Update the `spring-boot`, `spring-boot-autoconfigure` and
`spring-boot-actuator` project to generate configuration meta-data
files during compilation.

See gh-1001

Update `ConfigurationMetadataAnnotationProcessor` to include the
'defaultValue' of a property when possible. For example the
'defaultValue' or 'server.port' is '8080'.

Default values are detected by inspecting the field assignments of
@ConfigurationProperties items. In order to detect field values some
internals of the Java compiler are used. To save a dependency on
'tools.jar' internal javac classes are accessed using reflection.

See gh-1001

Adds an annotation processor to generates a JSON meta-data file at
compile time from @ConfigurationProperties items. Each meta-data file
can include an array or 'properties' and 'groups'.

A 'property' is a single item that may appear in a Spring Boot
'application.properties' file with a given value. For example,
'server.port' and 'server.context-path' are properties. Each property
may optionally include 'type' and 'description' attributes to provide
the data type (e.g. `java.lang.Integer`, `java.lang.String`) and
some short documentation (taken from the field javadoc) about what the
property is for. For consistency, the type of a primitive is translated
to its wrapper counterpart, i.e. `boolean` becomes `java.lang.Boolean`.

A 'group' provides a higher level grouping of properties. For example
the 'server.port' and 'server.context-path' properties are in the
'server' group.

Both 'property' and 'group' items may additional have 'sourceType' and
'sourceMethod' attributes to indicate the source that contributed them.

Users may use `META-INF/additional-spring-configuration-metadata.json`
to manually provide additionally meta-data that is not covered by
@ConfigurationProperties objects. The contents of this file will be
read and merged with harvested items. The complete meta-data file is
finally written to `META-INF/spring-configuration-metadata.json`.

See gh-1001

Update OnBeanCondition to use a new BeanTypeRegistry which includes
optimized code when using a DefaultListableBeanFactory. The optimized
version calculates bean types only once per bean and caches the result.

Prior to this change the sample "pet clinic" application would spend
400-500 milliseconds evaluating OnBeanConditions, after this change it
spends around 120 milliseconds.

Fixes gh-1803

Update ManagementSecurityAutoConfiguration to return
IgnoredPathsWebSecurityConfigurerAdapter rather than
WebSecurityConfigurer<WebSecurity>.

Conflicts:
	spring-boot-autoconfigure/src/test/java/org/springframework/boot/autoconfigure/security/SecurityAutoConfigurationTests.java

Instead of just blindly creating the default authentication manager, after
thic change we count the beans of type GlobalAuthenticationManagerConfigurer
and assume that if we detect more than we expect (one from Boot and one from
Spring Security) then the user is telling us they want to configure the
AuthenticationManager themselves.

Fixes gh-1801

Change InitializrService to use a late binding CloseableHttpClient
since the calling `HttpClientBuilder.create().build()` is slow.

Fixes gh-1764

Rename a few classes and methods and extract some logic into helper
classes. Also change 2 char shortcuts to a single char.

Closes gh-1751

Apply eclipse formatting rules to b2fe2dd9.

See gh-1751

See gh-1801

Fixes gh-1775

All metrics in the MetricRegistry have been added unconditionally for now.

Fixes gh-1795

This commit adds a new command to the CLI that allows to initialize a new
project from the command line. It uses the Spring initializr service to
actually generate the project.

The command offers two main operations:

1. Listing the capabilities of the service (--list or -l). This basically
dumps the defaults of a given service and the list of dependencies and
project types it supports
2. Generating a project. By default, http://start.spring.io is used and
its configured defaults are applied. Running spring init would therefore
have the same effect as clicking the 'generate project' on the UI without
entering any extra information. No file is overwritten by default.

The generation can be customized with the following options:

* --boot-version (-bv) Spring Boot version the project should use
* --dependencies (-d) comma separated list of dependencies to add to the
generated project
* --java-version (-jv) Java version to use
* --packaging (-p) the packaging for the project (jar, war)
* --target the url of the service to use

The actual type of the project can be defined in several ways:

1. Using the --type (-t) option that identifies a type that is supported
by the service
2. A combination of --build and/or --format that can be used to uniquely
identify matching these tags. Build represents the build system to use
(e.g. maven or gradle) while --format defines the format of the generated
project.

