Commit Graph
125 Commits
Author SHA1 Message Date
saphid d6f3d7c380 Set Windows DPI coordinates before native plugin initialization 2026-09-28 23:10:00 +10:00
saphid 838eaa4e6b Cover the available PipeWire converter and its D-Bus support 2026-09-28 23:09:04 +10:00
saphid 6104c3f192 Refuse window input when Windows cannot focus its target 2026-09-28 23:07:38 +10:00
saphid da5c954ce4 Retain the final idle-window update while the stream gate is closed 2026-09-28 23:05:23 +10:00
saphid 2466c5aa14 Include and smoke-test PipeWire video adapter factories 2026-09-28 23:00:13 +10:00
saphid 060cf9cd57 Bundle PipeWire client modules and record the real Frame probe 2026-09-28 22:58:41 +10:00
saphid cd202433bb Verify shared adaptation and harden PC stream lifecycle 2026-09-28 22:50:58 +10:00
saphid c5fc552c60 Add native PC capture adapters behind the shared panel viewer 2026-09-28 22:29:34 +10:00
saphid 559caa2dd0 Share the desktop stream rate controller across host platforms 2026-09-28 22:23:10 +10:00
saphidandClaude Opus 5.5 1b90c64b73 Docs: benchmark with the headset worn (Wi-Fi much rougher; controller bounded latency)
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-09-28 22:02:05 +10:00
saphidandClaude Opus 5.5 39afb23d97 Docs: Steam's StartDesktopStream pairs the Frame with the Mac (verified); stream test next
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-09-28 21:45:52 +10:00
saphidandClaude Opus 5.5 bda9d77d2d Mac in the headset: USB route keeps a configured host-key alias and falls back to the network
Review round 12: a failed USB-C tunnel now retries the normal path; an
existing HostKeyAlias wins; --host with --usb is rejected; the Steam
desktop-streaming claim is now 'untested' (Valve documents the desktop
showing when a game loses focus); the Show/cleanup overlap test blocks
for real (it fails without the lock). Verified live: with the cable out,
the route is the normal path.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-09-28 21:39:40 +10:00
saphidandClaude Opus 5.5 4fae62ba14 Mac in the headset: use the Frame's USB-C network when plugged in; Steam streaming findings
- Plugged into the Mac, the Frame is a USB network device ("Steam Frame",
  usb0 at ~0.9 ms). The tunnel uses it when the Frame's usb0 answers,
  with the usual host key; otherwise the normal path. Interleaved runs:
  content p50 7 vs 10 ms, click to drawn 17 vs 27 ms, scroll p95 23 vs
  31-37 ms. FRAME_MACVIEW_USB=0 turns it off; the card says "over USB-C".
- Bench: --usb, and the route is recorded per run.
- Docs: Steam's own streaming (no SteamVR host on macOS; Remote Play pairs
  but streams games, not windows; test blocked); the Frame's USB network;
  the 2026-09-28 health-check boot-loop recurrence.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-09-28 21:33:31 +10:00
saphidandClaude Opus 5.5 1d05f57fbf Mac in the headset: a Show waits for a viewer cleanup already in progress
Review round 11: the cleanup's check and pkill now hold a lock that Show
takes to count itself in, so a new viewer can't start between them.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-09-28 21:14:27 +10:00
saphidandClaude Opus 5.5 484a50a189 Mac in the headset: no viewer cleanup while a Show is launching; relay pump always ends
Review round 10: a Show replacing its own stream could have its new viewer
ended by the cleanup; a viewer reset left the delayed relay pending.
Verified: the relay delivers what's queued, then closes the agent side.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-09-28 21:11:26 +10:00
saphidandClaude Opus 5.5 7f769ca2f0 Mac in the headset: end the Frame's viewer browser after Stop; relay fixes
- Chromium on the Frame outlived its last viewer window (verified: 11
  processes left after a run). Once nothing is shown, Stop ends it, unless
  Show was pressed again meanwhile; its profile is Frame Control's own.
- Relay --delay: if the agent side fails, close the viewer side too
  (review round 9).
- Docs: the final scroll run captured 57 fps; don't blame ScreenCaptureKit.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-09-28 21:07:59 +10:00
saphidandClaude Opus 5.5 71200c5179 Merge origin/main into mac-in-headset
Brings in #4, #5, #8 (fbl100's verified Remmina/VNC mirror), APK
alternatives and the website. docs/streaming.md: Mac in the headset stays
the recommendation (now verified on the Frame); Remmina keeps #8's verified
evidence as the whole-screen fallback. docs/mac-in-headset.md cites #8's
lag finding.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-09-28 21:03:34 +10:00
saphidandClaude Opus 5.5 00b45bafc5 Mac in the headset: final benchmark numbers; relay delay is a delay line
Review round 8 (GPT-6 Astra xhigh): --delay paused upstream reads, so busy
streams saw 30-60 ms instead of 30. Verified locally: 60-62 ms round trip
with 30 ms each way under load. No saved result used --delay.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-09-28 21:02:26 +10:00
saphidandClaude Opus 5.5 9e4dcdd147 Mac in the headset: measure every frame, adapt to the network, separate windows
- Per-frame timing on the Mac's clock (capture, encode, network, decode,
  draw), viewer clock sync and reports, input echo, /stats and a HUD.
- scripts/macview-bench.py: repeatable runs on the real Frame, a shaping
  relay (no sudo), interleaved A/B between agent settings; results in
  bench/results/.
- Adaptive controller: ack-based send gate with jitter-aware slack, AIMD
  bitrate that knows when a stream is app-limited, fps then size tiers.
  On a 50->3->50 Mbit/s step, scroll p95 went from 4.7 s to 72 ms; no cost
  on a clean link.
- Separate mode: real AppKit event loop (HiDPI and NSScreen now work),
  cropped capture for fixed-size windows, windows kept on their display,
  graceful quit restores windows; stop/start races fixed.
- Encoder timeline clamp (no oversized frame after a pause).
- Frame Control shows each live stream's fps, delay, bitrate and tier.

Reviewed by GPT-6 Astra xhigh (read-only), 7 rounds; findings fixed.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-09-28 20:56:14 +10:00
Alex Southwell dcf9689f64 Merge pull request #11 from saphid/apk-alternatives
Offer older APK versions that fit when Lepton refuses an app
kdeconnect-frame-24.02.2-1
2026-09-28 17:38:35 +10:00
saphidandClaude Opus 5.5 224340edc9 APK alternatives: refresh the catalogue after an install job; pin the test's premise
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-09-28 17:29:29 +10:00
saphidandClaude Opus 5.5 c814cb95d0 Merge main into apk-alternatives: install other versions as background jobs
Main now runs Android installs as background jobs and maps SSH failures to
one offline message. Alternative-version installs go through the same job,
the alternatives dialog waits on it with runJob, and send_error_json keeps
the apk blocker that opens the dialog.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-09-28 17:21:55 +10:00
Alex Southwell ff2c4ebfe0 Merge pull request #8 from fbl100/mac-mirror-verified
Mac → Frame mirror: verified, with fixes for scaling, login and the cursor
2026-09-28 17:20:22 +10:00
Alex Southwell 03322d3166 Merge pull request #5 from saphid/iphone-app
iPhone and iPad app: the same features, served from the Frame
2026-09-28 17:20:18 +10:00
Alex Southwell 2d08486278 Merge pull request #4 from saphid/ui-tabs-reliability
Tabs, one offline banner, background installs, drop anywhere
2026-09-28 17:19:46 +10:00
Alex Southwell f780ab2c6a Merge pull request #14 from saphid/site-polish
Website: crop the Tools screenshot and tighten the phone footer
2026-09-28 16:43:21 +10:00
saphidandClaude Opus 5.5 396f2a830a Website: crop the empty half off the Tools screenshot; tighten wrapped footer links
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-09-28 16:39:01 +10:00
Alex Southwell 59761ae015 Merge pull request #12 from saphid/website-kofi
Website: Ko-fi donate buttons and visual fixes
2026-09-28 16:34:18 +10:00
saphidandClaude Opus 5.5 48eedbc2eb Website: let the hero platform list wrap on phones so large text isn't clipped
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-09-28 16:31:39 +10:00
saphidandClaude Opus 5.5 2b89eeba1d Website: fix visual bugs found in a page-by-page review
- Gradient buttons showed a dark sliver on the right: the background shorthand reset
  background-origin, so the gradient repeated under the transparent border.
- Phone header: keep the brand on one line and drop the GitHub button under 480px.
- Hero pill: put the platform list on its own line on phones instead of orphaning one item.
- Privacy page: plain sections instead of open accordions whose x looked like a close button.
- Same header (GitHub button) and footer links on every page; legend spacing on the form.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-09-28 16:26:43 +10:00
saphidandClaude Opus 5.5 75e92db2fa Website: point the donate buttons at ko-fi.com/alexsouthwell; add FUNDING.yml
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-09-28 16:18:12 +10:00
Alex Southwell 10f96656e3 Merge pull request #10 from saphid/website
Website, feedback form that opens issues, and pi's contributor gate
2026-09-28 16:07:23 +10:00
saphidandClaude Opus 5.5 a1fa4ce140 Fix the cross-provider review's findings on APK alternatives
One malformed or unreachable repo no longer hides the others; skip bad
index entries; a refreshed raw index outdates its reduced copy; style the
dialog like the others; validate package ids with PKG_RE.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-09-28 15:15:11 +10:00
saphidandClaude Opus 5.5 a7663b3a5e Website feedback: double backslashes so escapes can't rebuild references; queue every approval
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-09-28 14:51:18 +10:00
saphidandClaude Opus 5.5 a6fff434a4 Website feedback: attribution first, escape <, wait out the fill timer; maintainers only in the approval queue
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-09-28 14:45:54 +10:00
saphidandClaude Opus 5.5 cd40a194ed Website feedback: escape & so entities can't rebuild mentions; queue only lgtm comments
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-09-28 14:39:54 +10:00
saphidandClaude Opus 5.5 0037b1ef4b Website feedback: fixes from review
- Rate limiting fails open on KV errors instead of dropping feedback
- Break owner/repo#1, GH-1 and github.com references in user text
- Time the form with the browser's monotonic clock, not wall-clock
- Serialize lgtm approvals and rebase before pushing

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-09-28 14:33:42 +10:00
saphidandClaude Opus 5.5 50405ccf88 Add IzzyOnDroid to APK alternatives; keep the catalogue's index file
Reducing an index no longer deletes apk-catalog/data/index-v2.json, which
the catalogue build reads. Drop the measurement log from docs.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-09-28 14:30:15 +10:00
saphid 32196b4260 Reduce F-Droid indexes and fetch APK alternatives asynchronously 2026-09-28 14:26:40 +10:00
saphidandClaude Opus 5.5 7d328e20a5 Website with a feedback form that opens GitHub issues, and pi's contributor gate
- site/: landing page, /feedback/ and /privacy/ on Cloudflare Pages
  (frame-control.pages.dev). POST /api/feedback validates the form and opens
  a labelled issue with a fine-grained token; honeypot, minimum fill time and
  KV rate limits keep spam out. Ko-fi donate buttons appear once the page
  name is set in site/public/js/site.js.
- .github: the issue and PR gate from badlogic/pi-mono. New contributors'
  issues and PRs are auto-closed; a maintainer replying lgtmi/lgtm approves
  them via APPROVED_CONTRIBUTORS. Issue templates and CONTRIBUTING.md.
- CI runs the website tests; README points feedback at the form.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-09-28 14:26:33 +10:00
saphid fbe7ba9575 Find installable APK alternatives in F-Droid main and archive 2026-09-28 14:16:43 +10:00
saphidandClaude Opus 5.5 a3c6e5c984 Mac in the headset: stream Mac windows and screens into the Frame as panels
Frame Control can now show any Mac window, or a whole screen, as its own
SteamVR panel on the Steam Frame (Tools -> Mac in the headset, macOS only).
Place it with the SteamVR dashboard; the laser clicks and scrolls, and the
Mac's own keyboard types.

