Skip to content

Harden runtime packaging and native loading - #902

Draft
bmehta001 wants to merge 2 commits into
mainfrom
bhamehta/flcore/runtime-packaging-hardening
Draft

Harden runtime packaging and native loading#902
bmehta001 wants to merge 2 commits into
mainfrom
bhamehta/flcore/runtime-packaging-hardening

Conversation

@bmehta001

Copy link
Copy Markdown
Contributor

Summary

  • Package required ORT, GenAI, WinML, and DirectML runtime companions consistently across C++, JS, and Python packaging paths.
  • Make local package extraction compatible with older CMake versions and avoid echoing caller-provided package URLs in configure logs.
  • Preserve Python native loader handles and use ORT_LIB_PATH on POSIX instead of mutating installed package contents.

Validation

  • python sdk_v2/cpp/build.py --build --config RelWithDebInfo --parallel 4
  • cd sdk_v2/js; npm run build:ts -- --pretty false
  • cd sdk_v2/python; python src/foundry_local_sdk/_native/build_cffi.py; python -m pytest test/unit -q

@vercel

vercel Bot commented Jul 24, 2026

Copy link
Copy Markdown

The latest updates on your projects. Learn more about Vercel for GitHub.

Project Deployment Actions Updated (UTC)
foundry-local Ready Ready Preview, Comment Jul 26, 2026 8:04pm

Request Review

Package required ORT, GenAI, WinML, and DirectML runtime companions consistently across C++, JS, and Python packaging paths. Avoid logging caller-provided package URLs and make CMake package extraction work with older CMake versions.

Files changed:

- pipeline packaging/build templates

- sdk_v2/cpp/cmake/FindOnnxRuntime*.cmake and FindWinMLEpCatalog.cmake

- sdk_v2/cpp/nuget/pack.py

- sdk_v2/js/package metadata and prebuild scripts

- sdk_v2/python native lib loader

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Copilot-Session: 297bc539-5f7f-4a70-9411-48e91a5bf532
@bmehta001
bmehta001 force-pushed the bhamehta/flcore/runtime-packaging-hardening branch from 443182c to 01908f1 Compare July 24, 2026 22:26
The packaging pipeline requires DirectML.dll alongside Microsoft.Windows.AI.MachineLearning.dll, but CMake only copied the WinML library. Copy DirectML from the same architecture-specific runtime directory during post-build staging.

Files changed:

- sdk_v2/cpp/CMakeLists.txt

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Copilot-Session: 297bc539-5f7f-4a70-9411-48e91a5bf532
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant