docs/release.md
Collect all release artifacts in the Github draft release page
Create release notes draft (auto-generate?)
Create release videos
Open3D-ML dev_to_master PR and merge
Version bump PR, merge to main
Build all Python version wheels (Linux, macOS, Windows x86_64) (CI)
Build Open3D C++ libraries (Linux, macOS, Windows x86_64) (CI)
Build docs (CI) and push to Open3d_website repo
Build arm64 wheels (with ML Ops):
macOS 12. Python 3.10+ (desktop):
Configure: cmake -DCMAKE_BUILD_TYPE=Release -DDEVELOPER_BUILD=OFF -DCMAKE_EXPORT_COMPILE_COMMANDS=ON -DBUILD_TENSORFLOW_OPS=ON -DBUILD_PYTORCH_OPS=ON -DBUNDLE_OPEN3D_ML=ON ..
Ubuntu 20.04 Python (all versions) (follow docs/arm.rst) (desktop)
Build command: cd docker; ./docker_build.sh openblas-arm64-py311; ...
Build Open3D app
Ubuntu 20.04, Windows 10, macOS 10.15 x86_64: (CI)
macOS 12 arm64 (desktop)
macOS (x86_64, arm64) sign (desktop):
From build/bin directory: ../../cpp/apps/sign_open3d_app.sh Open3D.app ../../cpp/apps/Open3DViewer/Open3dViewer.entitlements <apple-id> <cert-name> <team-id> <app-password>
(TBD) Windows app sign
Testing: Run all (especially visualization) examples and Open3D viewer with (Suzanne, Khronos helmet, large point cloud - apartment), on Linux x86-64, macOS (x86-64, arm64), Windows x86-64.
PyPI: Upload wheels
Github: Publish release with auto tag version
Github: Create Open3D-ML release (auto tag version, auto-generate release notes)
git pull docs to website (Google cloud server open3d:instance1)
open3d.org (wordpress) : Update downloads table and post release notes
Upload release video to YouTube
Announce on Twitter, Discord, etc.