This commit improves `AbstractValueAdaptingCache` to throw a dedicated
exception if `allowNullValues` is `false` and a `null` value is provided
anyway. This avoid a lower-level exception from the cache library that
will miss some context.
Issue: SPR-15173
Prior to this commit, the `ShallowEtagHeaderFilter` could participate in
the response and set its ETag/Content-Length headers, even for HEAD
requests. Since the response body is empty, the filter implementation
would set a `"Content-Length: 0"`.
The RFC states that responses to HEAD requests should exhibit identical
response headers to GET (with the possible exception of payload related
headers such as Content-Length.
With this commit, `ShallowEtagHeaderFilter` now ignores HEAD requests
since the proper values may be set already for payload related headers
by the handler. The filter has no way to generate a proper ETag value
nor calculate the content length without the actual body.
Issue: SPR-15261
Clearly separate how the result of an exchange is represented before
the response body has been read (e.g. assertions on status and headers
only) vs later after the body is extracted to a representation
(assertions on the extracted body) or is decoded to Flux<T> (e.g. for
use with a StepVerifier).
The first fix for issue 15264 covered the case of using a single
variable (the case mentioned in the bug report). However, when
more than one variable is used a different PathElement is built.
This RegexPathElement needs a similar change that checks the
path includes data to bind.
Issue: SPR-15264
WebTestClient now defines all the steps from setup to performing
an exchange and applying expectations.
The order of expectations now ensures the response status and headers
are verified first since that's available before the body is consumed
and also because it determines how the body is to be decoded, i.e.
error vs success scenarios.
There is now a built-in option for verifying the response as a Map
along with Map-specific assertions.
There are similar options for verifying the response as a List as well
as whether to "collect" the list or "take" the first N elements from
the response stream.
Rather than returning ExchangeActions that contains ExchangeInfo and
applies a chain of assertions, the exchange operation in WebTestClient
now has an extra step to decode the response.
The outcome of that is ExchangeResult<T>, parameterized with the
decoded body type, and containing the request and response details,
also providing access to built-inassertions via an assertThat()
instance method.
This approach lends itself better to decoding and asserting response
body types with generecis. It is also more friendly to using any
assertion library such as AssertJ since you get the result first
and then deal with assertions.
Removed `json()`, `html()` and `xml()` from `RequestPredicates`, since
they were confusingly named and would also require a counterpart that
reads the request `Content-Type` instead of the `Accept` header.
Renamed `RouterFunction.andSame()` to `and()`, and `and()` to
`andOther()`. The reason for this change is that we can expect most
RouterFunctions to be parameterized over ServerResponse, and thus it
makes sense to use the shortest composition method (`and()`) for
composing router functions of the same type.
When a user composes different response types, such as composing a
`RouterFunction<RenderingResponse>` with an
`RouterFunction<EntityResponse<?>`, the `andOther` method is to be used,
but this is a less common scenario.
Juergen Hoeller
Stephane Nicoll
Sebastien Deleuze
Arjen Poutsma
Brian Clozel
Rossen Stoyanchev
Andy Clement
Kazuki Miyahara