limiting, with a Redis-backed distributed token bucket for multi-replica\n  deployments.\n\nBoth are leaf crates (no internal dependencies), so they compile and test on\ntheir own (audit: 6 unit tests; ratelimit: 9 tests, plus Redis integration tests\nthat are `#[ignore]` so they compile but don\u0027t require Docker in CI). They are\nappended to the workspace `members` list, and are consumed later by the `proxy`\ncrate.\n\n### Why are the changes needed?\n\nImplements the reference gateway from the approved SPIP (SPARK-58455), imported\none reviewable piece at a time.\n\n### Does this PR introduce _any_ user-facing change?\n\nNo.\n\n### How was this patch tested?\n\n`cargo fmt --all --check`, `cargo clippy --workspace --all-targets -- -D warnings`,\n`cargo build --workspace`, and `cargo test --workspace` all pass locally\n(audit 6 tests, ratelimit 9 tests; Redis integration tests are `#[ignore]`).\n\n### Was this patch authored or co-authored using generative AI tooling?\n\nYes, co-authored with Claude Code.\n\nCloses #9 from viirya/import-06-audit-ratelimit.\n\nAuthored-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cliangchi.hsieh@databricks.com\u003e\n"
    },
    {
      "commit": "e6a8cba030414c90e5734fb837cea4c0f1c14738",
      "tree": "154a53fc0864e1e975458a9aa4e3f0da0b0938a3",
      "parents": [
        "c28f2038a7e59b7ecd4533492b4850d4bb537e41"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Sun Aug 23 18:12:15 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "liangchi.hsieh@databricks.com",
        "time": "Sun Aug 23 18:12:15 2026 -0700"
      },
      "message": "[SPARK-58930] Add scg-observability crate\n\n### What changes were proposed in this pull request?\n\nPart of importing the Spark Connect Gateway reference implementation into\n`apache/spark-connect-gateway` as a series of small, bottom-up PRs (tracked in\nthe do-not-merge reference PR #1).\n\nThis adds the `scg-observability` crate — the metrics, tracing, and admin\nsurface for the gateway:\n\n- Prometheus metrics for the gateway\n- Distributed tracing — an OpenTelemetry exporter with W3C traceparent\n  propagation\n- An admin HTTP server exposing `/metrics`, `/healthz`, and `/readyz`\n\n`scg-observability` is a leaf crate (no internal dependencies), so it compiles\nand tests on its own (8 unit tests). It is appended to the workspace `members`\nlist, and is consumed later by the `proxy` and `gateway` crates.\n\n### Why are the changes needed?\n\nImplements the reference gateway from the approved SPIP (SPARK-58455), imported\none reviewable piece at a time.\n\n### Does this PR introduce _any_ user-facing change?\n\nNo.\n\n### How was this patch tested?\n\n`cargo fmt --all --check`, `cargo clippy --workspace --all-targets -- -D warnings`,\n`cargo build --workspace`, and `cargo test --workspace` all pass locally\n(scg-observability has 8 unit tests).\n\n### Was this patch authored or co-authored using generative AI tooling?\n\nYes, co-authored with Claude Code.\n\nCloses #7 from viirya/import-05-observability.\n\nAuthored-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cliangchi.hsieh@databricks.com\u003e\n"
    },
    {
      "commit": "c28f2038a7e59b7ecd4533492b4850d4bb537e41",
      "tree": "b87cbcf21c10cef401dc4280381f912d91f3a658",
      "parents": [
        "3d796c446a6203445fb029e24bfae7897536fdef"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Fri Aug 21 00:29:01 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "liangchi.hsieh@databricks.com",
        "time": "Fri Aug 21 00:29:01 2026 -0700"
      },
      "message": "[SPARK-58901] Add scg-auth crate\n\n### What changes were proposed in this pull request?\n\nPart of importing the Spark Connect Gateway reference implementation into\n`apache/spark-connect-gateway` as a series of small, bottom-up PRs (tracked in\nthe do-not-merge reference PR #1).\n\nThis adds the `scg-auth` crate — the authentication layer for the gateway. It\ndefines the `Authenticator` abstraction and the verified `Identity` attached to\nan authenticated request, a tonic interceptor that runs an authenticator and\nstamps the identity onto the request, and several authenticator implementations:\n\n- `AnonymousAuthenticator` — used when auth is explicitly disabled\n- `StaticTokenAuthenticator` — a Bearer-token allowlist\n- `JwtAuthenticator` — verifies a signed JWT against a local key\n- `OidcAuthenticator` — verifies JWTs against a remote JWKS endpoint\n\n`scg-auth` is a leaf crate (no internal dependencies), so it compiles and tests\non its own (30 unit tests). It is appended to the workspace `members` list, and\nis a prerequisite for the `tenant` and `proxy` crates that arrive in later PRs.\n\n### Why are the changes needed?\n\nImplements the reference gateway from the approved SPIP (SPARK-58455), imported\none reviewable piece at a time.\n\n### Does this PR introduce _any_ user-facing change?\n\nNo.\n\n### How was this patch tested?\n\n`cargo fmt --all --check`, `cargo clippy --workspace --all-targets -- -D warnings`,\n`cargo build --workspace`, and `cargo test --workspace` all pass locally\n(scg-auth has 30 unit tests).\n\n### Was this patch authored or co-authored using generative AI tooling?\n\nYes, co-authored with Claude Code.\n\nCloses #5 from viirya/import-04-auth.\n\nAuthored-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cliangchi.hsieh@databricks.com\u003e\n"
    },
    {
      "commit": "3d796c446a6203445fb029e24bfae7897536fdef",
      "tree": "9597fc6253682860871c3bb6d830a4df85a199c1",
      "parents": [
        "f4a8a40b5a6f9eb8d323cec93c97f392e2c952f0"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Wed Aug 19 23:23:49 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "liangchi.hsieh@databricks.com",
        "time": "Wed Aug 19 23:23:49 2026 -0700"
      },
      "message": "[SPARK-58858] Add scg-config crate\n\n### What changes were proposed in this pull request?\n\nPart of importing the Spark Connect Gateway reference implementation into\n`apache/spark-connect-gateway` as a series of small, bottom-up PRs (tracked in\nthe do-not-merge reference PR #1).\n\nThis adds the `scg-config` crate — the YAML configuration layer for the gateway.\nIt defines the config schema (deserialized with serde) and the loader that reads\nand validates the gateway\u0027s YAML configuration at startup.\n\n`scg-config` is a leaf crate (no internal dependencies), so it compiles and\ntests on its own (45 unit tests covering the schema and loader). It is appended\nto the workspace `members` list and is consumed later by the gateway binary.\n\n### Why are the changes needed?\n\nImplements the reference gateway from the approved SPIP (SPARK-58455), imported\none reviewable piece at a time.\n\n### Does this PR introduce _any_ user-facing change?\n\nNo.\n\n### How was this patch tested?\n\n`cargo fmt --all --check`, `cargo clippy --workspace --all-targets -- -D warnings`,\n`cargo build --workspace`, and `cargo test --workspace` all pass locally\n(scg-config has 45 unit tests).\n\n### Was this patch authored or co-authored using generative AI tooling?\n\nYes, co-authored with Claude Code.\n\nCloses #4 from viirya/import-03-config.\n\nAuthored-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cliangchi.hsieh@databricks.com\u003e\n"
    },
    {
      "commit": "f4a8a40b5a6f9eb8d323cec93c97f392e2c952f0",
      "tree": "e3e16adaaa78f01cae7c318697eae04f9bb4e150",
      "parents": [
        "97e5d973c39aa8b000730a053b9d7499deb7a23a"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Tue Aug 18 18:59:37 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "liangchi.hsieh@databricks.com",
        "time": "Tue Aug 18 18:59:37 2026 -0700"
      },
      "message": "[SPARK-58773] Add scg-routing crate\n\n### What changes were proposed in this pull request?\n\nPart of importing the Spark Connect Gateway reference implementation into\n`apache/spark-connect-gateway` as a series of small, bottom-up PRs (tracked in\nthe do-not-merge reference PR #1).\n\nThis adds the `scg-routing` crate — the routing core of the gateway. It defines\nthe central traits and types that the pool and store layers implement and that\nthe proxy routes through:\n\n- `SessionKey` — the routing key derived from a request\n- `Pool` — the backend-pool trait (implemented later by pool-static / pool-k8s)\n- `AffinityStore` — the session-affinity store trait (implemented later by\n  store-memory / store-redis)\n- `Router` — ties them together\n\n`scg-routing` is a leaf crate (no internal dependencies), so it compiles and\ntests on its own. It is appended to the workspace `members` list. Subsequent\nPRs add the crates that depend on it.\n\n### Why are the changes needed?\n\nImplements the reference gateway from the approved SPIP (SPARK-58455), imported\none reviewable piece at a time.\n\n### Does this PR introduce _any_ user-facing change?\n\nNo.\n\n### How was this patch tested?\n\n`cargo fmt --all --check`, `cargo clippy --workspace --all-targets -- -D warnings`,\n`cargo build --workspace`, and `cargo test --workspace` all pass locally\n(scg-routing has 15 unit tests).\n\n### Was this patch authored or co-authored using generative AI tooling?\n\nYes, co-authored with Claude Code.\n\nCloses #3 from viirya/import-02-routing.\n\nAuthored-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cliangchi.hsieh@databricks.com\u003e\n"
    },
    {
      "commit": "97e5d973c39aa8b000730a053b9d7499deb7a23a",
      "tree": "34223acaec6db52c6017fd9c3dd35df7f0149a79",
      "parents": [
        "50e1c1425ff4adbcd34e53481b9d448db54023d2"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Thu Aug 13 20:14:30 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "liangchi.hsieh@databricks.com",
        "time": "Thu Aug 13 20:14:30 2026 -0700"
      },
      "message": "[SPARK-58753] Add Cargo workspace skeleton, vendored protos, and genproto crate\n\n### What changes were proposed in this pull request?\n\nThis is the first of a series of small, bottom-up PRs importing the Spark\nConnect Gateway reference implementation (tracked in the do-not-merge reference\nPR #1). It introduces the Cargo workspace:\n\n- Root `Cargo.toml` with the full `[workspace.package]` and\n  `[workspace.dependencies]` tables. `members` lists only `crates/genproto` for\n  now; subsequent PRs append one or more crates.\n- `Cargo.lock` (tracked — this project ships a binary), `rust-toolchain.toml`.\n- The vendored `spark.connect.*` protos under `proto/spark/connect/` — a\n  read-only mirror of the corresponding files in `apache/spark`.\n- The `scg-genproto` crate, whose `build.rs` generates tonic bindings from the\n  protos at build time via `tonic-prost-build`.\n\nEach PR in this series keeps `main` green under the full CI matrix\n(fmt / clippy --all-targets / build / test --workspace).\n\n### Why are the changes needed?\n\nImplements the reference gateway from the approved SPIP (SPARK-58455), imported\ninto the `apache/spark-connect-gateway` subproject repo one reviewable piece at\na time.\n\n### Does this PR introduce _any_ user-facing change?\n\nNo.\n\n### How was this patch tested?\n\n`cargo fmt --all --check`, `cargo clippy --workspace --all-targets -- -D warnings`,\n`cargo build --workspace`, and `cargo test --workspace` all pass locally.\n\n### Was this patch authored or co-authored using generative AI tooling?\n\nYes, co-authored with Claude Code.\n\nCloses #2 from viirya/import-01-genproto.\n\nAuthored-by: Liang-Chi Hsieh \u003cviirya@gmail.com\u003e\nSigned-off-by: Liang-Chi Hsieh \u003cliangchi.hsieh@databricks.com\u003e\n"
    },
    {
      "commit": "50e1c1425ff4adbcd34e53481b9d448db54023d2",
      "tree": "29a28d5b4fe3b05a267d274f8f94fc6762c847c2",
      "parents": [
        "b1eb448764a0a99148bd0b3c1ec28c3576403d92"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Wed Aug 12 14:10:43 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Wed Aug 12 14:10:43 2026 -0700"
      },
      "message": "Add .gitattributes\n\nNormalize line endings to LF on checkin so cross-platform contributions don\u0027t\nintroduce CRLF noise, and mark common binary types as binary. Release\nexport-ignore rules can be added later alongside the source-release tooling.\n\nCo-authored-by: Claude Code\n"
    },
    {
      "commit": "b1eb448764a0a99148bd0b3c1ec28c3576403d92",
      "tree": "7628a4b6382e60445f03473ccf2b68b68903dee7",
      "parents": [
        "99fcd1febc6631ce2fdd5551355f41098a6adffd"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Wed Aug 12 10:11:59 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Wed Aug 12 10:11:59 2026 -0700"
      },
      "message": "Use an allowlisted CI toolchain setup\n\nThe Apache Actions allowlist rejects dtolnay/rust-toolchain, so install the\nRust toolchain through a local composite action (.github/actions/setup-builder)\nthat uses rustup, matching apache/datafusion and apache/iceberg-rust. Pin\narduino/setup-protoc and swatinem/rust-cache to the commit SHAs those Apache\nprojects use (which are on the allowlist), and pin actions/checkout.\n\nCo-authored-by: Claude Code\n"
    },
    {
      "commit": "99fcd1febc6631ce2fdd5551355f41098a6adffd",
      "tree": "1122b8ec6ae8ffaf100ad60867e157ac687a182d",
      "parents": [
        "ce35815a2a5e0c7a37e28d2a59fd3ea31d07dcd7"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Wed Aug 12 07:19:03 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Wed Aug 12 07:19:03 2026 -0700"
      },
      "message": "Exclude dev/.rat-excludes from license header check\n\nThe RAT exclude list is a plain data file with no place for an Apache license\nheader, but the skywalking-eyes check flagged it as missing one. Add\n**/.rat-excludes to the licenserc paths-ignore, alongside the other headerless\nconfig files (.gitignore, .gitattributes, .helmignore).\n\nCo-authored-by: Claude Code\n"
    },
    {
      "commit": "ce35815a2a5e0c7a37e28d2a59fd3ea31d07dcd7",
      "tree": "1ca562669717bca02aa8e59752e3c0577057617f",
      "parents": [
        "96b6d77ab0c8aa5939cbc45a365ced299775b9cd"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Tue Aug 11 15:26:35 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Tue Aug 11 16:11:07 2026 -0700"
      },
      "message": "Add dev/merge_spark_pr.py\n\nCopy Apache Spark\u0027s PR merge tool, which Spark projects use to merge pull\nrequests (squash + well-formed commit message + optional JIRA resolution)\ninstead of the GitHub merge button. Subproject changes: point GITHUB_BASE /\nGITHUB_API_BASE at apache/spark-connect-gateway, treat main as the default\nbranch for fix-version selection, filter SPARK JIRA fix versions by the\nconnect-gateway- prefix, and adjust the header comments. The SPARK JIRA\nintegration is kept as-is since this subproject uses the SPARK project.\n\nCo-authored-by: Claude Code\n"
    },
    {
      "commit": "96b6d77ab0c8aa5939cbc45a365ced299775b9cd",
      "tree": "8e7c43f82ea9bafbef7250ce26a2ee5202992a0a",
      "parents": [
        "468317c42eb4a03ca1f308825f5d6f9fe1fd1157"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Tue Aug 11 14:49:09 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Tue Aug 11 16:11:07 2026 -0700"
      },
      "message": "Add a minimal README\n\nA bootstrap-stage stub: what the gateway is, that it is an Apache Spark\nsubproject, its current bootstrapping status, contribution process, and\nlicense. Deliberately avoids linking to docs/crates/deploy paths that do not\nexist yet; a full README follows when the implementation is imported.\n\nCo-authored-by: Claude Code\n"
    },
    {
      "commit": "468317c42eb4a03ca1f308825f5d6f9fe1fd1157",
      "tree": "0a407e4eb74e65cc7cac8f8106a96db996f73187",
      "parents": [
        "007ccd25e530c1f9c28b9c6dbeed8af33f4a1a9e"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Tue Aug 11 14:20:39 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Tue Aug 11 16:11:07 2026 -0700"
      },
      "message": "Add .gitignore\n\nBased on the reference implementation (target/, Cargo.lock.bak, *.swp),\nextended with editor/OS noise, Helm chart packaging artifacts, Apache RAT\naudit output, and the Apache release distribution directory. Cargo.lock is\ntracked (this project ships a binary), so only the .bak is ignored.\n\nCo-authored-by: Claude Code\n"
