## Provider-specific Notes

* AWS
  * Support for ap-northeast-2 region (Seoul)
  * Allow cross-region image pulling with ECR
  * More reliable kube-up/kube-down
  * Enable ICMP Type 3 Code 4 for ELBs
  * ARP caching fix
  * Use /dev/xvdXX names
  * ELB:
    * ELB proxy protocol support 
	* mixed plaintext/encrypted ports support in ELBs
    * SSL support for ELB listeners

#     v2: Build TF v2
#
# Feature and Third party library support options:
#     xla:          Build TF with XLA
#     tpu:          Build TF with TPU support
#     cuda:         Build with CUDA support.
#     cuda_clang    Build with CUDA Clang support.
#     rocm:         Build with AMD GPU support (rocm)
#     mkl:          Enable full mkl support.
#     tensorrt:     Enable Tensorrt support.
#     noaws:        Disable AWS S3 storage support

        write_to_bazelrc(
            'build --copt="-DEIGEN_ALTIVEC_ENABLE_MMA_DYNAMIC_DISPATCH=1"')

  with_xla_support = environ_cp.get('TF_ENABLE_XLA', None)
  if with_xla_support is not None:
    write_to_bazelrc('build --define=with_xla_support=%s' %
                     ('true' if int(with_xla_support) else 'false'))

  set_action_env_var(
      environ_cp, 'TF_NEED_ROCM', 'ROCm', False, bazel_config_name='rocm')

config_setting(
    name = "stackdriver_support",
    define_values = {"stackdriver_support": "true"},
    visibility = ["//visibility:public"],
)

config_setting(
    name = "with_xla_support",
    define_values = {"with_xla_support": "true"},
    visibility = ["//visibility:public"],
)

# By default, XLA GPU is compiled into tensorflow when building with
# --config=cuda even when `with_xla_support` is false. The config setting

_2019-02-05_

 *  Fix: Don't crash when using a custom `X509TrustManager` or `SSLSocket` on Android. When we
    removed obsolete code for Android 4.4 we inadvertently also removed support for custom
    subclasses. We've restored that support!


## Version 3.13.0

_2019-02-04_

 *  **This release bumps our minimum requirements to Java 8+ or Android 5+.** Cutting off old

        : new FilteredEntrySetMultimap<K, V>(checkNotNull(unfiltered), entryPredicate);
  }

  /**
   * Support removal operations when filtering a filtered multimap. Since a filtered multimap has
   * iterators that don't support remove, passing one to the FilteredEntryMultimap constructor would
   * lead to a multimap whose removal operations would fail. This method combines the predicates to

            """
            OpenAPI-specific examples.

            It will be added to the generated OpenAPI (e.g. visible at `/docs`).

            Swagger UI (that provides the `/docs` interface) has better support for the
            OpenAPI-specific examples than the JSON Schema `examples`, that's the main
            use case for this.

            Read more about it in the

   * TODO(cpovirk): Is there any way for us to support TypeParameter instances for type parameters
   * that have nullable bounds? Unfortunately, if we change the parameter to TypeParameter<? extends
   * @Nullable X>, then users might pass a TypeParameter<Y>, where Y is a subtype of X, while still
   * passing a TypeToken<X>. This would be invalid. Maybe we could accept a TypeParameter<@PolyNull

    private SetView() {} // no subclasses but our own

    /**
     * Returns an immutable copy of the current contents of this set view. Does not support null
     * elements.
     *
     * <p><b>Warning:</b> this may have unexpected results if a backing set of this view uses a
     * nonstandard notion of equivalence, for example if it is a {@link TreeSet} using a comparator

                }
            }
        }
        return list;
    }

    /**
     * Returns the elements placed on the classpath for compilation.
     * This method can be invoked when the caller does not support module-path.
     *
     * @throws DependencyResolutionRequiredException if an artifact file is used, but has not been resolved
     *