MySQL used to have "pluggable" extensions in a form of dynamically loaded functions since version 3.21.24 released in February 1998. Despite being quite limited in functionality, they were useful and people were using them. In early 2005, one of the authors of this book together with another MySQL developer, Sergey Vojtovich, were working on loadable parsers for MySQL full-text search, to be able to load a very specialized parser that one of their customers wanted. And Brian Aker, who was MySQL Director of Architecture at that time, suggested creating a unified interface for loadable "modules". Based on this idea we developed the MySQL Plugin API—a generic framework that allowed loading of any functionality in the server—and Full-text Parser plugins were the first plugin type.

Storage Engine API already existed in MySQL at that time—Michael "Monty" Widenius, the original author of MySQL, had it since the very first MySQL version, although he only added the handler class few years later, in 1999. This made Storage Engine plugins an easy target, and we added them as the next plugin type. Soon after that I, and another MySQL developer, Antony Curtis, extended Plugin API with the autotools support, the infamous plug.in file and MYSQL_PLUGIN_* macros that go in it, and implemented support for server variables, MYSQL_SYSVAR_* and MYSQL_THDVAR_* macros. Brian Aker added two more plugin types—Information Schema Table plugins and Daemon plugins.

Life was going on even after MySQL 5.1 was released—Antony Curtis and I have developed Audit plugins. And very recently I and an external contributor, MIT student R.J. Silk, have completed the work on pluggable authentication and Authentication plugins were born.

Meanwhile, Michael "Monty" Widenius had left MySQL and started a new company to work on MySQL fork, that he named MariaDB. Another former MySQL developer, Sanja Byelkin, and I have implemented the latest (at the time of writing) feature in the Storage Engine API, the engine defined attributes in the CREATE TABLE statement.

Today, the MySQL Plugin API is a robust and proven feature. There are many third-party plugins both open and closed source, the most popular being Storage Engines, often accompanied by Information Schema tables, and Full-text parsers.

However, the API documentation is not very helpful. If you are anything like me, you prefer fiction to a dictionary and a few good examples to a grammar description in the Backus-Naur form. The Plugin API documentation describes the functions and the structures but does not show how to use them. Tutorials, on the other hand, help the reader to understand how to use the API. Examples are important to illustrate the concepts and to bootstrap a new plugin project easily.

This is where the idea of this book came from. We wanted to create a book that would allow readers to start writing plugins right away. With detailed tutorials and practical plugin examples, thoroughly explained line by line, highlighted common mistakes and clarified design decisions. And with code samples that you can start using in your projects. Not just the code you can copy, but more importantly, the code you understand—every line, every variable—as if you had written it yourself.

But this book is not a reference manual. It does not contain an exhaustive list of all functions, classes, and macros of the MySQL Plugin API. The API is documented in the header files and in the MySQL manual. But to use it, you need to know what to look for. It is often said that asking the right question is half the right answer. This book teaches you to ask right questions. It gives detailed understanding—not just knowledge—of the MySQL Plugin API, and even if you will not have every piece of the puzzle, you will have most of them, you will know how they fit together, and you will be able to see the whole picture.

The book encourages consecutive reading, but chapters can be read in any order too. They are mostly independent, and, if needed, you can start reading from, for example, storage engine chapters without reading about full-text search parsers or UDFs. The book is structured as follows.

Chapter 1, Compiling and Using MySQL Plugins lays the necessary foundation for the rest of the book, you will need it in all of the following chapters. It describes how to compile, link, and install UDFs and plugins. Even if you are only interested in, say, full-text parsers or storage engines, you may want to read this chapter first. It is not called Read Me First!!! only because we suspected that the editor may not have wanted a lot of exclamation marks in the chapter title.

Chapter 2, User Defined Functions deals with UDFs - these dynamically loaded server extensions that first appeared in the server in 3.21.24, the great-grandparents of the MySQL Plugin API. Although, strictly speaking, UDFs are not MySQL Plugins—not part of the MySQL Plugin API—they are still used to load functionality in the server at runtime, just like plugins are, and sometimes they are used to complement the plugin functionality.

Chapter 3, Daemon Plugins introduces the reader to the MySQL Plugin API. It talks about the most simple plugin type—Daemon plugins. It starts with the basic structure of a plugin—what a plugin declaration should look like, what plugin types are, and so on. Then it describes features common to all plugin types—initialization and de-initialization callbacks, status variables, and configuration system variables. After that it describes and analyzes line by line four Daemon plugin examples—from a simple plugin that prints Hello World! when loaded, to a system monitoring plugin that periodically logs the number of connections, to a system usage status plugin that displays the memory and I/O usage of the MySQL server.

Chapter 4, Information Schema Plugins is dedicated to plugins that add tables to INFORMATION_SCHEMA. It describes all of the necessary data structures and ends with two plugin examples—a simple INFORMATION_SCHEMA table with versions of different MySQL subsystems and system usage statistics presented as an INFORMATION_SCHEMA table.

Chapter 5, Advanced Information Schema Plugins delves more into the topic started in the previous chapter. It explains how to use condition pushdown and how to extract and display information from the server internal data structures. It presents three plugins that demonstrate condition pushdown, list all user variables, and all binary log files.

Chapter 6, Full-text Parser Plugins is about plugins that extend the MySQL built-in full-text search. It describes all of the data structures and the code execution flow and illustrates all that with an example plugin that can parse PHP scripts.

Chapter 7, Practical Full-text Parsers is devoted to the advanced applications of the plugins of this type. It explains how the search in Boolean mode works and contains more plugin examples—an Exif parser that allows users to search within embedded comments in image files, a Soundex parser that post-processes all words with a Soundex algorithm making the search invulnerable to typos and misspelled words, and a Boolean search parser plugin that supports AND and OR operators.

Chapter 8, Storage Engine Plugins starts the discussion about the most complex and versatile plugin type in MySQL. It gives an overview of the main concepts of the Storage Engine API and thoroughly analyzes sources of the very simple read-only storage engine.

Chapter 9, HTML Storage Engine - Reads and Writes continues the Storage Engine series. It presents a storage engine plugin that keeps table data in HTML tables and uses it to explain how to implement an updatable data stores.

Chapter 10, TOCAB Storage Engine - Implementing Indexes concludes the Storage Engine part of the book. In this chapter, we develop a storage engine that supports indexes, using it to explain how the indexing part of the MySQL Storage Engine API works, how to build an engine that uses an external indexing library, and how to work around the incompatibilities of their APIs.

Appendix talks about new MySQL Plugin API features, those that did not make it into MySQL 5.1. It describes Server Services, what they are and why they were introduced, the Audit plugins, the example of a plugin that audits security violations, Authentication plugins, with a plugin that uses USB devices to identify users, and engine attributes in the CREATE TABLE, demonstrating the feature with the help of the storage engine from Chapter 10.

The book assumes basic knowledge of SQL and MySQL in particular, and until MySQL developers implement support for plugins in scripting languages, which would be great but can hardly happen any time soon, a certain level of familiarity with C, and for storage engines C++, will be required.

We wrote this book for people who want to create MySQL plugins. They could be developers with a great idea for a new storage engine. But more often than not they will be application developers that need to solve a specific problem, whether it is searching text within Microsoft Word or Open Office documents, monitoring the database server with their company-wide monitoring framework, querying with SQL the multi-gigabyte files created with a 20 year old custom data storage library and joining them with new relational data, or adding MySQL to the company-wide single sign-on setup. All this and much more can be done with MySQL plugins.

In this book, you will find a number of styles of text that distinguish between different kinds of information. Here are some examples of these styles, and an explanation of their meaning.

Code words in text are shown as follows: "The second argument of the name_init() function is a pointer to the UDF_ARGS structure."

A block of code is set as follows:

typedef struct st_field_info
{
const char* field_name;
uint field_length;
enum enum_field_types field_type;
int value;
uint field_flags;
const char* old_name;
uint open_method;
} ST_FIELD_INFO;

When we wish to draw your attention to a particular part of a code block, the relevant lines or items are set in bold:

static int tocab_init(void *p)
{
handlerton *tocab_hton = (handlerton *)p;
tocab_hton->create = tocab_create_handler;
tocab_hton->table_options = table_option_list;
return 0;
}

Any command-line input or output is written as follows:

shell$ mysql_config --cflags

New terms and important words are shown in italics. Words that you see on the screen, in menus or dialog boxes for example, appear in the text like this: "Then in the C/C++ section we need to add the MySQL include path to Additional Include Directories".

Feedback from our readers is always welcome. Let us know what you think about this book—what you liked or may have disliked. Reader feedback is important for us to develop titles that you really get the most out of.

To send us general feedback, simply send an e-mail to<[email protected]>, and mention the book title via the subject of your message.

If there is a book that you need and would like to see us publish, please send us a note in the SUGGEST A TITLE form on www.packtpub.com or e-mail<[email protected]>.

If there is a topic that you have expertise in and you are interested in either writing or contributing to a book, see our author guide on www.packtpub.com/authors.

Now that you are the proud owner of a Packt book, we have a number of things to help you to get the most from your purchase.