Skip to content

test(cua-driver): verify live Windows executable grants - #3600

Merged
f-trycua merged 1 commit into
mainfrom
codex/3330-windows-grant-regression
Sep 6, 2026
Merged

test(cua-driver): verify live Windows executable grants#3600
f-trycua merged 1 commit into
mainfrom
codex/3330-windows-grant-regression

Conversation

@f-trycua

@f-trycua f-trycua commented Sep 6, 2026

Copy link
Copy Markdown
Collaborator

Issue #3330 reports that Windows bounded application grants cannot match processes without installed-app identity fields. Main already canonicalizes live Windows executable fingerprints (#3299), and shared authorization enriches resources with those fingerprints before grant matching.

Add a Windows-native regression across that boundary: collect the live test process fingerprint, load a v3 executable manifest, and verify both observation and input grants with null bundle_id and launch_path. The exact executable is allowed and a different executable is denied. This is test coverage; no authorization behavior changes.

Refs #3330.

Validation:

  • cargo fmt --all -- --check and git diff --check pass.
  • Windows run 34056147716 passed the complete job and explicitly executed live_executable_grant_matches_without_installed_app_identity. Its PR-merge checkout tree exactly matches candidate d301ae0287986474c2823ded6217da55a371c473.
  • Tested candidate 5996b0667d14a441506909453126885f725d1cfa is rebased onto main after ci(cua-driver): execute complete Windows core tests #3596 with an identical patch. Windows run 34065623076 passed, including all 586 core tests and the live executable-grant regression. All ordinary PR checks passed or were skipped by workflow conditions.
  • Final head b9077d6cf68fede6e6c5905fbf7487b4f257b8f0 rebases over intervening Fleet documentation merges. Compared with the passing candidate, only Fleet documentation and its dedicated tutorial test/workflow differ. The grant patch, complete Driver source/tests, and every Driver CI workflow are identical. Existing Windows behavioral evidence remains applicable; fresh required merge checks are revalidated.
  • The window ID in this test is synthetic. This proves native identity collection plus grant matching, not actual window ownership, capture, input, or bounded enumeration. Server 2025 end-to-end replay remains required before resolving the issue.

no-release accounts for the test-only edit in a release-tracked source file. Release publication is excluded.

@f-trycua f-trycua added the no-release Skip auto-release on merge label Sep 6, 2026
@f-trycua
f-trycua force-pushed the codex/3330-windows-grant-regression branch from d301ae0 to 0045bc9 Compare September 6, 2026 22:37
@f-trycua
f-trycua marked this pull request as ready for review September 6, 2026 22:37
@f-trycua
f-trycua requested a review from injaneity as a code owner September 6, 2026 22:37
@f-trycua
f-trycua force-pushed the codex/3330-windows-grant-regression branch 3 times, most recently from 40b304f to 4484839 Compare September 6, 2026 23:13
@f-trycua
f-trycua force-pushed the codex/3330-windows-grant-regression branch from 4484839 to b9077d6 Compare September 6, 2026 23:15
@f-trycua
f-trycua merged commit d03461a into main Sep 6, 2026
19 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

no-release Skip auto-release on merge

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant