obs-websocket: Show session durations of a minute and longer

[rdoupe]

Fixes #1116

What and why

The Session Duration column in Tools > WebSocket Server Settings now keeps counting after the first minute instead of going blank.

• SettingsDialog::FillSessionTable() (plugins/obs-websocket/src/forms/SettingsDialog.cpp line 252) formatted the duration with QTime(0, 0, sessionDuration).toString("hh:mm:ss"). QTime(h, m, s) only takes s from 0 to 59, so from 60 seconds on the QTime was invalid, toString() returned an empty string, and the cell stayed empty for the rest of the session while the client was still connected.
• The fork fix, forms: Show session durations of a minute and longer obs-websocket#15, moves the formatting into FormatSessionDuration() (new header src/forms/SessionDuration.h) and builds hh:mm:ss from the seconds count. The hours count on past 24 (QTime(0, 0).addSecs() would wrap after a day), and a negative count (the clock set back below the connect time) shows 00:00:00. This PR moves the submodule to it.
• Nothing else in the dialog changes; durations under a minute look as before.
• Reported upstream by bryce-seifert in Bug: Session Duration in Serve Settings disappears after 1 minute obsproject/obs-websocket#1208 (March 2024, macOS, still open; gitagogaming saw it too). DanielLester83 found the QTime range limit in CI: Possible bugfix for the Status Duration. Please Test. obsproject/obs-websocket#1179 (closed; that patch did not build, so this is a new fix).

Reproduced in release v0.1.0-beta.20261014 (AppImage, headless: private Xvfb, private HOME, dbus-run-session, --multi), websocket server with a password and one identified client that stays connected; Tools > WebSocket Server Settings opened through the menu with XTest clicks, the session row read from a screenshot:

Build	~30 s after connecting	~70 s after connecting
v0.1.0-beta.20261014	00:00:32	(empty), messages still counting
This branch, C libobs lane	00:00:31	00:01:12
This branch, Rust libobs lane	00:00:32	00:01:12

RED → GREEN

• RED commit: 6f2980e test/cmocka: Add failing test for the websocket Session Duration text. It moves the submodule from master's pin f06b040 to 6f3efa3, the fork PR's first commit, which moves the formatting into FormatSessionDuration() unchanged (still QTime(0, 0, seconds)). test_ws_session_duration builds that header with Qt Core only (no libobs, no display) and fails 4 of 5 cases in both lanes: one_minute_and_more ("" != "00:01:00"), hours, more_than_a_day and clock_set_back. The control under_a_minute passes. Registered on every platform with cmocka_qt_runtime().
• GREEN: 32c7028 moves plugins/obs-websocket from 6f3efa3 to 8bca4ac, the fix (8bca4ac once forms: Show session durations of a minute and longer obs-websocket#15 merges; forward only). 5 of 5 with ENABLE_RUST_LIBOBS OFF and ON. Full OBS build (frontend + obs-websocket) in both lanes. Full ctest -j8 (private Xvfb, system libobs-opengl masked, TMPDIR unset): 267/267 in both lanes. Every PipeWire/PulseAudio test ran; skipped only open_crashes_without_obs_core (both lanes) and test_startup_shutdown_cycles (C lane), the same skips earlier master runs here showed. A first full run under heavy load from other jobs on the machine had timing failures in test_pulse_stream_restart, test_pulse_monitor_reset_shared_failed_context_reset_recovers, test_pipewire_screencast_shm and test_ws_transform_bounds (none touched here); the rerun at normal load passed them all. clang-format 22.1.3, gersemi and ci/test_registration.py clean.

Checklist

• Base is master
• I claimed the issue, or no one else has an open PR for it
• Characterization test (if the unit had none), then RED, then GREEN commits
• (Optional) Local gate passed and the pushed head is stamped; otherwise PR CI runs the full suite
• Warning or flag changes (-W..., ratchets): none
• Intentional differences from C listed in the parity test header and rust/README.md (ports): not a port

AI note: Claude Code (Claude Opus 5.5) wrote the code, the test, the commit messages and this text under the submitter's direction, and ran the tests above on the submitter's machines.

🤖 Generated with Claude Code

[zackees]

Submodule check: this PR moves plugins/obs-websocket from f06b040 to 8bca4ac (#15 only). #1146 moves it to d79d2c2 (#15 + #16). If #1146 merges first, this PR then moves the pointer back to 8bca4ac and silently removes the screenshot fix. We had this happen before with #492.

Please bump this one to d79d2c2 as well, so both PRs agree and the order doesn't matter. #16 only touches the screenshot code, so it doesn't affect test_ws_session_duration. Until then I'll queue this PR before #1146.

[zackees]

Part one of the speed up refactor has gone in. C code is amalgamated and relies on one single PCH

Part two, going in now, is rust optimization:

• unit test amalgamation
• artifact caching

You are going to have to rebase, but it's worth it!

[rdoupe]

Bumped: the GREEN commit (now 5f9d492) moves plugins/obs-websocket to d79d2c2, the same commit #1146 uses. Whichever of the two merges first, the other no longer moves the pointer back. d79d2c2 adds only the #1141 screenshot change on top of 8bca4ac. I'll rebase onto master once part two of the refactor is in.