These three methods were delegating to delegate; their implementation
was duplicated across both subclasses of DelegatingHttpsURLConnection.
This change moves them into the base class in order to
- avoid duplication
- separate the responsibility of delegation (base class) from the
responsibility of change to behavior (subclasses).
Notes:
- Since these three methods are not available in Java 6, each of
these need a @IgnoreJRERequirement annotation to pass the
animal sniffer test. This appears like it should have been
necessary before but for some reason it used to pass without.
- This in turn requires a new dependency from okhttp-urlconnection
onto animal-sniffer-annotations. The dependency was added onto
version 1.11 of that library (consistent with okhttp itself);
okio depends on version 1.10, the latest version is 1.15.
This is an internal class so nobody should see the name. But having a
completely different name will make it much easier to diagnose future
stacktraces because the internal implementation has changed so much
recently.
Also share fewer internal details with the rest of OkHttp. In particular
remove the OkHttp-Selected-Protocol and OkHttp-Response-Source headers
from the core.
I'm currently keeping these as longs. That's easy an unambiguous, but it also
feels rather primitive.
This was more work than I expected, but it doesn't seem particularly risky.
It's also potentially more efficient, since there's fewer places where we're
converting from string to long and vice versa.
Closes: https://github.com/square/okhttp/issues/2035
Still some work to do on method naming, but this was the interesting part.
Also very much need to rename OkUrlFactory in tests to urlFactory, otherwise
the client.client stuff becomes madness.
There's a big consequence to this for compatible with RFC 2965 Cookies. In particular,
this drops our ability to support 'Set-Cookie2' headers, and some features used there
including quoted attributes like 'Max-Age="25"'. This is the right move for interop with
browsers, and likely to make things better for application developers, but some
people who strictly followed the RFC 2965 spec will be broken. (That spec was never
seriously adopted anywhere, which is the entire motivation of RFC 6265.)
The upside is the CookieJar interface is now much more straightforward. This is
particularly good going forward, and for clients who don't have to worry about the
strangeness of things like $Version=1 in the RFC 2965 spec.
No good first-party implementation yet.
There is an implementation in okhttp-urlconnection that delegates to the java.net
CookieHandler. That should be good enough for AOSP.
This is awkward because it adds another field to the cache. I've
decided to try to keep old values in the cache working. We will
probably do a change that'll disrupt the cache later, but that
is invasive and we don't need that yet.
There's a mostly-academic corner case on URLs like http://host/%xx that
have a malformed escape sequence. This changes the exceptions on those
URLs from checked to unchecked.
I think we may want a separate change to HttpUrl to fix those URLs to
be encoded as http://host/%25xx which avoids the problem altogether.
Closes: https://github.com/square/okhttp/issues/2106
We aren't yet ready to make FileSystem a public type, but I don't mind
making it _almost_ available via an internal API for those brave enough
to try that.
Also migrate all of our tests to use the in memory file system. It's simpler.
https://github.com/square/okhttp/issues/1459
Keep the embedded OkHttp on Android working with Vary headers:
Vary headers were broken with android.net.http.HttpResponseCache:
the OkHttp caching requires a networkResponse to check the vary
headers and it was not being provided, leading to a
NullPointerException. Further changes were needed to actually
retain the request headers needed for the vary check itself.
Support for other stacks using the RequestCache has been
dropped for requests that include Vary headers.
To avoid regressions the existing ResponseCacheTest has been
changed to use the real cache. Previously, the use of
InMemoryResponseCache was hiding the Vary problem.
The ResponseCacheTest is now run with an AndroidShimResponseCache
to prove the tests work with real cases. The case without the shim
is covered with the CacheTest. ResponseCacheTest has been updated
to include many tests from CacheTest that were missing.
Switching the test to a real cache revealed a problem in the
JavaApiConverter and bad headers, which has been fixed by
making Headers.Builder.addLenient(String, String) accessible.
Including the code in
okhttp-android-support/src/test/java/com/squareup/okhttp/android
provides easy verification of API or behavior changes that might
affect Android.
There are some small JavaApiConverter improvements.
okhttp-android-support contains classes needed for Android embedding and
not for normal OkHttp development. The classes here can be excluded from
okhttp-urlconnection, shrinking that artifact. More classes will be
added to this component to make Android maintenance easier.
