XIRUSS-T API and Implementation


com.glaforge.i18n.io Third-party code for determining the text encoding of files.
org.xiruss Top-level classes for the XIRUSS repository HTTP API server.
org.xiruss.demos.simpleui Provides a simple graphical user interface for interacting with a XIRUSS repository.
org.xiruss.repository Defines the base interfaces and implementation classes for XIRUSS repository servers and clients.
org.xiruss.repository.api Defines the repository-level interfaces common to clients and servers.
org.xiruss.repository.client.importers Provides classes that support importing data objects into the repository.
org.xiruss.repository.client.importers.linkmanagement Classes that support the creation of "referent tracking documents" used to manage element-to-element links via the SnapCM dependency mechanism.
org.xiruss.repository.client.importers.xml Provides classes for supporting the import of generic XML documents as well as XSD schemas and XSLT style sheets.
org.xiruss.repository.common Holds implementation classes usable by both clients and servers.
org.xiruss.repository.multiuser.api Defines APIs for multi-user repositories.
org.xiruss.repository.multiuser.server Server-side implementation classes for multi-user server objects.
org.xiruss.repository.multiuserschema_aware API for schema-aware repositories.
org.xiruss.repository.schema_aware Server-side base implementation of a multi-user, schema-aware repository.
org.xiruss.repository.schema_aware.api API for schema-aware repositories.
org.xiruss.repository.server Implementation classes for the core server-side objects that make up a SnapCM repository.
org.xiruss.repository.server.linkmanagement Server-side base implementations of the link-management-specific interfaces.
org.xiruss.repository.server.reuse_support Contains classes that support the management of use-by-reference relationships among versions.
org.xiruss.repository.server.schema_aware Server-side base implementation of the schema-aware repository classes.
org.xiruss.repository.server.storagemanagement Contains classes that manage the storage of Versions that are storage objects (that is, they contain data and act like files).
org.xiruss.repository.server.storagemanagement.specializations Defines specializations of StorageObject that understand the details of specific kinds of data.
org.xiruss.snapcm Defines the SnapCM API.
org.xiruss.snapcm.api Defines the SnapCM API as a set of interfaces and exception types.
org.xiruss.snapcm.api.helpers Defines types that are not part of the core SnapCM data model but that make working with SnapCM objects more convenient.
org.xiruss.util Utility classes that are used by all parts of the code.
org.xiruss.util.xpointer Support for processing and resolution of XPointers.
org.xiruss.xirussrepository.api Defines the API for the XirussRepository.
org.xiruss.xirussrepository.client This package provides client-side proxies that implement all the relevant Repository and XirussRepository APIs and manage all communication with the HTTP Xiruss Repository API (provided by the JettyXirussHttpApiRunner class.
org.xiruss.xirussrepository.server Implementation of the XirussRepository API.
org.xiruss.xirussrepository.server.jetty Implements the necessary Jetty-defined APIs, implementing a Jetty-based HTTP server that implements the Xiruss HTTP REST API.
org.xiruss.xirussrepository.server.jetty.apiserver Implements the Xiruss HTTP API handler and repository-object-type specific Jetty Resources that manage the translation of specific Xiruss Repository objects into HTTP actions that result in XML API response messages (i.e., the data contents of a version or a list of resources or whatever.
org.xiruss.xirussrepository.server.jetty.viewer Implements the Xiruss HTTP "viewer" server, that is the server that allows direct access to repository resources for "viewing", including access to version contents via a direct and obvious URL (i.e., the version ID in any conetext or resource ID within a snapshot context.


The XIRUSS-T system is an educational and demonstration tool that has several primary goals:

The XIRUSS-T code has been implemented using the basic engineering principle of "make it work, make it right, make it fast, in that order". It has been implemented almost entirely using test-driven development and subsequent refactoring. I have also applied the eXtreme Programming principle of "the simplest thing that could possibly work" to the code. In addition, as a demonstration system, I have tried to make the code as direct and clear as possible, often at the expense of even obvious optimization techniques that experienced programmers might use simply by habit. Thus, while I'm reasonably sure that the behavior of the system is correct per the relevant specs or requirements, I cannot claim that the code as written represents the best implementation approach. This is to a large degree an ongoing experiment in CMS implementation by "simplest possible means". If you come across any code that offends you, please report it to me--I do not claim to be an expert programmer, in Java or any other language (except possible XSLT), and am always eager to learn. I have tried to apply fundamental design patterns where appropriate but you know how it goes.

All of this means in part that the XIRUSS-T system has not been optimized in any way for performance or scalability. It is explicitly a toy for demonstration and experimental purposes. A production-quality system that provided the same features would require a significant amount of additional engineering to make it fast, scalable, and reliable. By the same token, I have not done anything in particular to ensure proper support for concurrency or thread safety. XIRUSS-T as currently implemented is essentially a single-user system intended for non-stressful demonstrations. Any actual correct concurrency or multi-user behavior is accidental.

Code Overview

The XIRUSS-T system provides the following sets of features:

Understanding the XIRUSS-T Code

The best way to start understanding how this code works and how it is intended to be used is to look at the unit tests.

The top-level tests XirussRepositoryTest and JettyServerTest show how to interact with the XIRUSS system from the client perspective, that is, how to put things into the repository and get them back out again. This is the best place to start.

The importer tests (xiruss.importers) show how to use the import framework to do importing of files and documents into the XIRUSS system.

The SnapCM semantics are demonstrated and tested through the xiruss.snapcm tests.

See the description for a specific package for more details about what that package does and the design approach behind the code in that package.