…to try-with-resources

TemplateFile.java:
In retrieveFirstAndLastNames(), fixed a bug.

Reformat sdk toolkit descriptor
Move some descriptor informations in the toolkit pom.xml

Remove maven version from the sdk toolkit extract folder.

The copyright maven was broken with files which were not commited.
This is a regression introduced by 84e18a7.

The invalidCredRecentCount counter was maintained but not used anywhere since commit 6e1f91e.

* Move copyright-maven-plugin executions to the repositority parent pom
to allow copyright checks fail fast.
* Remove duplicated reporting section in opendj-sdk-parent/pom.xml file

* Creates a parameter for the Git branch which is used against HEAD in
the diff command. And change its default value from master to
origin/master (Thanks to JNR for the tip).
* Also fix little bug (duplicate file name in the incorrect file list)

Other code cleanups.
Reset the clirr api changes against 3.0.0-SNAPSHOT (waiting for 3.0.0)

* Update THIRDPARTYREADME file according to the BOM 4.1.1
* Update copyright year when it was needed

Copyright plugin now supports .properties files.
Thanks to JNR for the RFE.

Removes useless code in copyright maven plugin.

SSLContextBuilder.java:
Added constant PROTOCOL_TLS1_2 (RFC 5246: TLS version 1.2)
Code cleanup

Ensure default attribute types are created with the provided syntax

AttributeType.java:
These changes allow the server to create custom AttributeTypes with a non strict schema.

Schema.java:
Consequence of the changes to AttributeType.
Added getAttributeType(String name, Syntax).

…cable.

…ame is checked.

Moves the excluded files list to the repository top pom (sdk-bom) since
the copyright maven plugin execution has been moved.
This will enable the exclude file list which was unexpectedly ignored
since c9b49f7.

Creates a property name to allow user to set the reference git branch to
use in the diff command.
Also log a quick message to explain how to use the property.
Extracs some constants to mutualize the code.

… launching perf test

For all *rate tools, abort the run if the tool cannot connect to the server (wrong bind options, server unreachable, etc.)

PerformanceRunner.java:
Added validateCanConnectToServer()

…and line arguments

Add missing base64 decoding to match the documented behaviour/syntax

In remove*() methods, renamed parameter from name to nameOrOid where applicable.

Added create(AttributeType, Collection<String> options), needed by OpenDJ server.

…pe, Collection<String> options)

…atching rule.

SchemaBuilder.java:
In removeSyntax(Syntax), also removed the ordering matching rule associated to an enum syntax.
Removed code duplication by extracting method addEnumerationSyntax0().

…ments.

If for some reasons, all rate tool threads fail with the same error, we will have the same error printed n times on console.
This is fixed by adding synchronization mechanism with stopTool() method.
To ensure that reported statistics remain consistent if tool is stopped by error, the statistic thread is now scheduled by a ScheduledExecutorService.

This change consists in three steps:

* Reboot the MultiColumnPrinter
** In order to help code readability and to prepare metrics core
integration, the MultiColumnPrinter has been refactored.
** Add a Column class to store metrics informations related to UI
** The new printer can be used to print formatted data line in separated
columns or just raw CSV line.

* Use metrics to render stats
** The metrics framework helps to make the code more readable.
** It also use LongAdder for counters.

* Use HdrHistogram to compute quantiles
** To compute three, four and five nines stats, we now use HdrHistogram
instead of our own histogram implementation
** This API is designed for recording histograms of value measurements
in latency and performance sensitive applications.
** We use a bridge to integrate HdrHistogram with metrics framework (see
THIRDPARTYREADME file)

* Minor changes
** Remove child method must call super anti pattern in UpdateStatsResultHandler class hierarchy
** Extract StatsThread class inits own file
** Various other cleanups

This option was (almost) never used.
In addition, it just consists to perform a DoS attack, so Matt suggests
to remove it.

… attribute type

The solution consists in adding AVA.getAttributeName().

AVA.java:
Added attributeNakme field + getter + modified toString()
Addeed constructor AVA(AttributeType, String attributeName, Object attributeValue).
In readAttributeName(), changed return type from AttributeType to String.
Improved javadoc, by ading a note about the name of this class.
Extracted method illegalCharacter().
Simplified needEscaping().

AVATestCase.java: ADDED

ByteString.java:
Made toHexString() compatible with valueOfHex() - necessary to have a correct AVA.toString()

ByteStringTestCase.java:
Added a test for encoding/decoding as hex.

CollationSubstringMatchingRuleTest.java:
Consequence of the change to ByteString.toHexString().

The ConnectionFactoryProvider is storing the port in two places: an IntegerArgument and an int field.
Remove the int field and only keep the IntegerArgument to be consistent with other arguments.

ConnectionFactoryProvider.java:
Removed port field.
In getPort(), conditionally read the port from standard input.
if running in interactive mode and the port argument was not provided before
In getConnectionFactory(), avoid repeting conditions.

The tests were carried over from the server's test suite.

PR-45
Remove unused code.
Remove unnecessary JavaDoc.
Use interface type instead of implementation type as declarative type.
Other minor cleanup actions.

PR-45
Create a fluent builder class hierarchy to allow fluent argument
creation.

* An argument has a mandatory long identifier
    * The name attribute has been merged with long identifier and removed
    * Remove useless null checks on argument long identifiers
    * Change impacted error messages
    * Create some constants (remove hard coded old lower case argument name)
* Now all argument can be set in a property file.
  The argument property name is always the same as the argument long identifier (i.e can not be set any more)
* Default builder values are null for object attributes and false for boolean attributes.
* Remove the possiblity to set if an argument needs a value since it was a non hidden
implementation detail.
* For IntegerArgument default lower bound is Integer.MIN_VALUE
and the default upper bound is Integer.MAX_VALUE
* Once every needed attributes are contributed, the argument can be
created by the builder or added to an argument parser or a subcommand.
This commit also includes additional related cleanup.

sdk PR-45
Rename CommonArguments.getFooBar() to CommonArguments.fooBarArgument().
Static import com.forgerock.opendj.cli.CommonArguments.* everywhere
CommonArguments class is used.
Plus in few places, related minor cleanup actions.

PR-45
Those parameters were copied from the server tools but are not used any
more since API changes.

* ldapmodify
    * defaultAdd (-a, --defaultAdd) which is enabled by default in the sdk.
* ldapmodify and ldapsearch
    * -i (--encoding {encoding}) which is actually completely ignored,
      a user can provide it with any value he wants and this will have
      no consequences at all.

Merged LoadBalancer and LoadBalancingAlgorithm hierarchies into a single
class hierarchy in order to make it easier to implement request load-
balancers:

* LoadBalancer - abstract class providing common functionality for
connection and request load-balancers
* ConnectionLoadBalancer - abstract class for connection load-balancers
(round-robin and fail-over)
* ConnectionLoadBalancer - abstract class for connection load-balancers
(round-robin and fail-over)
* Connections - moved common load-balancing options to this class.

Added RequestLoadBalancer as common base class for for request based
load-balancers. Sharded load-balancer is implemented in Connections via
a static factory method.

* Align the copyright headers with the rest of ForgeRock source code.
* Ensure that we always use "ForgeRock AS."
* org/forgerock/opendj/ldap/schema/GenerateCoreSchema.java
    * Change generated copyright
    * Code cleanup

Apdapts the copyright maven plugin according to changes performed on
copyright header in the previous commit.
Also adapts unit tests.

Use enum to describe what a comment should looks like for supported
extensions.
Also renames test files in consequences.
Add support for the following extensions: java.stub, mc, security,
xml.vm

Changes needed for refactoring in server.
    * Moves user validation input for port number in Utils to allow
server code to use it.
    * Little enhancements in CommandBuilder and in CommonsArguments

* Add resource bundle for locales: ca_ES, de, es, fr, ja, ko, pl, zh_CN, zh_TW.
* For each locales, add the translation of ERR_TOOL_CONFLICTING_ARGS message

Add useful methods to handle arguments conflicts.

…ng in an unindexed search.

Aligning Indexer and SubString Assertion with regards to generated name for
substring indexes, who must contain the substring length configured for
the index, since it is used for the name of the tree in the Storage and
for the index itself.

This patch indicates the default for each "edition" delivered for version 3.

As we were using the buildnumber-maven-plugin in few places of the sdk
and server, factor out the 'generate-buildnumber' execution in opendj-
sdk-parent pom:

* Use the 1.4 version
* Use buildRevision as property name for build number
* Keep the useLastCommittedRevision property to false (default value)

Generalize the way that tool version and revision are computed and move
it to opendj-cli module.
Server's DSConfig class will then be able to use
ToolVersionHandler.newToolVersionHandler("opendj-config")

SchemaBuilder.java:

* In addSchema0(), do not call addAttributeType(), but
buildAttributeType() like for Syntax and MatchingRule

SchemaBuilderTestCase.java:

* Added a non regression test

These changes are required to complete the migration of the server to the SDK's DN and RDN classes.

