feat(macOS): libdispalydevice integration by martona · Pull Request #5338 · LizardByte/Sunshine

martona · 2026-06-26T03:22:52Z

Description

This is a follow-up from last week's libdisplaydevice work: macOS integration in Sunshine.

macOS now uses libdisplaydevice for display enumeration as well as applying/reverting changes.

I also removed the patch that kept the display awake; Sunshine now relies on the new libdisplaydevice functionality to do the same. Note that while libdisplaydevice gained the ability to wake displays on Windows, this PR does not take advantage of it; the scope was kept to macOS.

I had to bump the libdisplaydevice submodule.

Screenshot

Issues Fixed or Closed

Roadmap Issues

Type of Change

Checklist

Code follows the style guidelines of this project
Code has been self-reviewed
Code has been commented, particularly in hard-to-understand areas
Code docstring/documentation-blocks for new or existing methods/components have been added or updated
Unit tests have been added or updated for any new or modified functionality

AI Usage

None: No AI tools were used in creating this PR
Light: AI provided minor assistance (formatting, simple suggestions)
Moderate: AI helped with code generation or debugging specific parts
Heavy: AI generated most or all of the code changes

ReenigneArcher · 2026-06-26T03:28:03Z

Apologies, but I merged a change a little while ago that causes conflicts with this one.

martona · 2026-06-26T03:42:53Z

No worries, should merge cleanly now.

Co-authored-by: Dave Lane <42013603+ReenigneArcher@users.noreply.github.com>

sonarqubecloud · 2026-06-27T00:18:55Z

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

codecov · 2026-06-27T00:27:03Z

Bundle Report

Bundle size has no change ✅

ReenigneArcher · 2026-06-27T00:24:58Z

+#endif
+    }
+
+    std::unique_ptr<DisplayPowerInterface> make_display_power() {


Wasn't this added for Windows as well?

https://github.com/LizardByte/libdisplaydevice/blob/65616076ba881046085438c80fddead9beedf74f/src/windows/display_power.cpp#L43

Also, found this which should maybe replaced with the new library code?

Sunshine/src/platform/windows/display_base.cpp

Lines 241 to 248 in 9f645a9

// Keep the display awake during capture. If the display goes to sleep during

// capture, best case is that capture stops until it powers back on. However,

// worst case it will trigger us to reinit DD, waking the display back up in

// a neverending cycle of waking and sleeping the display of an idle machine.

SetThreadExecutionState(ES_CONTINUOUS | ES_DISPLAY_REQUIRED);

auto clear_display_required = util::fail_guard([]() {

SetThreadExecutionState(ES_CONTINUOUS);

});

Yeah that's what I meant in part when I said I could make a power management pass after the interface-factories. Technically it could have been a part of this PR, but I did not want it out of control, touching multiple OSes. A note regarding it is also in the body text of the PR: it's definitely not forgotten about, it just felt out of scope.

Obviously, I have a reading problem today. Thanks for the clarification!

Maybe just throw a couple TODO comments in? (Sonar will complain about the TODOs, but I'll bypass it)

codecov · 2026-06-27T01:22:56Z

❌ 2 Tests Failed:

Tests completed	Failed	Passed	Skipped
346	2	344	4

View the top 2 failed test(s) by shortest run time

DisplayDeviceConfigTest/ParseHdrOption::IntegrationTest/2

Stack Traces | 0s run time

.../tests/unit/test_display_device.cpp:122
Expected equality of these values:
  std::get<display_device::SingleDisplayConfiguration>(result).m_hdr_state
    Which is: (nullopt)
  expected_value
    Which is: (4-byte object <01-00 00-00>)

DisplayDeviceConfigTest/ParseHdrOption::IntegrationTest/3

Stack Traces | 0s run time

.../tests/unit/test_display_device.cpp:122
Expected equality of these values:
  std::get<display_device::SingleDisplayConfiguration>(result).m_hdr_state
    Which is: (nullopt)
  expected_value
    Which is: (4-byte object <00-00 00-00>)

To view more test analytics, go to the Test Analytics Dashboard
_{📋 Got 3 mins? Take this short survey to help us improve Test Analytics.}

martona · 2026-06-27T01:30:24Z

Everything going red is on libdisplaydevice. EnumeratedDevice is an aggregate and does not define a constructor, so calls like:

devices.emplace_back(device_id, display_name, friendly_name, edid, info);

rely on C++20 parenthesized aggregate initialization through std::construct_at. Newer toolchains accept that, which is why it built for me locally, but the Xcode 15.4 CI toolchain rejects it here.

