-
Notifications
You must be signed in to change notification settings - Fork 3.9k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
release v4.6.0 #6796
release v4.6.0 #6796
Conversation
/gha run r-configure |
I think it's time to start another release! We need to at least release the R package by February 14th, to stay on CRAN (#6791). But even without that, it's been about 6 months since v4.5.0 and a lot of good changes have piled up. I've listed the PRs / issues that I think we should try to get into the release. Please comment if there are others you want to try to get in. @guolinke @shiyu1994 @StrikerRUS @jmoralez @borchero @btrotta |
This failed with the following:
@shiyu1994 can you please fix that in the repo settings? I can update |
Did this. Here's the first build: https://readthedocs.org/projects/lightgbm/builds/26974345/ |
I guess we need to replace old |
@shiyu1994 I tried messaging you in the maintainer Slack as well, trying again here in case you check GitHub more frequently. Can you please help us with the workflow token issues here (described in the comments above)? |
I cannot access the URL https://github.com/settings/tokens/503169623. I will figure it out tomorrow. |
@shiyu1994 Do you have rights to create new tokens in the repo? It should be enough... |
/gha run r-configure |
/gha run r-configure |
I think it works now. |
/gha run r-valgrind Status: failure ❌. |
Thanks @shiyu1994 , looks like that worked! This revealed that the valgrind workflow is broken for other, unrelated reasons. Put up #6816 to try to fix it. |
/gha run r-valgrind Workflow R valgrind tests has been triggered! 🚀 Status: success ✔️. |
Docs look good to me after merging the last required PR (#6714)
With that, I think we are ready to release! @guolinke @shiyu1994 @StrikerRUS @jmoralez @borchero can you please approve? Remember that we have to get a release up to CRAN by February 14th, Vienna time... so in about 19 hours 😬 (#6791) Very sorry for the time pressure, I tried to get this ready further ahead of that deadline 😭 I'll merge this once we have at least 2 reviews (including at least 1 from either @guolinke or @shiyu1994). If we don't have enough approvals to merge normally by around 15 hours from now, I'll just upload the R package to CRAN. |
It's currently around 10:30p in Vienna. I have some other obligations later today, so we are really out of time to at least release the R package. I built it on my laptop today and submitted to CRAN. I'm sorry to do that without approvals on this PR, but I thought it was ok to avoid disrupting all of Built locally like this: git clone \
--recursive \
--branch release/v4.6.0 \
-o upstream \
[email protected]:microsoft/LightGBM.git
cd ./LightGBM
sh build-cran-package.sh And tested it locally as one last smoke test. R CMD check --as-cran ./lightgbm_4.6.0.tar.gz 'R CMD check' results (click me)
![]() |
@jameslamb |
At
UPD: Sorry, didn't see those errors are already fixed #6714 (comment). |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Good to go from my POV :)
No problem @StrikerRUS , I'm sorry you had to go through that! Thank you for all your reviews to help get the last couple PRs into this release, it was really really helpful (as always). And thanks for looking @borchero! I hope you'll have more time for LightGBM for the next release cycle, we could really use your help (for example with #5739). I'd still like either @guolinke or @shiyu1994 to also approve before we merge this and continue with the release. |
Thanks friends! I'll merge this and push a tag, to trigger the CI job that produces release artifacts. I'll come back and continue the release (publishing packages to PyPI, etc.) later tonight my time. |
Just did this, using the code in the description of this PR. ![]() ref: https://github.com/microsoft/LightGBM/tags That triggered this run: https://dev.azure.com/lightgbm-ci/lightgbm-ci/_build/results?buildId=17596&view=results Which succeeded 🎉 , and created this draft release: https://github.com/microsoft/LightGBM/releases/tag/untagged-0fb853a31a098d52f907 |
Done. Build: https://readthedocs.org/projects/lightgbm/builds/27196379/
Done. Build: https://readthedocs.org/projects/lightgbm/builds/27196384/
Done. |
Done: https://github.com/microsoft/LightGBM/releases/tag/v4.6.0
Done. Copied from here: https://github.com/microsoft/LightGBM/actions/runs/13336935748 |
This was done earlier: #6796 (comment)
The R package was accepted to CRAN, and so far checks are succeeding. For example, the ![]() And a few precompiled binaries have already been posted: ![]() ref: https://cran.r-project.org/web/packages/lightgbm/index.html So I've updated |
Opened conda-forge/lightgbm-feedstock#70. Also included the new |
Looks like a bot in |
Done: https://www.nuget.org/packages/LightGBM Also opened #6827 to document something I've noticed in the last couple releases. |
Release checklist:
Copied from #6538 with a few changes.
deadline
February 14, 2025 (ref: #6791)
before merge
configure
file of R-package:/gha run r-configure
.development.mode
fromunreleased
torelease
in pkgdown config file.python-package/pyproject.toml
versionadded::
annotations (docs on those)config.h
have*New in version {version}*
comments addedpyarrow
andcffi
are installed #6785/gha run r-valgrind
(docs link)after merge
v*
tag to triggerGitHubRelease
action at Azure Pipelines.stable
tag at GitHub.code to update tags (click me)
stable
version docs on readthedocs (see [ci] Fix version matching between RTD pages and R-package pages #6673 (comment))release/v{n}.{n}.{n}
from RTD versionsafter the release is created by Azure DevOps job
release-drafter
to the automatically-created GitHub release, convert it fromDraft
to published.arm64
macOS wheel to the releasecode to publish to PyPI (summary)
(
gh
is the GitHub CLI, see https://cli.github.com/manual/gh_release_download)Then confirmed that installing the latest wheel works.
pip install -i https://test.pypi.org/simple/ "lightgbm==${NEW_VERSION}" python ./examples/python-guide/logistic_regression.py
Then pushed them to real PyPI.
conda-forge
recipecode to open that dev version PR (click me)
Notes for Reviewers
I believe this should be
v4.6.0
instead ofv4.5.1
because of some small breaking changes:scikit-learn>=0.24.2
, make scikit-learn estimators compatible withscikit-learn>=1.6.0dev
#6651FetchContent_Populate
withFetchContent_MakeAvailable
#6550See https://github.com/microsoft/LightGBM/releases.