Release candidate: a build of main for testing, not a full release. Its Windows files are unsigned and report version 6.1.0. In-app updates don't offer it; install it with the setup by hand.
Built from 16f3e87 on main.
Changes since v6.1.0
- ci(release): CI mirrors Nightjar and gives both apps one RC version per commit (475814b)
- docs(perf): fix the A/B runs path in the AMF output entry (dc43e7a)
- docs(amf): the output thread costs about 1 to 1.5 % of a core since it parks (07f4bb9)
- feat(release): publish a release candidate from every green main run (5b360e7)
- fix(amf): the output thread stops querying once the last frame is taken (5a299f9)
- perf(amf): collect encoder output on its own thread by default (e318581)
- diag(windows): take the settled window look after the stream's display is gone (#129) (8e3ea94)
- diag(windows): log the end's final displays, scale and HDR commits and a settled window look (#129) (1467e11)
- diag(windows): log the displays and moved windows while an end restores the layout (#129) (ffb4a2b)
- feat(input): Automatic gives on-screen Steam controls a Steam Controller in Steam apps (8762cf5)
- fix(nightjar): macOS fullscreen covers the notch by leaving Spaces (4b291f0)
- fix(library): list a Battle.net game once and look it up by uid first (236ea91)
- fix(library): find Battle.net products by product code, not only the uid (2eb8ef3)
- fix(nightjar): tighten the macOS FFmpeg patch's wording and application (f1cd0df)
- fix(nightjar): macOS media builds keep CMake build folders outside the source paths (f484555)
- fix(nightjar): macOS VideoToolbox decodes into Metal-compatible buffers (1a6c535)
- fix(nightjar): the Controllers Off notice always shows, also for a controller connected later (9380d24)
- feat(nightjar): a stream says when Controllers is Off with a controller connected (29cc9d4)
- test(nightjar): the held-request tests use a 400 ms window (5d9f893)
- fix(nightjar): a held key frame request goes out after its window without another frame (9315983)
- fix(nightjar): a failing decoder asks for one key frame per window, not one per frame (852a4f8)
- docs(release): the probe adapter fix joins the host notes (73d5933)
- test(capture): the headless adapter choice is tested where the probe makes it (0cedf8a)
- fix(capture): the probe without a display opens the GPU the stream encodes on (1b06616)
- fix(console): the touchpad option names the on-screen Steam controls exception (4b2274e)
- fix(input): on-screen Steam Controller controls never make player 1 a DualSense (ebcc9b0)
- fix(stream): an encoder rebuild teaches the source phase only from a focus picture (70578fe)
- docs(performance): record the focus-hold fix on hardware (ccc21cf)
- test(stream): name the desktop pair of the burst test; say only what was measured (3fa4bbe)
- fix(stream): the source phase judges each capture by its own focus label (f796d01)
- test(check): a stream with the strip behind a window repainting in bursts (b1bf611)
- fix(stream): a window repainting in bursts in front no longer holds the picture behind it (da4b56e)
- test(check): the keyframe-request case runs its strip as the foreground window (90f8658)
- docs(release): the 6.1.0 notes get their Tested paragraph (32f6b53)