- mac/frame-mac-view (Swift, no dependencies): ScreenCaptureKit capture per
  window or display, VideoToolbox H.264 with low-latency rate control (JPEG
  fallback), a loopback HTTP/WebSocket server, CGEvent/AX input playback,
  and a display-awake assertion while anyone watches.
- ui/frame_macview.py: starts the agent, runs an ssh -R tunnel with a
  supervisor that reopens it on the same port, and launches a Chromium app
  window per stream on gamescope's :0, tagged with STEAM_GAME for its own panel.
- Frame Control's key never leaves the Mac: viewers get single-use,
  per-source tickets and reconnect keys that Stop revokes.
- ui/mac-view.html: WebCodecs decode, keyframe recovery, pointer/wheel/keys back.
- Bundled in the Mac app build; tests/test_macview.py builds and drives the
  agent on macOS.

Verified on the Frame (build 20260925.6191901) with the test pattern: panel
in about 1.5 s, about 60 fps, Mac-to-window about 11-17 ms, tunnel recovery
in 4 s. Laser input and real window capture still need a person in the headset.

Also commits the other thread's first-party rule (steam-frame skill) and
the first-party options table in docs/streaming.md.

Reviewed by GPT-6 Astra (xhigh, read-only) over six rounds; all findings fixed.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-09-28 13:33:46 +10:00
Frank LevineandClaude Opus 5.5 5e12245932 Streaming doc: Mac → Frame mirror verified; move answered open questions
Tested on Frame BUILD_ID 20260925.6191901 with macOS 27.0.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-09-27 13:49:12 -04:00
Frank LevineandClaude Opus 5.5 df1810bae3 Add mac-cursor-ring.lua: show the Mac pointer in the VNC mirror
macOS leaves the pointer out of the Screen Sharing framebuffer, and neither Remmina showcursor setting brings it back. A Hammerspoon ring around the pointer is a real window, so it gets mirrored.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-09-27 13:49:12 -04:00
Frank LevineandClaude Opus 5.5 3f0e7b198a install-apps: scale the Mac screen profile to fit; warn about the Mac login
Without scale-to-fit a Retina Mac shows 1:1 as a zoomed-in corner. macOS offers Apple auth ahead of plain VNC auth, so Remmina asks for the Mac account login.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-09-27 13:49:12 -04:00
saphidandClaude Opus 5.5 9110557e23 README: link the Frame Control trailer
A thumbnail under the screenshot opens the 66-second trailer, which is attached to the 'trailer' release.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
v0.3.1
2026-09-27 22:07:44 +10:00
saphid 0016d9200c Merge remote-tracking branch 'origin/iphone-app' into iphone-app 2026-09-27 21:27:10 +10:00
saphidandClaude Opus 5.5 fe87a9d826 Keep the page clear of iOS safe areas everywhere; research typing, pointing and mirroring into the Frame
- Header, content, tab bar, status strip, drawer and toasts add the notch,
  Dynamic Island, rounded-corner and home-indicator insets on every side
  (zero on desktops). Phones on their side use the bottom tab bar layout.
- The phone tab bar hides while a text field has focus, instead of riding
  on the keyboard.
- DEBUG hook FRAME_TEST_LANDSCAPE for checking this in the Simulator.
- docs/streaming.md: iPhone mirroring (UxPlay, broadcast extension) and
  keyboard/mouse input (uinput needs no sudo on the Frame, verified).

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-09-27 21:26:32 +10:00
Alex Southwell 1b26c4f92c Merge pull request #7 from saphid/fake-frame-tests
Regression tests against a fake Frame, plus a headset smoke test
2026-09-27 21:25:36 +10:00
saphidandClaude Opus 5.5 3db311da13 iPhone app: declare photo saving so Save Image appears; record what was tested
The share sheet only offers Save Image when the app declares
NSPhotoLibraryAddUsageDescription; without it screenshots couldn't be saved to
Photos. docs/iphone.md now lists what was verified against the Frame (Bonjour
discovery, waiting and reconnecting, upload, share sheet, install links,
opening Steam Link) and the two permission prompts iOS shows.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-09-27 18:49:25 +10:00