[github-actions]

This PR was opened by the Changesets release GitHub action. When you're ready to do a release, you can merge this and the packages will be published to npm automatically. If you're not ready to do a release yet, that's fine, whenever you add more changesets to main, this PR will be updated.

Releases

@maple-dev/alchemy@0.6.1

Patch Changes

• Updated dependencies [c022d7f]
• Updated dependencies [613df48]
• Updated dependencies [58c2b10]
• Updated dependencies [04fd697]
  • @maple-dev/effect-sdk@0.14.1

@maple-dev/browser@0.14.2

Patch Changes

• c022d7f: Network capture no longer rejects a fetch call that succeeded:

  • fetch accepts anything it can stringify, such as fetch(window.location) or a URL from another frame. Capture read a url property those values do not have, threw after the response had arrived, and the app's await fetch(...) rejected with a TypeError. The URL is now read the way fetch reads it.
  • A request that does not ask for keepalive no longer has its body encoded just to measure it.

• 613df48: Session replay no longer changes how a page's own stylesheet calls fail, and benign browser notices no longer count as session errors:

  • A page whose code calls insertRule (or deleteRule, replace, replaceSync, setProperty, removeProperty) with something the browser rejects gets the browser's exception again. While a recording ran, the call returned undefined instead of throwing and the console showed [maple] session replay recording failed. CSS-in-JS libraries that insert vendor-prefixed rules and rely on catching the error hit this on every page load. The recording itself was not interrupted.
  • The ResizeObserver loop notices and errors thrown by browser extensions are no longer recorded as session errors, so they stop counting toward a session's error total and its "has errors" filter. These are the default error filters that already keep them out of issues; errors: { defaultFilters: false } in @maple-dev/browser keeps them in both.

• 58c2b10: Capture no longer changes what the app sees, and both SDKs agree on one page:

  • Network capture cannot reject a fetch that succeeded. Anything capture throws after the response arrived (a URL filter, a body pattern, reading the body in the background) is dropped, the app gets the real response, and a failed request still rejects with its own error. An XHR's loadend bookkeeping can no longer raise an uncaught error on the page.
  • Session events carry the time, page URL and trace id of the moment they happened. Clicks, inputs, errors, navigations, console and network events were stamped when the batch was sent, up to a minute later: a click that navigated was attributed to the destination page, and the trace id was almost always empty.
  • MapleBrowser.init() with no argument starts on defaults and warns about the missing serviceName instead of throwing. If startup throws, init() warns, undoes what had started and returns a handle that does nothing, so telemetry cannot take the app's entrypoint down.
  • With @maple-dev/browser and the Effect client on one page, an uncaught error is recorded once and web vitals are reported by one of the two. The Effect client now uses the page-wide error claim and the webVitals lease, and tells the page's error listeners (breadcrumbs, replay on error) about the errors it records.
  • The Effect client records no replay for crawlers, as @maple-dev/browser already did. They still get a session row and events.
  • The Effect client removes its privacy.sanitizeUrl when it stops, so a remount no longer stacks a second copy, and it stops capturing when another Maple SDK on the page turns privacy.requireConsent on later.
  • Maple.layer (client and server) sends the x-maple-sdk header like the other presets; the server layer also names itself in user-agent.
  • The session record is no longer JSON-parsed and validated for every span and replay input event: while the stored string is unchanged, the record parsed last time is reused.

• 04fd697: A failed export no longer fills the console:

  • When an export failed (for example a browser's tracking protection blocking the ingest request), the SDK printed flush skipped (cooldown ...) on every flush tick of the 60s cooldown, per signal, and flush failed again after every failed retry. A blocked endpoint produced about 24 lines a minute for the whole session. Skipped flushes now print nothing, an outage is logged once per signal, and one flush recovered line follows when export works again. Buffering, restore and the 60s cooldown are unchanged.
  • Console capture (session replay, breadcrumbs and logs.captureConsole) no longer records the SDKs' own [maple] and [MapleClientSDK] diagnostics as the app's console output, so an export failure is not counted as a session error.

