ci(device_info_plus): migrate android workflows to ubuntu-24.04 and kvm

[mdmahendri]

Description

This PR migrates the Android workflows for device_info_plus from macos-15-intel to ubuntu-24.04 environment.

The Problem

Currently, the repo utilize macOS runner which make android build and android integration tests running on API level 36 consistently time out and fail across the repository's packages. This PR serves as one part of a plan to migrate all nine packages to ubuntu-24.04 to resolve this issue.

Consideration

• This aligns with the android-emulator-runner README.md, which references the official GitHub Blog regarding hardware-accelerated Android virtualization on Linux:

  "...Testing on a 4-core machine with hardware acceleration is around 2-3 times faster than not using hardware acceleration and around 2 times faster than using MacOS."

• The community has reported substantial performance improvements on Linux runners compared to macOS instances (see the Kiwix Android migration PR and GitHub Runner Images Issue #1336.

Benchmark & Validation Data on android_example_build

The table below presents the total runtime of the workflow job for build. Control baseline were retrieved from the repo historical runtime logs, compared against two validation runs on the Linux environment on my other branch:

No	Package Name	Control (macos-15-intel)	Test 1 (ubuntu-24.04)	Test 2 (ubuntu-24.04)
1	android_alarm_manager_plus	10m 2s	5m 31s	5m 23s
2	android_intent_plus	10m 57s	4m 30s	4m 19s
3	battery_plus	13m 1s	4m 46s	4m 33s
4	connectivity_plus	17m 21s	4m 24s	4m 24s
5	device_info_plus	12m 1s	4m 53s	4m 40s
6	network_info_plus	12m 2s	4m 48s	4m 55s
7	package_info_plus	8m 14s	4m 41s	4m 41s
8	sensors_plus	6m 6s	4m 25s	4m 41s
9	share_plus	12m 6s	5m 21s	5m 22s

Benchmark on integration test targeting API Level 36

Because the current macOS CI baseline consistently times out (30m+) or fails to boot the API 36 image entirely, no repo historical runtime logs are retrieved. On ubuntu-24.04 with KVM enabled, API 36 consistently runs to completion. This represents a ~70% reduction in integration test times where it previously fail.

No	Package Name	Job Link (Test 1 Verification)	Test 1 Duration	Test 2 Duration
1	android_alarm_manager_plus	Job Log	8m 30s	8m 18s
2	android_intent_plus	Job Log	3m 4s	3m 15s
3	battery_plus	Job Log	7m 28s	7m 13s
4	connectivity_plus	Job Log	7m 24s	7m 24s
5	device_info_plus	Job Log	7m 6s	7m 27s
6	network_info_plus	Job Log	7m 43s	7m 23s
7	package_info_plus	Job Log	7m 23s	7m 33s
8	sensors_plus	Job Log	7m 25s	7m 8s
9	share_plus	Job Log	8m 0s	8m 11s

Related Issues

No tracking issue is open, but ongoing pull requests show persistent fails and timeouts during the AVD stage. This creates confusion for contributors diagnosing whether a test failure stems from their code changes or the infrastructure itself. Fixing this workflow provide better tool for all future contributions.

Checklist

• I read the Contributor Guide and followed the process outlined there for submitting PRs.
• I titled the PR using Conventional Commits.
• I did not modify the CHANGELOG.md nor the plugin version in pubspec.yaml files.
• All existing and new tests are passing. (CI/Workflow infrastructure update only).
• The analyzer (flutter analyze) does not report any problems on my PR. (CI/Workflow infrastructure update only).

Breaking Change

Does your PR require plugin users to manually update their apps to accommodate your change?

• No, this is not a breaking change.

[mdmahendri]

android_example_build and android_integration_test (36) ran successfully in 4m 52s and 7m 15s respectively. These times are well within the range of the reported results above.