[ric-oliv]

On macOS, the MAUI cases in cli.Tests.ps1 are the longest part of the integration tests. They build three Release apps and check what sentry-cli uploads for each:

Case	#5788	#5793
Android build	186 s	244 s
iOS build	197 s	253 s
Mac Catalyst build	85 s	81 s
Whole integration test step	13.4 min	14.6 min

With the other open CI PRs applied, the macOS leg is the slowest leg in the build workflow at 32 to 37 minutes, so those 8 to 10 minutes are on the critical path.

This PR:

• moves the MAUI (<framework>) block unchanged from cli.Tests.ps1 to a new cli-maui.Tests.ps1, with the same header,
• skips that block when SKIP_MAUI_TESTS is true, which the build legs set only on macOS,
• adds a MAUI integration test (macos) job that packs the six packages the block registers and runs cli-maui.Tests.ps1 with the same github-workflows action.

The Windows and Linux legs keep running the Android case, because it covers host-specific parts like the sentry-cli binary and Sentry.targets' Exec calls.

The new job tests packages it builds itself, not the ones the macOS leg publishes. For these cases that's equivalent. The macOS leg builds with CI_PUBLISHING_BUILD=true, which only adds the other desktop platforms' native libraries and sentry-cli binaries to the Sentry package. MAUI apps don't use the native libraries, and a macOS arm64 build always includes the Darwin arm64 sentry-cli.

Checked locally:

• Pester discovery finds the same 39 tests, 36 in cli.Tests.ps1 and 3 in cli-maui.Tests.ps1, and skips the 3 MAUI cases when SKIP_MAUI_TESTS=true.
• The moved block is identical to the original, line for line.
• actionlint reports the same findings as on main.

On release branches the new job builds the Cocoa SDK from source, like the macOS leg does.

The first run's MAUI job failed in pack with NU5026. Packing Sentry.Maui while building it on macOS looks for its Windows target's .pri file, which macOS doesn't generate. The job now builds first and packs with --no-build, as the build legs do.

Measured on the second run, against 3 runs from the same evening without this PR:

	Without this PR	This PR
macOS leg	37.1 to 51.1 min	24.1
Its integration test step	12.3 to 15.5	3.7
MAUI integration test (macos) job		15.3
Android, iOS and Mac Catalyst cases	7.8 to 9.6 in the macOS leg	7.0 in the new job
Pester totals on the macOS leg	69 passed, 34 skipped	66 passed, 37 skipped
Pester totals in the new job		3 passed, 0 skipped

The 51.1 was a run whose Build step took 23.6 minutes, against 14.5 and 17.0 in the other two.

So the 3 MAUI cases moved from the macOS leg to the new job, and every other case still runs in the leg. The new job builds for 4.0 minutes, packs in 0.2 and finishes well before the macOS leg.

#skip-changelog

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 76.13%. Comparing base (b41b96c) to head (9e33f11).

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #5797   +/-   ##
=======================================
  Coverage   76.13%   76.13%           
=======================================
  Files         528      528           
  Lines       19579    19579           
  Branches     3829     3829           
=======================================
  Hits        14906    14906           
  Misses       3791     3791           
  Partials      882      882           

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.