Both connection managers agree on the meaning of a zero-duration
validateAfterInactivity value. Previously the documentation for
both suggested that zero resulted in no validation, however
the classic client would validate prior to each request and
the async client would never validate.
This commit standardizes behavior on the classic client, which
is consistent with hc4.x, allowing zero to force validation prior
to every request.
Previously HttpClientBuilder methods addExecInterceptorFirst and
addExecInterceptorLast resulted in NullPointerExceptions because
the execInterceptors list had not been initialized yet.
Modeled after StandardCharsets, the new class indicates a non-exhaustive
list of standard cookie specifications by name supported by the HttpClient.
The previous enum suffered from two issues:
* it was exhaustive implying that no more cookie specifications can be supported
* its sole purpose was to contain an id for the declared cookie specification;
the enum values theirselves were never used directly
This also reuses the naming approach of HttpMultipartMode of symbolic names for
implementation behavior.
This closes#197
This change ensures that during hostname verification the public suffix
list is only used to prevent wildcard matching against entire TLDs (e.g.
`*.com`). Currently, private domains are also being matched against,
which is preventing reasonable wildcards (such as
`*.s3.eu-central-1.amazonaws.com`) from being respected.
Throughout the code the terms 'authScheme' and 'scheme' have been used
synonymously for an AuthScheme instance and a string-based auth scheme
name. To avoid confusion, fields, methods and variable have been adapted
to distinguish both properly. If necessary, Javadoc has been modified to
denote the nature of the input.
Also an auth scheme name is retained as-is, but normalized to lowercase
if comparsion is required.
This closes#193
Modeled after StandardCharsets, the new class indicates a non-exhaustive
list of standard auth schemes by name supported by the HttpClient. The
previous enum suffered from two issues:
* it was exhaustive implying that no more auth schemes can be supported
* its sole purpose was to contain an id for the declared auth scheme;
the enum values theirselves were never used directly
This closes#196
