-
-
Notifications
You must be signed in to change notification settings - Fork 1.6k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
meta: Update CHANGELOG for 8.18.0 #12932
Conversation
[Gitflow] Merge master into develop
Only add this if this is some other value, e.g. `CLIENT` or `SERVER`. It is useless to have the internal value there and potentially confusing (why is this there? what does internal even mean?)
Adds an option to trace long animation frames as per #11719. This tracing feature is disabled by default. Currently, this feature https://developer.mozilla.org/en-US/docs/Web/API/Performance_API/Long_animation_frame_timing is only supported by Chromium browsers. Usage is opt-in: ```js Sentry.init({ dsn: '__PUBLIC_DSN__', integrations: [ Sentry.browserTracingIntegration({ enableLongAnimationFrame: true, }), ], tracesSampleRate: 1, }); ```
Remove `dom` types from `packages/typescript/tsconfig.json` and add it to the respective packages that need them. This is important because we should only be adding `dom` types to packages that will be used in the browser.
Co-authored-by: Francesco Novy <[email protected]>
Removes the usage of the `codecov` package as we use the v4 version of the codecov github action which uses the CLI directly. Helps reduce the dependency count!
…caught in submodules (#12868)
… custom page extensions (#12858) Co-authored-by: Luca Forstner <[email protected]>
For now this only tests client side errors. It also doesn't use the solidrouter integration yet because it fails to resolve the `@sentry/solidstart/solidrouter` import properly in our repo. I have yet to figure out what exactly is causing this (my guess is some pnpm shenanigans). I couldn't reproduce this in a sample app outside of our repo yet. Closes: #12547
Fixes #12855 Previously, we stored the route name of the pageload in a map for INP interactions. However, in some frameworks - e.g. remix, but also others, we update the pageload span name later (since we rely on e.g. react hooks for this etc). Since we store the name of the pageload span at the time the first interaction is recorded, it can thus happen that these run out of sync. This PR changes this so that instead of the routename itself, we store the pageload span in a map, and pick the last name of this when generating the INP span. I added tests in a remix e2e tests that show the now correct behavior, these used to fail (because `transaction` on the pageload INP was `/`).
See https://docs.github.com/en/actions/using-workflows/events-that-trigger-workflows#pull_request_target You can see e.g. here: https://github.com/getsentry/sentry-javascript/actions/runs/9888865345/job/27313696886 that this lacks permissions when run as `pull_request`. Now, this should ran in the context of the base branch, which grants proper rights etc. and also prevents leaking etc.
resolves #12685 Initial attempt: #12733 This PR introduces `@sentry/cloudflare`, our package for cloudflare pages and cloudflare workers. It also adds a basic test with [`miniflare` ](https://miniflare.dev/), the local simulator for cloudflare workers. In the next step, I'll add basic SDK initialization + error monitoring for the cloudflare pages use case.
Small optimization using the new hook cleanup capabilities to remove unused hooks. Ref PR to do this in angular: #12786
This PR adds the external contributor to the CHANGELOG.md file, so that they are credited for their contribution. See #12879
…ing in Node environments (#12892) Add a guard for referencing `ErrorEvent` which is only available in browser environments but not in Node.
* Uses clean `Array.from` implementation * Revert css parsing (player) * Implements multitouch gestures (player) * Many upstream fixes
…fig (#12842) Not all our e2e tests require a `startCommand` as we sometimes directly invoke an application from within tests. For example in our AWS lambda e2e tests. This PR makes the `startCommand` option optional so that we can avoid a larger override of the entire `webServer` playwright config object.
Adding a note here too as first time users are likely to be following this doc when they see the build fail due to python version incompatibility
) Add better build docs on how to build profiling binaries from source #12899
Closes #12878 I added this feature to `import-in-the-middle` which was released in v1.9.0: - nodejs/import-in-the-middle#124 This PR changes the hook from `@opentelemetry/instrumentation/hook.mjs` to `import-in-the-middle/hook.mjs` as it was only pasing though anyway and the otel hook doesn't pass the `initialize` export.
…n when used in solidstart sdk (#12919) The `@sentry/solid/solidrouter` integration is used within the solid start sdk as well, so we need a way to update the `sentry.origin` for navigation spans to be for the correct framework. I opted to read this out of the sdk metadata instead of exposing extra api surface and passing it through several levels.
While we specifically ignored dependabot, we did not ignore github bot. So now, just generally skipping this for any bot users.
Add the `window.navigator.userAgent` string as the [`user_agent.original`](https://opentelemetry.io/docs/specs/semconv/attributes-registry/user-agent/) attribute to INP standalone spans.
Add a simple, optional peer dependency on Vite to `@sentry/sveltekit` to ensure that it picks up whatever version of Vite is being used in the specific package it's installed in. This is necessary because `pnpm` otherwise could resolve the `vite` import to a wrong version, especially in monorepo environments
…n*` (#12928) Noticed this while writing docs, this makes this a bit harder to understand. With this change you can say that `parentSpan` behaves the same and accepts the same as `withActiveSpan`. See getsentry/sentry-docs#10729
…allback (#12929) Fix an oversight in our Node `httpIntegration`. It looks like we assumed that the `request` object being passed to `ignoreIncomingRequestHook` and `ignoreOutgoingRequestHook` was of the same type. However, it's not: - `request` is of type `IncomingMessage` in `ignoreIncomingRequestHook` - `request` is of type `RequestOptions` in `ignoreOutgoingRequestHook` fix the bug by simply taking the request.url property instead and adds integration tests to properly test the two options.
bb6ceae
to
fbd8ddb
Compare
fbd8ddb
to
c57e363
Compare
- fix(tracing): Ensure you can pass `null` as `parentSpan` in `startSpan*` (#12928) | ||
- ref(core): Small bundle size improvement (#12830) | ||
|
||
Work in this release was contributed by @GitSquared and @mcous. Thank you for your contributions! |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Hey @ziyadkhalil, sorry we missed this and thanks for letting us know. I've opened up a PR to add you in. Apologies again!
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thank you! 🙏🏻
No description provided.