Repository navigation
JavaScript bindings: @aptabase/tauri 0.7.0.
Added
- Error reporting (beta), to the error endpoint (
/api/v0/error); reports appear on the Errors page of the dashboard ErrorTrackertrait onApp,AppHandleandWindow:track_error(&err)reports anystd::error::Errorwith its type name and causes, andtrack_error_report(ErrorReport)reports one described field by field (ErrorReport::with_fatalfor severityfatal)trackError(error, { fatal })in the JavaScript bindings, with theaptabase:allow-track-errorpermissionBuilder::with_crash_reporting(): a panic on the main thread is reported as a fatalcrash(the app waits up to three seconds for the report before it exits), one in an async task as ataskExceptionand one on another thread asunhandled; the webview's uncaught errors and unhandled promise rejections are reported too- Reports are sent right away, one per request, with the same retry rules as the other SDKs: network errors, 408, 429 and 5xx are retried, a 403 (monthly error quota exhausted) and other 4xx are dropped; at most 25 wait for delivery
- Tests for event delivery (batches of 25, what's kept for the next flush and what's dropped, the event format, polling) and for App Key validation, run in CI
- README sections on delivery and on the TLS backend features
Fixed
- Use the system proxy settings on macOS and Windows again. Turning off reqwest's default features to add the TLS feature flags (#34) also turned off its
system-proxy,http2andcharsetfeatures; they're back on, and the TLS backend is still chosen withdefault-tls,rustls-tlsornative-tls rust-versionis 1.85.0: reqwest 0.13 and rand 0.10 don't build on 1.81flush_events_blocking(), and the flush after a custom panic hook, panicked with "there is no reactor running" when events were queued and the calling thread was outside Tauri's async runtime (the main thread handlingRunEvent::Exit, as the README shows); the flush now runs on the runtime