+1 (binding)

- Verified signature, checksum, license
- Build + test passed
- LICENSE/NOTICE files exist and valid
- No unexpected binary files
- Ran a few examples on Spark
- Ran against pyiceberg CI [1]

LICENSE/NOTICE looks good after the recent fixes.
I tasked my agent with scanning the repository for correctness issues, and
it flagged a few [2]. None were introduced in 1.12; they also exist in 1.11
or earlier. I haven't had time to fully triage them, but I don't believe
they are release blockers.

Thanks again for running the release!

Best,
Kevin Liu

[1] https://github.com/kevinjqliu/iceberg-python/pull/48
[2] #18004 / #18005, #18173, #18224, #18272


On Tue, Sep 29, 2026 at 9:59 AM <[email protected]> wrote:

> LLM Disclaimer: I generated this using an Agent Skill for validating OSS
> RC's
>
>
> +1 (non-binding)
>
>
> Verified:
>
>
>    - sha512 and GPG signature (Amogh Jahagirdar's key from KEYS)
>
>
>    - source tarball matches the apache-iceberg-1.12.0-rc2 tag (212aa894)
>    apart from the generated iceberg-build.properties and version.txt, and
>    contains no compiled artifacts
>
>
>    - LICENSE and NOTICE, dev/check-license (RAT)
>
>
>    - staged binaries in orgapacheiceberg-1286: signatures on the 12
>    runtime and bundle jars, each jar's LICENSE and NOTICE against the source
>    module, and the third-party packages inside each jar against its LICENSE
>    entries. The RC1 gaps are closed in the jars themselves: Caffeine (#18218)
>    and CEL-Java (#18226) in gcp-bundle, Schubfach and fast_double_parser
>    (#18227) in all eleven Jackson-carrying runtimes. The aws-bundle change
>    from #18195 (HttpClient 4, commons-codec and commons-logging dropped) is
>    reflected in its LICENSE.
>
>
>    - ./gradlew build with tests on Amazon Linux 2023 x86_64, Corretto
>    21.0.12.1: BUILD SUCCESSFUL, 28676 tests, 2590 skipped, 0 failures
>    (kafka-connect integrationTest excluded, no docker compose on the host).
>    The aws integration tests ran against the RustFS container from #18247.
>
>
> On Sep 29, 2026, at 09:39, Russell Spitzer <[email protected]>
> wrote:
>
> 
> +1 (Binding)
> Tested Sigs, Keys
> Tags match
> Staging artifacts are all good (Excluding Spark 4.2 as expected)
> Licenses check out
> Build from source working and all tests passing
>
> On Tue, Sep 29, 2026 at 10:21 AM Alex Dutra <[email protected]> wrote:
>
>> +1 (non-binding)
>>
>> I verified:
>>
>> - Git tag and SHA: OK
>> - GPG signatures: OK
>> - Checksums (sha512): OK
>> - Source tarball matches git tag
>> - version.txt: OK - 1.12.0
>> - RAT license check: OK
>> - Source build: OK
>> - Tests: OK (:iceberg-api:test and :iceberg-core:test)
>> - LICENSE and NOTICE files: OK
>> - Maven artifacts: OK
>>
>> Thanks,
>> Alex
>>
>> On Tue, Sep 29, 2026 at 4:46 PM Manu Zhang <[email protected]>
>> wrote:
>> >
>> > +1 (non-binding)
>> >
>> > Verified:
>> > - Signature and SHA-512 checksum
>> > - Tag apache-iceberg-1.12.0-rc2 points to
>> 212aa894e919ff9fdf89160da4aadcbe6858ac01; source tarball matches the tag
>> > - No binaries in the source tarball; RAT check passes
>> > - Built all modules with JDK 21 (Spark 3.5/4.0/4.1/4.2, Flink
>> 1.20/2.1/2.2/2.3)
>> > - Ran api, core, data and Spark 4.1 tests
>> > - Tested the staged iceberg-spark-runtime-4.1_2.13 on Spark 4.1.1: MoR
>> DML, partition/schema evolution, tag time travel, rewrite_data_files,
>> expire_snapshots, and v3 deletion vectors, row lineage and variant
>> >
>> > On Tue, Sep 29, 2026 at 9:03 AM Anoop Johnson <[email protected]> wrote:
>> >>
>> >> +1 (non-binding)
>> >>
>> >> Thanks Neelesh for running the release!
>> >>
>> >> Verified the source release candidate (commit
>> 212aa894e919ff9fdf89160da4aadcbe6858ac01, tag apache-iceberg-1.12.0-rc2):
>> >>
>> >> - Signature: gpg --verify -> Good signature (Amogh Jahagirdar, key in
>> KEYS)
>> >> - Checksum: shasum -a 512 --check -> OK
>> >> - License: dev/check-license (Apache RAT 0.17) -> passed, no missing
>> headers
>> >> - Built from source
>> >> - Ran :iceberg-api and :iceberg-core test suites -> all green
>> >>   (api: 1287 tests, core: 9075 tests, 0 failures)
>> >>
>> >> Note: I did not run the downstream Spark/Flink/Hive test suites or
>> test the
>> >> staged convenience binaries.
>> >>
>> >> Tested against:
>> >> - Java: OpenJDK 17.0.15
>> >> - OS: Linux (x86_64)
>> >>
>> >> On Mon, Sep 28, 2026 at 3:08 PM Andrei Tserakhau via dev <
>> [email protected]> wrote:
>> >>>
>> >>> +1 (non-binding)
>> >>>
>> >>>
>> >>> On Mon, Sep 28, 2026 at 10:19 PM huaxin gao <[email protected]>
>> wrote:
>> >>>>
>> >>>> +1 (non-binding)
>> >>>>
>> >>>> Verified:
>> >>>> - git tag apache-iceberg-1.12.0-rc2 ==
>> 212aa894e919ff9fdf89160da4aadcbe6858ac01
>> >>>> - sha512 matches
>> >>>> - gpg --verify: good signature from Amogh Jahagirdar
>> (80763BA3EDCCE745), key in KEYS
>> >>>> - version.txt is 1.12.0
>> >>>> - RAT (dev/check-license) passed
>> >>>> - ./gradlew build -x test -x integrationTest passed
>> >>>> - ./gradlew :iceberg-core:test passed
>> >>>> - Staged binaries in orgapacheiceberg-1286 are present with
>> signatures
>> >>>>
>> >>>> The LICENSE findings from RC1 are all fixed: Caffeine now has an
>> entry in
>> >>>> gcp-bundle/LICENSE, and Schubfach and fast_double_parser are present
>> in all 11
>> >>>> bundle and runtime LICENSE files. CEL-Java is in gcp-bundle/LICENSE
>> as well.
>> >>>> The variant NoSuchMethodError fix is also in; ParquetVariantReaders
>> no longer
>> >>>> references commons-lang3.
>> >>>>
>> >>>> Tested with Java 17.0.16 (OpenJDK) on macOS.
>> >>>>
>> >>>> Thanks,
>> >>>> Huaxin
>> >>>>
>> >>>> On Mon, Sep 28, 2026 at 12:59 PM Neelesh Salian <
>> [email protected]> wrote:
>> >>>>>
>> >>>>> Thanks for sorting this out.
>> >>>>> Please continue the voting process.
>> >>>>>
>> >>>>> On Mon, Sep 28, 2026 at 12:54 PM huaxin gao <[email protected]>
>> wrote:
>> >>>>>>
>> >>>>>> I just remembered we intentionally not to release 4.2. Sorry about
>> the false alarm.
>> >>>>>>
>> >>>>>>
>> >>>>>> On Mon, Sep 28, 2026 at 12:33 PM Amogh Jahagirdar <
>> [email protected]> wrote:
>> >>>>>>>
>> >>>>>>> I missed this thread , sounds like it was done intentionally so
>> that development work could continue on 4.2 but there were doubts on
>> releasing it due to evolving spark APIs.
>>
>

Reply via email to