"tests",
                        false,
                        JavaPathType.CLASSES,
                        JavaPathType.PATCH_MODULE),
                new DefaultType(Type.MODULAR_JAR, Language.JAVA_FAMILY, "jar", null, false, JavaPathType.MODULES),
                new DefaultType(Type.CLASSPATH_JAR, Language.JAVA_FAMILY, "jar", null, false, JavaPathType.CLASSES),
                // j2ee types

# For projects which use TensorFlow as part of a Bazel build process, putting
# nothing in a bazelrc will default to a monolithic build. The following line
# opts in to modular op registration support by default.
build --define framework_shared_object=true
build --define tsl_protobuf_header_only=true

build --define=use_fast_cpp_protos=true
build --define=allow_oversize_protos=true

It is very simple and intuitive. It's designed to be easily extensible, and have modular components.

It has:

* Seriously impressive performance.
* WebSocket support.
* In-process background tasks.
* Startup and shutdown events.
* Test client built on HTTPX.

	<target name="install.module">
		<get dest="${target.dir}">
			<url url="${repo.url}/${module.groupId}/${module.name.prefix}${module.name}/${module.version}/${module.name.prefix}${module.name}-${module.zip.version}.zip" />
		</get>
		<unzip dest="${modules.dir}/${module.name}" src="${target.dir}/${module.name.prefix}${module.name}-${module.zip.version}.zip">
			<patternset>
				<include name="**" />
			</patternset>
			<cutdirsmapper dirs="1" />

#include "absl/strings/match.h"
// Required for IS_MOBILE_PLATFORM
#include "tensorflow/core/platform/platform.h"  // NOLINT

#if !defined(IS_MOBILE_PLATFORM) && !defined(IS_SLIM_BUILD)
#include "tensorflow/c/experimental/filesystem/modular_filesystem.h"
#include "tensorflow/cc/framework/gradients.h"
#include "tensorflow/cc/framework/ops.h"
#include "tensorflow/cc/framework/scope_internal.h"
#include "tensorflow/cc/ops/while_loop.h"

        override fun afterInvalidation(modules: Set<KtModule>) {
            modules.forEach { analysisSessionProvider.cache.remove(it) }
        }

        override fun afterGlobalInvalidation() {
            // Session invalidation events currently don't report whether library modules were included in the global invalidation. This is

        selective op registration headers else execution would fail with
        unregistered kernels error.

*   Modular File System Migration:

    *   Support for S3 and HDFS file systems has been migrated to a modular file
        systems based approach and is now available in
        https://github.com/tensorflow/io. The `tensorflow-io` python package

            }

            for (String module : parent.getModules()) {
                module = module.replace('\\', '/');

                if (module.regionMatches(true, module.length() - 4, ".xml", 0, 4)) {
                    module = module.substring(0, module.lastIndexOf('/') + 1);
                }

                String moduleName = module;
                if (moduleName.endsWith("/")) {

            .java files in c.g.c.collect.Collect, a module that the c.g.c.base test module inherits,
            and .java files in c.g.c.collect.testModule, one that it doesn't[*]. Consequently,
            c.g.c.base.testModule transitively pulls in .java files from both modules while reading
            the module description of only the prod module. And the prod module doesn't inherit all
            the modules that the test module classes use, so we get errors.

  <modules>
    <module>maven-api-meta</module>
    <module>maven-api-di</module>
    <module>maven-api-xml</module>
    <module>maven-api-model</module>
    <module>maven-api-plugin</module>
    <module>maven-api-settings</module>
    <module>maven-api-toolchain</module>
    <module>maven-api-metadata</module>
    <module>maven-api-core</module>
    <module>maven-api-spi</module>
  </modules>