SensaHelp

Verification records

简体中文 | English

Results below belong to the named commits. Open current CI, Pages and CodeQL for later source. How checks run · Browser test workflow

Migration capture and image review: 197af206

Candidate run 40996282389 for source 197af206 executed 218/218 cases; independent Firefox/WebKit compatibility executed 8/8, with zero failures, skips, retries or unexecuted cases. Candidate mode generates images and records capture behavior; it is not ordinary screenshot-comparison evidence.

The per-image record covers 164 application baselines and eight site originals: 20 opened new application images and eight delegated-review originals were accepted, 142 images reuse identical-hash prior observations, and two OSC Hub images retain their reviewed source versions. The Hub endpoint/send-status display regression found during review was not accepted. All 38 documentation images match the selected baseline bytes and identify their actual capture source/run.

Capture and visual review establish provenance and image selection. For the current pinned source, matching CI and Pages summaries/artifacts identify actual ordinary comparison, native CLI/Electron, preview and deployment outcomes. The tables below retain each historical snapshot's verified scope.

Accepted snapshot: b225ee3a

Source b225ee3a7074a3a1e6a6c81d5e2bf8da824f10de, CI #213, attempt 1:

Check Actual result
Web / .NET / Python 252 / 1207 / 401 passed
Full ordinary browser comparison 176/176 passed, 88 per viewport, zero retries, failures, skipped or unexecuted cases
Pages 20/20 passed and deployed; 98 routes, 154 assets and 55 selected online files matched
Native packages Windows, Linux and macOS, each x64/ARM64: isolated TCode smoke passed
Windows formats Portable, single file, installed app, reinstall/uninstall and user-data retention passed on both architectures
WinRT Both architectures imported 8 namespaces and passed native-buffer byte roundtrip; no adapter/GATT operation
Publication All 12 preview #213 downloads matched official SHA-256 and SHA256SUMS; 54 bridge-tool source copies and 169 Unity files matched the build plan
CodeQL Four languages uploaded and processed; all 249 tracked C# files covered, two quality metrics 100%, none-mode 0

That historical browser run checked every selected case and compared the reviewed baselines with updates disabled. Its eight site originals and eight rendered-diagram originals were opened individually. The deployed diagrams were also tested for keyboard scrolling, source fallback and rejected active content.

Coverage of the automated suites

Suite What it checks
Unit / contracts Injected clocks, real loopback HTTP/WebSocket, isolated PTY streams and mock serial/BLE/WinRT
Protocol inputs Schema negotiation, independent source sessions, atomic channel batches, durations, task lifetime and stop order
Browser workflows Configuration, control, recording, script playback, shared tabs and simulated device transport
Site Bilingual navigation, direct downloads, offline assets, responsive layout, gallery and diagram behavior
Native package smoke Actual platform binaries with temporary configurations and simulated network targets
Source analysis Real application/Unity-reference compilation, source coverage and scanner evidence

The Contact cases include contact loss and frozen activity, with a fixed 0.210 m/s screenshot sample followed by ten distinct velocities while Sample remains frozen. Device and protocol regressions include large/fragmented Intiface messages, script loop selection, nonzero script starts, finite task timing and Python bridge roundtrips.

Images and review

The per-image review record identifies the current 164 application baselines and eight site originals, with hashes, observations and reuse chains. Ordinary comparison uses the selected application baselines. The 38 documentation images retain original bytes and capture provenance. Visual changes pass candidate capture and image review before comparison for their matching source. Review process

Environment-specific checks

Actual websites, vendor cloud services, physical OSR/BLE devices, Unity clients and VRChat/HMD operation are checked in their corresponding environments. The latest recorded Unity Editor 2022.3.22f1 attempt stopped at license validation with zero tests executed; 17 Tracking Frame source cases were not run and no NUnit XML was produced. Compilation and package structure results remain useful for their own scope.

Scanner execution and SARIF results are recorded separately from GitHub alert dispositions. In the accepted b225 snapshot, four uploads completed, but no remote alerts were closed or risks accepted in that round; current panel counts and alert state require their own access and review.

Earlier accepted snapshot

0bf1a8e9 passed 144/144 ordinary browser cases, six-platform checks and preview #210 publication. Its Pages run passed 14/14 site checks. These historical runs remain available for comparing later changes.