The `generateClientInfo` API, used to set client identification attributes
within traces, is an experimental API and is subject to removal or change in
a future (major) Apollo Server release.
Ref: #1631
* Provide ability to specify client info in traces
Adds the createClientInfo to apollo-engine-reporting, which enables the
differentiation of clients based on the request, operation, and
variables. This could be extended to include the response. However for
the first release. It doesn't quite make sense.
* Use extensions and context in createClientInfo
* Remove support for clientAddress
The frontend will not support it in the near future
* create -> generate and make default generator
createClientInfo -> generateClientInfo
* Clarify default values
* Allow an optional function to resolve the rootValue
Passes the parsed DocumentNode AST to determine the root value,
useful when providing a different rootValue for query vs mutation
* Add API docs for rootValue
Some changes are easier to mentally consume before they've been prettier'd.
While I'll certainly allow Prettier to do what it wants to do, I've confined
it's changes to this commit for easier reviewability.
Note to the reviewer: this commit is exactly the result of running the
`npm run lint-fix` command on a clean Git working tree.
Ref: c9375c405b
The techniques previously used to skip particular tests for particular
Node.js versions (for example, file uploads, Hapi, etc.) were problematic in
a couple ways:
1. `return`-ing outside a function (say, at the top level of a module) is
not valid ECMAScript. While it works in this current state, it only works
because of Babel's wrapping scope. As we move toward less and less
transpiled modules, this would become problematic anyway, but it's a
short-term blocker for me since I intend on dropping the `skipBabel` option
we currently pass to` `ts-jest` in its current (`package.json`)
configuration. (My editor also dislikes top-level `return`, expectedly.)
2. By `return`-ing from a test which we wish to skip conditionally, we avoid
the awareness which is provided by Jest (automatically) when using its
built-in `.skip` helpers. By switching to technique I've employed here,
we're able to see `<num> skipped` in Jest's output and (rightfully) be able
to question why something is being skipped.
By using the `atLeastMajorNodeVersion` helper, we'll also be able to more
easily find where this pattern is used in test-cases, to more easily
re-evaluate their use-cases when we add support newer Node.js versions.
I'm reverting apollographql/apollo-server#1698 not because it's been problematic in any way, but because I'd like to give it a bit more thought and don't want this to accidentally get cut into a release prior to that consideration.
More specifically: The `graphql-tools` update on its own shouldn't really cause any problems, but the [4.x version of `graphql-tools`](https://github.com/apollographql/graphql-tools/releases/tag/4.0.0) is intended to support and enable the latest `graphql@14` which contains [breaking changes](https://github.com/graphql/graphql-js/releases/tag/v14.0.0).
I believe most of those breaking changes would be show-stoppers and the failures would surface immediately (meaning that servers would completely fail to start, rather than being a surprise in other, more delayed scenarios), but it's still worth pausing and carefully considering versioning to avoid any surprises.
That said, the 14.x version of `graphql` has been an acceptable range in the `peerDependencies` of `apollo-server-*` since before its final release came out, and I don't believe we've caught wind of anything that a major version bump would have prevented or made more clear. In the end, `graphql` is a peer dependency and any problems should only surface if consumers also update their `graphql` dependency — a clear major version bump, which deserves review by the upgrader — so perhaps we can avoid bumping the major version after all?
Input welcomed, but again, merging this now to give this a bit more thought first.
cc @hwillson
`esModuleInterop` was enabled in
e4164c8892
to help with importing from packages that use default exports.
Those changes were reverted in
https://github.com/apollographql/apollo-server/pull/1210
to work around a few reported issues. Those issues are no longer
relevant, so this commit re-enables `esModuleInterop`, and
updates all default imports to use the more common (standard)
import syntax.
* Allow an optional function to resolve the rootValue
Passes the parsed DocumentNode AST to determine the root value,
useful when providing a different rootValue for query vs mutation
* Add API docs for rootValue
* Specify explicit `cursorShape` to avoid missing cursor in GraphQL Playground.
This solves a problem with the text cursor caret not being visible in the
operation input box within GraphQL Playground by explicitly setting a
GraphQL Playground configuration option called `cursorShape` to `line`.
All credit for the actual solution goes to @lpellegr for their discovery in
https://github.com/prisma/graphql-playground/issues/790#issuecomment-409221277
* Update CHANGELOG.md
The `generateClientInfo` API, used to set client identification attributes
within traces, is an experimental API and is subject to removal or change in
a future (major) Apollo Server release.
Ref: #1631
* Updated request handler to handle null or undefined req.path
* Added test for GCF handling request.path being null
* Updated Changelog
* Add note about checking for null path
* Provide ability to specify client info in traces
Adds the createClientInfo to apollo-engine-reporting, which enables the
differentiation of clients based on the request, operation, and
variables. This could be extended to include the response. However for
the first release. It doesn't quite make sense.
* Use extensions and context in createClientInfo
* Remove support for clientAddress
The frontend will not support it in the near future
* create -> generate and make default generator
createClientInfo -> generateClientInfo
* Clarify default values
* Pass the context along to all the extension methods
Addresses #1343
With this change you should now be able to implement an extension like so:
```javascript
class MyErrorTrackingExtension extends GraphQLExtension {
willSendResponse(o) {
const { context, graphqlResponse } = o
context.trackErrors(graphqlResponse.errors)
return o
}
}
```
Edit by @evans :
fixes#1343fixes#614 as the request object can be taken from context or from requestDidStart
fixes#631 ""
* Remove context from extra extension functions
The context shouldn't be necessary for format, parse, validation, and
execute. Format generally outputs the state of the extension. Parse and
validate don't depend on the context. Execution includes the context
argument as contextValue
* Move change entry under vNext
Fixes#1613.
We always send traces that includes an error node if the trace has an
error. In the case that sending errors is disabled, we replace the
message and remove the location.
Note that the Engine proxy strips all error information from the traces
with noErrorTraces set. To get errors to show up in the ui, the proxy
sends error counts inside of the aggregated stats reports. To get
similar behavior inside of the apollo server metrics reporting, we
always send a trace and mask out the PII.
Playground can either use external settings or its default settings (browser settings), in incompatible ways. The original implementation of default settings here would always provide 'settings' for Playground, in turn make its own settings mechanism disabled, with consequences such as altering settings in-browser cannot work as expect. Besides, once the default settings here go different with Playground provided, it would cause other problems.
The intention here is to use Playground setting as default, while explicit passing Playground settings in Apollo server would still work with defaults defined here merged.