The project is saved on disk with the name provided by the server through
the Content-Disposition header, if any. It is possible to force it with
the --output option. It is possible to overwrite existing files by adding
the --force (-f) flag.

The --extract (-x) option allows to extract the project instead of saving
the zip archive. By default, the project is extracted in the current
working directory but it is possible to specify an alternate directory
using the --output option.

Fixes gh-1751

Update LogbackLoggingSystem to programmatically configure logback
defaults rather than parsing XML. This change shaves about 100ms off
the start up time.

Fixes gh-1796

GroovyWebConfiguration creates a GroovyTemplateViewResolver which is
a UrlBasedViewResolver sub-class. UrlBasedViewResolver is provided
by spring-webmvc. Previously, if a user configured a web application
but did not have spring-webmvc on the classpath, the application
would fail to start with a NoClassDefFoundError for
UrlBasedViewResolver.

This commit makes GroovyWebConfiguration conditional on
UrlBasedViewResolver being on the classpath so that it backs of in
the absence of spring-webmvc.

Fixes gh-1793

Previously, Spring Security's filter had no configured order. Due to
the use of AnnotationAwareOrderComparater this meant that its order
defaulted to LOWEST_PRECEDENCE. This meant that a user had to declare
a FilterRegistrationBean for the filter and explicitly set its order
if they want another filter to run after Spring Security's.

This commit updates the security auto-configuration to assign a
default order of zero to Spring Security's filter, allowing filters
to be easily configured to run before it or after it. This default
value can overridden using the server.filter-order property. The
default order is also exposed as a constant on SecurityProperties,
allowing it to be referenced from other filter declarations.

Closes gh-1640

Prior to this commit LoggingSystem initialization would happen multiple
times. Once to configure "quiet" logging, and again to configure correct
settings once the Application was initialized. This could cause problems
if `logging.groovy` logback files were used.

The logging system is now only initialized once (when possible) by
following these steps:

- Standard logging initialization occurs via the actual logging
  implementation used (e.g. logback will load a logback.xml file if it
  exists)
- beforeInitization() is called to prevent early log output.
  Implementations now either use a Filter or simply set the root logging
  level.
- initialize() is called with an optional log configuration file (e.g
  a custom logback.xml location) and an optional log output file (the
  default is null indicating console only output).

The initialize() method will attempt to prevent double initialization
by checking if a standard configuration file exists. Double
initialization now only occurs in the following situations:

- The user has a standard configuration file (e.g. classpath:logback.xml)
  but also specifies a logging.config property. Double initialization is
  required since the specified configuration file supersedes the default.
- The user has a standard configuration file (e.g. classpath:logback.xml)
  and specifies a logging.file property. Double initialization is
  required since the standard configuration may use a ${LOG_FILE}
  reference.

In addition this commit removes the `logging.console` option and now
assumes that logging either occurs only to console or to both the
console and a file. This restriction helps simplify the LoggingSystem
implementations. If file only logging is required a custom logback.xml
can be used.

Fixes gh-1091
See gh-1612, gh-1770

Closes gh-1787

Closes gh-1774

Closes gh-1767

Spring IO Platform already provides dependency management for JavaMail.
This commit updates Boot’s new JavaMail dependency management to align
with the Platform, thereby allowing the Platform to inherit Boot’s
dependency management instead of defining its own.

This commit adds a new starter to auto-configure a MailSender when the
necessary classes are present and when the property "spring.mail.host" is
set.

The auto-configuration also accepts any arbitrary properties that
JavaMail might need using the "spring.mail.properties" prefix.

Fixes gh-1760

There actually isn't anything much we can (or should) change here. Spring always
copies PropertySources and profiles down into the child context when you set its parent.
That seems sensible. So you can add a new profile in a child context but it will always
"inherit" from its parent.

See gh-1776

* whitelabel-typo:
  Fix a typo in error.whitelabel.enabled

Conflicts:
	spring-boot-dependencies/pom.xml

Closes gh-1782

Fixes #1781