@maple-dev/effect-sdk@0.14.1

Patch Changes

• c022d7f: Network capture no longer rejects a fetch call that succeeded:

  • fetch accepts anything it can stringify, such as fetch(window.location) or a URL from another frame. Capture read a url property those values do not have, threw after the response had arrived, and the app's await fetch(...) rejected with a TypeError. The URL is now read the way fetch reads it.
  • A request that does not ask for keepalive no longer has its body encoded just to measure it.

• 613df48: Session replay no longer changes how a page's own stylesheet calls fail, and benign browser notices no longer count as session errors:

  • A page whose code calls insertRule (or deleteRule, replace, replaceSync, setProperty, removeProperty) with something the browser rejects gets the browser's exception again. While a recording ran, the call returned undefined instead of throwing and the console showed [maple] session replay recording failed. CSS-in-JS libraries that insert vendor-prefixed rules and rely on catching the error hit this on every page load. The recording itself was not interrupted.
  • The ResizeObserver loop notices and errors thrown by browser extensions are no longer recorded as session errors, so they stop counting toward a session's error total and its "has errors" filter. These are the default error filters that already keep them out of issues; errors: { defaultFilters: false } in @maple-dev/browser keeps them in both.

• 58c2b10: Capture no longer changes what the app sees, and both SDKs agree on one page:

  • Network capture cannot reject a fetch that succeeded. Anything capture throws after the response arrived (a URL filter, a body pattern, reading the body in the background) is dropped, the app gets the real response, and a failed request still rejects with its own error. An XHR's loadend bookkeeping can no longer raise an uncaught error on the page.
  • Session events carry the time, page URL and trace id of the moment they happened. Clicks, inputs, errors, navigations, console and network events were stamped when the batch was sent, up to a minute later: a click that navigated was attributed to the destination page, and the trace id was almost always empty.
  • MapleBrowser.init() with no argument starts on defaults and warns about the missing serviceName instead of throwing. If startup throws, init() warns, undoes what had started and returns a handle that does nothing, so telemetry cannot take the app's entrypoint down.
  • With @maple-dev/browser and the Effect client on one page, an uncaught error is recorded once and web vitals are reported by one of the two. The Effect client now uses the page-wide error claim and the webVitals lease, and tells the page's error listeners (breadcrumbs, replay on error) about the errors it records.
  • The Effect client records no replay for crawlers, as @maple-dev/browser already did. They still get a session row and events.
  • The Effect client removes its privacy.sanitizeUrl when it stops, so a remount no longer stacks a second copy, and it stops capturing when another Maple SDK on the page turns privacy.requireConsent on later.
  • Maple.layer (client and server) sends the x-maple-sdk header like the other presets; the server layer also names itself in user-agent.
  • The session record is no longer JSON-parsed and validated for every span and replay input event: while the stored string is unchanged, the record parsed last time is reused.

• 04fd697: A failed export no longer fills the console:

  • When an export failed (for example a browser's tracking protection blocking the ingest request), the SDK printed flush skipped (cooldown ...) on every flush tick of the 60s cooldown, per signal, and flush failed again after every failed retry. A blocked endpoint produced about 24 lines a minute for the whole session. Skipped flushes now print nothing, an outage is logged once per signal, and one flush recovered line follows when export works again. Buffering, restore and the 60s cooldown are unchanged.
  • Console capture (session replay, breadcrumbs and logs.captureConsole) no longer records the SDKs' own [maple] and [MapleClientSDK] diagnostics as the app's console output, so an export failure is not counted as a session error.

@maple/cli@0.3.6

Patch Changes

• Updated dependencies [c022d7f]
• Updated dependencies [613df48]
• Updated dependencies [58c2b10]
• Updated dependencies [04fd697]
  • @maple-dev/effect-sdk@0.14.1

[coderabbitai]

Important

Review skipped

Bot user detected.

To trigger a single review, invoke the @coderabbitai review command.

Run configuration

• Source mode: Auto
• Repository: Unavailable
• Branch: Unavailable
• File: No configuration file from this repository applied
• Commit: Unavailable
• Layers: defaults
• Review profile: CHILL
• Plan: Advanced
• Run ID: 7a85b779-e882-4118-b0b1-60aa268644ce

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• Trigger review

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

Share

• X
• Mastodon
• Reddit
• LinkedIn