+1 (binding)
I've done the usual validation steps, with some minor changes due to the
different process:
- compared the commit id to the minifi-cpp-1.1.0-RC1 tag, that's no
longer specified in the initial email
- only checking SHA512 since SHA256 is no longer there, nor is there a
hash in the email
- the hash files now contain the filename in addition to the hash, I've
just modified my script to cut it off
Built with GCC 16.2 and Clang 22.1 on Arch Linux, and manually tested
the Clang build on the same system. I've tested the new ClassifyImage
processor with a yolo26n-cls based model that I've been working with in
the past, and it worked as expected.
Also built with the same compilers on Fedora Asahi Remix (aarch64).
After all builds, I've run the unit tests too, to verify that everything
works as expected.
A few (not release blocking) concerns for the future:
- The ATR system flagged a few issues, like a few missing license
headers, but I don't think it's critical, I think we can fix them for 1.2.
- Ballooning build times. Now that I've built everything all at once,
multiple times, in multiple environments, it's becoming harder and
harder to ignore that all these new features are leading to longer build
times. I'm thinking we could make Conan builds the default in the near
future, so we can avoid compiling all of our dependencies, and use
binary packages instead, to speed up builds. We could publish binary
packages in the Apache artifactory for the most common environments, to
maximize cache hit rate for everyone.
- There are a few warnings on GCC 16, so builds are failing when
compiling with -Werror (-DMINIFI_FAIL_ON_WARNINGS=ON, only enabled
Werror for our code, not dependencies).
- We should change the base image of the docker builds, so that
potential third party extensions built on a glibc-based system can be
loaded by our docker image. Currently the base image is Alpine, which
uses musl libc, which makes it incompatible with glibc-linked extensions.
Overall I think this package is good to go out to our users.
Thanks for taking the RM role, and for test driving the new ATR system
for us.
Márton
On 10/28/26 3:11 PM, [email protected] wrote:
Hello NiFi,
I'd like to call a vote on releasing the following artifacts as
Apache NiFi Minifi Cpp 1.1.0. This vote is being conducted using a
Beta version of the Apache Trusted Releases (ATR) platform.
Please report any bugs or issues to the ASF Tooling team.
The page of the release candidate from ATR revision 00017,
which includes downloads, can be found at:
https://releases.apache.org/vote/nifi-minifi-cpp/1.1.0
The file manifest for ATR revision 00017, including file digests,
can be found at:
https://releases.apache.org/manifest/nifi-minifi-cpp/1.1.0/00017
The release artifacts are signed with one or more OpenPGP keys from:
https://downloads.apache.org/nifi/KEYS
The recorded source commit is:
1a4b84eb8b53eebcf7c92d83f6ae818f440e33ee
Please review the release candidate and vote accordingly.
[ ] +1 Release this package
[ ] +0 Abstain
[ ] -1 Do not release this package (please provide specific comments)
You can vote on ATR at the URL above, or manually by replying to this email.
The vote is open for 72 hours.
Thanks,
Ferenc Gerlits (fgerlits)
This email was sent by [email protected] on the Apache Trusted Releases
platform