The cleanest fix is to go back to libdisplaydevice and undo the SONAR fix that caused this by changing the affected EnumeratedDevice insertions back to explicit braced aggregate construction, e.g. push_back(EnumeratedDevice { ... }). I'll submit a PR for that, then update the Sunshine submodule pin.

Before I do that though... does this sound like the best course of action to you too?

ReenigneArcher · 2026-06-27T02:10:42Z

Everything going red is on libdisplaydevice. EnumeratedDevice is an aggregate and does not define a constructor, so calls like:

devices.emplace_back(device_id, display_name, friendly_name, edid, info);

rely on C++20 parenthesized aggregate initialization through std::construct_at. Newer toolchains accept that, which is why it built for me locally, but the Xcode 15.4 CI toolchain rejects it here.

The cleanest fix is to go back to libdisplaydevice and undo the SONAR fix that caused this by changing the affected EnumeratedDevice insertions back to explicit braced aggregate construction, e.g. push_back(EnumeratedDevice { ... }). I'll submit a PR for that, then update the Sunshine submodule pin.

Before I do that though... does this sound like the best course of action to you too?

That sounds fine, maybe we should add an older macOS runner to the libdisplaydevice build/test matrix as well.

I think there's a few other things going on as well.

Unused function on Linux... you can apply [[maybe_unused]]?

[34](https://github.com/LizardByte/Sunshine/actions/runs/28272130054/job/83772706710?pr=5338#step:6:2835)
Error: /home/runner/work/Sunshine/Sunshine/src/display_device.cpp:144:10: error: ‘bool display_device::{anonymous}::is_unsigned_integer(std::string_view)’ defined but not used [-Werror=unused-function]
  144 |     bool is_unsigned_integer(std::string_view value) {
      |          ^~~~~~~~~~~~~~~~~~~

Two tests failing on macos-latest runners:

[ RUN      ] DisplayDeviceConfigTest/ParseHdrOption.IntegrationTest/2
/Users/runner/work/Sunshine/Sunshine/tests/unit/test_display_device.cpp:122: Failure
Expected equality of these values:
  std::get<display_device::SingleDisplayConfiguration>(result).m_hdr_state
    Which is: (nullopt)
  expected_value
    Which is: (4-byte object <01-00 00-00>)

[2026-06-27 01:07:15.242]: Tests: From /Users/runner/work/Sunshine/Sunshine/tests/unit/test_display_device.cpp:110
[2026-06-27 01:07:15.242]: Tests:   DisplayDeviceConfigTest/ParseHdrOption/IntegrationTest/2 started
[2026-06-27 01:07:15.242]: Info: Ignoring HDR display device request on macOS because macOS HDR changes are not supported by libdisplaydevice.
[2026-06-27 01:07:15.242]: Tests: At /Users/runner/work/Sunshine/Sunshine/tests/unit/test_display_device.cpp:122
[2026-06-27 01:07:15.242]: Tests:   Non-fatal failure: Expected equality of these values:
  std::get<display_device::SingleDisplayConfiguration>(result).m_hdr_state
    Which is: (nullopt)
  expected_value
    Which is: (4-byte object <01-00 00-00>)
[2026-06-27 01:07:15.242]: Tests: DisplayDeviceConfigTest/ParseHdrOption/IntegrationTest/2 failed
[  FAILED  ] DisplayDeviceConfigTest/ParseHdrOption.IntegrationTest/2, where GetParam() = ((4-byte object <01-00 00-00>, true), (4-byte object <01-00 00-00>)) (0 ms)
[ RUN      ] DisplayDeviceConfigTest/ParseHdrOption.IntegrationTest/3
/Users/runner/work/Sunshine/Sunshine/tests/unit/test_display_device.cpp:122: Failure
Expected equality of these values:
  std::get<display_device::SingleDisplayConfiguration>(result).m_hdr_state
    Which is: (nullopt)
  expected_value
    Which is: (4-byte object <00-00 00-00>)

[2026-06-27 01:07:15.242]: Tests: From /Users/runner/work/Sunshine/Sunshine/tests/unit/test_display_device.cpp:110
[2026-06-27 01:07:15.242]: Tests:   DisplayDeviceConfigTest/ParseHdrOption/IntegrationTest/3 started
[2026-06-27 01:07:15.242]: Info: Ignoring HDR display device request on macOS because macOS HDR changes are not supported by libdisplaydevice.
[2026-06-27 01:07:15.242]: Tests: At /Users/runner/work/Sunshine/Sunshine/tests/unit/test_display_device.cpp:122
[2026-06-27 01:07:15.242]: Tests:   Non-fatal failure: Expected equality of these values:
  std::get<display_device::SingleDisplayConfiguration>(result).m_hdr_state
    Which is: (nullopt)
  expected_value
    Which is: (4-byte object <00-00 00-00>)
[2026-06-27 01:07:15.242]: Tests: DisplayDeviceConfigTest/ParseHdrOption/IntegrationTest/3 failed
[  FAILED  ] DisplayDeviceConfigTest/ParseHdrOption.IntegrationTest/3, where GetParam() = ((4-byte object <01-00 00-00>, false), (4-byte object <00-00 00-00>)) (0 ms)

I didn't check every failure though, this was only a spot check.

martona force-pushed the integrate-macos-libdisplaydevice branch from ede0ab1 to b733db8 Compare June 26, 2026 03:39

ReenigneArcher reviewed Jun 26, 2026

View reviewed changes

Comment thread src/display_device.cpp Outdated

Comment thread src/display_device.cpp

martona and others added 3 commits June 26, 2026 20:04

libdisplaydevice macos integration

ef5ddb8

address sonar concern

95efad6

Update src/display_device.cpp

5700431

Co-authored-by: Dave Lane <42013603+ReenigneArcher@users.noreply.github.com>

ReenigneArcher force-pushed the integrate-macos-libdisplaydevice branch from 6c46fdc to 5700431 Compare June 27, 2026 00:04

ReenigneArcher reviewed Jun 27, 2026

View reviewed changes

	// Keep the display awake during capture. If the display goes to sleep during
	// capture, best case is that capture stops until it powers back on. However,
	// worst case it will trigger us to reinit DD, waking the display back up in
	// a neverending cycle of waking and sleeping the display of an idle machine.
	SetThreadExecutionState(ES_CONTINUOUS \| ES_DISPLAY_REQUIRED);
	auto clear_display_required = util::fail_guard([]() {
	SetThreadExecutionState(ES_CONTINUOUS);
	});

Uh oh!

Uh oh!

Conversation

martona commented Jun 26, 2026

Description

Screenshot

Issues Fixed or Closed

Roadmap Issues

Type of Change

Checklist

AI Usage

Uh oh!

ReenigneArcher commented Jun 26, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

martona commented Jun 26, 2026

Uh oh!

Uh oh!

Uh oh!

sonarqubecloud Bot commented Jun 27, 2026

Quality Gate passed

Uh oh!

codecov Bot commented Jun 27, 2026

Bundle Report

Uh oh!

ReenigneArcher Jun 27, 2026

Choose a reason for hiding this comment

Uh oh!

martona Jun 27, 2026

Choose a reason for hiding this comment

Uh oh!

ReenigneArcher Jun 27, 2026

Choose a reason for hiding this comment

Uh oh!

ReenigneArcher Jun 27, 2026

Choose a reason for hiding this comment

Uh oh!

codecov Bot commented Jun 27, 2026

❌ 2 Tests Failed:

Uh oh!

martona commented Jun 27, 2026

Uh oh!

ReenigneArcher commented Jun 27, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

ReenigneArcher commented Jun 26, 2026 •

edited

Loading

ReenigneArcher commented Jun 27, 2026 •

edited

Loading