DN.java:
In toNormalizedByteString(), changed the representation to include a leading NORMALIZED_RDN_SEPARATOR
 when for all DNs other than the root DN.
This is consistent with how the server stores data in dn2id or dn2uri.
Moved RDN.AVA_CHAR_SEPARATOR here, to be consistent with NORMALIZED_* constants.
Added RDN_CHAR_SEPARATOR and used it throughout this class.
Added valueOf(ByteString), rdn(int) and toUUID().
Moved hasAttributeType() here from server's RDN.
Inlined static method compareTo(DN, DN)

RDN.java:
Added MIN_VALUE and minValue(), and modified code to take care of it.
Moved AVA_CHAR_SEPARATOR to DN class and used it throughout this class.

DNTestCase.java:
Added test cases for using RDN.maxValue() and DN.minValue() with a Map keyed by DNs

DistinguishedNameEqualityMatchingRuleTest.java:
Consequence of the change to DN.toNormalizedByteString()

…tions)

The plugin is moved to it's own repo, as a consequence all source and test files are removed from this repository.

UUIDSyntaxImpl.java:
Implemented getEqualityMatchingRule() and getOrderingMatchingRule().

* makeldif now supports line folding as explained in LDIF RFC 2849 (https://www.ietf.org/rfc/rfc2849.txt)
* Add unit test to ensure non regression

* Add -w, --wrapColumn {wrapColumn} option to makeldif to allow a user
to ask for line wrapping in the result file.
This is a new feature which might be useful for a user since line folding is supported in ldif template files (see OPENDJ-2505).
* Add a unit test to ensure non regression

In order to keep the same behavior as deleted server make-ldif, use a random seed
number to initialize random number generator instead of '1'.

* Prints a correct line number in warning/error messages
The line number variable is used to access the lines array so the
first line number is 0. It's misleading for UI messages beacause file
editors usually starts with line 1.
* Do not try to replace constants on a commented line
* Improve makeldif error messages readability

Move build related sections (build, reporting and profiles) from opendj-sdk-bom/pom.xml to opendj-sdk-parent/pom.xml

Since the plugin has been moved, it's first version has been released.
Make the build more stable by using 1.0.0 released version

Do not meld reactor pom (opendj-sdk-bom) which should be used to declare
module to be build and parent pom.
The only effective difference introduced by this change in the build
process is that the following command:
`cd opendj-sdk-parent && mvn clean install`
will only build opendj-sdk-parent module instead of all opendj-sdk
modules before.
Thanks to Guillaume for the tip.

In order to have a more concise BOM, relocate it to a sub-module and
move the parent POM to the parent directory. Logically, the BOM is still
the parent of the parent POM, but it now no longer needs to include
project management related information such as license, scm,
repositories, etc.

(manually reviewed by Gaetan)

Remove unreachable code paths from toolkit _script-util.bat

Leading and trailing whitespace in RDN attribute values must be escaped
according to RFC 4514, otherwise the RDN is invalid. However, it is
often best practice to be tolerant of malformed data, so we should
really trim off any leading or trailing whitespaces. The AVA class was
already trimming leading whitespace, but not trailing.

… DNs

Strictly speaking RFC 4514 does not allow whitespace before or after RDN
or AVA separators. Whilst it is often good practice to be tolerant of
malformed input, we should not produce it.

The DN class was tolerating whitespace characters around the RDN and AVA
separators, but was not removing them, causing the toString() method to
product an invalid string representation.

This change provides a simple fix: do not cache the user provided
string, but instead regenerate it lazily when toString() is called.
Whilst the fix is simple it does incur a slight performance penalty in
the common case where the input DN string is valid.

According to RFC 4514 the '=' character does not need escaping. In
addition, re-enabled the unit tests for DN.toString() which were
surprisingly commented out.

…difyDN operations

The unit tests exposed several bugs:

* RDN.valueOf() ignored trailing RDN sequences or garbage
* RDN constructors allowed zero AVAs
* RDN constructors allowed AVAs having duplicate attribute types
* AVA decoding allowed OIDs with trailing "." characters.