    },
    {
      "commit": "007ccd25e530c1f9c28b9c6dbeed8af33f4a1a9e",
      "tree": "b59bd8f120cf1d24beb5365fa5acc265999f1a96",
      "parents": [
        "9df079898de0a96ce8dfa224c60bd3c72cad0ab4"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Tue Aug 11 14:07:15 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Tue Aug 11 16:11:07 2026 -0700"
      },
      "message": "Add basic Rust CI workflow\n\nfmt, clippy (-D warnings), build, and test across the workspace on push and\nPR to main. Installs protoc for the tonic-generated proto crate. Action\nversions follow the toolchain used by the reference implementation.\n\nThis is a baseline; integration tests and a broader matrix can come later.\nThe workflow is a no-op until the gateway Cargo workspace is imported.\n\nCo-authored-by: Claude Code\n"
    },
    {
      "commit": "9df079898de0a96ce8dfa224c60bd3c72cad0ab4",
      "tree": "5a26f5f88767f35e3370d35cd09e45fe3d7a5931",
      "parents": [
        "4a77c4ec21e8222d735c683c724984e10df9b244"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Tue Aug 11 13:20:46 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Tue Aug 11 16:11:07 2026 -0700"
      },
      "message": "Add license header checking\n\nEnforce the Apache-2.0 source header on every tracked source file via\napache/skywalking-eyes, driven by .github/.licenserc.yaml and run in a\ndedicated GitHub Actions workflow. Ignore list covers docs, generated Helm\nNOTES, lockfiles, and build artifacts. dev/.rat-excludes mirrors the ignore\nset for Apache RAT release audits.\n\nThe skywalking-eyes action is pinned to the v0.8.0 commit SHA.\n\nCo-authored-by: Claude Code\n"
    },
    {
      "commit": "4a77c4ec21e8222d735c683c724984e10df9b244",
      "tree": "3ecd9fe8605e186b9c460f9d480fd3c0c69e1f85",
      "parents": [
        "5a8582970a66c381fb187ed88c2a30547081b2b5"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Tue Aug 11 13:10:12 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Tue Aug 11 16:11:07 2026 -0700"
      },
      "message": "Add pull request template\n\nReuse the standard Apache Spark PR template (What changes / Why / user-facing\nchange / How tested / Generative AI tooling), matching apache/spark-connect-rust.\n\nCo-authored-by: Claude Code\n"
    },
    {
      "commit": "5a8582970a66c381fb187ed88c2a30547081b2b5",
      "tree": "9aa41aee04fd50d2d2af729b5f71b2886d285786",
      "parents": [
        "d04be825e344992b86d24bec8388525248161d72"
      ],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Tue Aug 11 12:54:06 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Tue Aug 11 16:11:06 2026 -0700"
      },
      "message": "Add LICENSE and NOTICE\n\nApache License 2.0 (source) and a NOTICE aligned with the Apache Spark\nsubproject convention. The NOTICE retains the Spark Export Control notice\nsince the gateway ships cryptographic functionality (TLS, JWT/OIDC), with\nthe Spark-specific third-party acknowledgements dropped as they do not\napply to this Rust project.\n\nLICENSE-binary / NOTICE-binary are intentionally deferred until a release\nbinary with a confirmed bundled-dependency set exists.\n\nCo-authored-by: Claude Code\n"
    },
    {
      "commit": "d04be825e344992b86d24bec8388525248161d72",
      "tree": "bbbd005eb67e5c069e6519c190db3f828027bf9f",
      "parents": [],
      "author": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Tue Aug 11 12:34:34 2026 -0700"
      },
      "committer": {
        "name": "Liang-Chi Hsieh",
        "email": "viirya@gmail.com",
        "time": "Tue Aug 11 16:11:06 2026 -0700"
      },
      "message": "Add .asf.yaml with Spark subproject metadata\n\nConfigure repo metadata for the Apache Spark Connect Gateway subproject:\nnotifications routed to the Spark mailing lists, SPARK JIRA autolink, and\nsquash/rebase merge buttons. Modeled on apache/spark-connect-rust\u0027s .asf.yaml.\n\nCo-authored-by: Claude Code\n"
    }
  ]
}
