Currently, in order to use replay you have to run:
And then use it in your codebase like this:
import * as Sentry from '@sentry/browser';
import { Replay } from '@sentry/replay';
Sentry.init({
dsn: '__DSN__',
integrations: [
new Replay({})
],
});
We want to simplify this, so the user can do e.g.:
import * as Sentry from '@sentry/browser';
Sentry.init({
dsn: '__DSN__',
integrations: [
new Sentry.Replay({})
],
});
And does not need to run yarn add @sentry/replay at all.
Goals
- We do not affect bundle size for users who do not enable replay via CDN
- We do not affect bundle size for users who do not enable replay via a bundler with tree shaking
- This should happen in a backwards-compatible way (e.g. in 7.x)
- Ideally, users that have replay currently setup with the manual import from
@sentry/replay can continue using this for the forseeable time
Non-goals
- For users that have, for whatever reason, no tree shaking, the bundle size may increase
Related
We may want to do the same for @sentry/tracing in the future.
Currently, in order to use replay you have to run:
And then use it in your codebase like this:
We want to simplify this, so the user can do e.g.:
And does not need to run
yarn add @sentry/replayat all.Goals
@sentry/replaycan continue using this for the forseeable timeNon-goals
Related
We may want to do the same for
@sentry/tracingin the future.