34,79 €
Mehr erfahren.
- Herausgeber: Packt Publishing
- Kategorie: Wissenschaft und neue Technologien
- Sprache: Englisch
MySQL has introduced a Plugin API with its latest version – a robust, powerful, and easy way of extending the server functionality with loadable modules on the fly. But until now anyone wishing to develop a plugin would almost certainly need to dig into the MySQL source code and search the Web for missing bits of the information.This is the first book on the MySQL Plugin API. Written together with one of the Plugin API primary architects, it contains all the details you need to build a plugin. It shows what a plugin should contain and how to compile, install, and package it. Every chapter illustrates the material with thoroughly explained source code examples.Starting from the basic features, common to all plugin types, and the structure of the plugin framework, this book will guide you through the different plugin types, from simple examples to advanced ones. Server monitoring, full-text search in JPEG comments, typo-tolerant searches, getting the list of all user variables, system usage statistics, or a complete storage engine with indexes – these and other plugins are developed in different chapters of this book, demonstrating the power and versatility of the MySQL Plugin API and explaining the intricate details of MySQL Plugin programming.
Das E-Book können Sie in Legimi-Apps oder einer beliebigen App lesen, die das folgende Format unterstützen:
Seitenzahl: 326
Veröffentlichungsjahr: 2010
Ähnliche
Table of Contents
MySQL 5.1 Plugin Development
MySQL 5.1 Plugin Development
Copyright © 2010 Packt Publishing
All rights reserved. No part of this book may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, without the prior written permission of the publisher, except in the case of brief quotations embedded in critical articles or reviews.
Every effort has been made in the preparation of this book to ensure the accuracy of the information presented. However, the information contained in this book is sold without warranty, either express or implied. Neither the authors, nor Packt Publishing, and its dealers and distributors will be held liable for any damages caused or alleged to be caused directly or indirectly by this book.
Packt Publishing has endeavored to provide trademark information about all of the companies and products mentioned in this book by the appropriate use of capitals. However, Packt Publishing cannot guarantee the accuracy of this information.
First published: August 2010
Production Reference: 1190810
Published by Packt Publishing Ltd. 32 Lincoln Road Olton Birmingham, B27 6PA, UK.
ISBN 978-1-849510-60-8
www.packtpub.com
Cover Image by Asher Wishkerman (<[email protected]> )
Credits
Authors
Sergei Golubchik
Andrew Hutchings
Reviewer
Giuseppe Maxia
Acquisition Editor
Sarah Cullington
Development Editor
Swapna Verlekar
Technical Editors
Priya Darwani
Chris Rodrigues
Indexer
Monica Ajmera Mehta
Editorial Team Leader
Akshara Aware
Project Team Leader
Ashwin Shetty
Project Coordinator
Zainab Bagasrawala
Proofreader
Kevin McGowan
Graphics
Geetanjali Sawant
Production Coordinator
Arvindkumar Gupta
Cover Work
Arvindkumar Gupta
About the Authors
Sergei Golubchik started modifying MySQL source code in 1998, and has continued as a MySQL AB employee since 2000. Working professionally with MySQL sources, he has had the opportunity to get to know and extend almost every part of the server code—from the SQL core to the utility functions. He was one of the primary architects of the Plugin API. After working for ten years in the ever-growing MySQL AB, and later in Sun Microsystems as a Principal Software Developer, he resigned to join a small startup company that works on a MariaDB—an extended version of the MySQL server, where he continues to do what he likes most—hack on MySQL, architecting, and developing MySQL/MariaDB Plugin API, making it even more powerful, safe, and easy to use.
He works and lives in Germany, near Cologne, with his lovely wife and two kids.
Andrew Hutchings is currently one of the top MySQL Support Engineers working at Oracle. He came from failing Computer Science at A-Level (British exams for 17-18 year olds) to working on, pretty much, every field of computing. His first development job was as an 8-bit assembly firmware developer for an environment monitoring company. He then went on to become a senior PHP and C/C++ developer as well as a DBA and system administrator for a large UK magazine chain. From there he was snapped up by Sun Microsystems as a MySQL Support Engineer specializing in MySQL Cluster and C/C++ APIs, much of this work involving deep analysis of the MySQL source code. Sun has since been bought by Oracle and Andrew is continuing his role there and was a tutorial speaker at the 2010 O’Reilly MySQL Conference & Expo. In his spare time Andrew is an active community developer of MySQL, MySQL Cluster, Drizzle, and MySQL Data Dumper (mydumper for short) as well as other small, related projects.
I’d like to thank my wife, Natalie, and my children, Tomos and Oliver, for putting up with me while I was writing this book. I also wish to thank my colleagues, ex-colleagues, and others in the MySQL community (you know who you are) for their help and support in getting started with writing a book. And, of course, my co-author Sergei, without whom this whole book would not have been possible.
About the Reviewer
Giuseppe Maxia, a.k.a. The Data Charmer, is the MySQL Community Team Lead at Sun Microsystems. He is an active member of the MySQL community and a long time open source enthusiast. For the past 23 years he has worked in various IT related fields, with focus on databases, object-oriented programming, and system administration. He is fluent in Italian, English, Perl, SQL, Lua, C, Bash, and a good speaker of C++, French, Spanish, and Java.
He works in cyberspace, with a virtual team.
Preface
Plugin based architecture is not something new, many popular software products use it. It is good both for the software product itself—if done properly it forces developers to structure the code and think about clean interfaces, which helps to keep the code maintainable over years—and for the users—as they can extend it without waiting for the vendor or choose from numerous third-party extensions.
History of the Plugin API
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.
Idea of this book
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.
What this book covers
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.
What you need for this book
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.
Who this book is for
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.
Reader feedback
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.
Customer support
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.
Tip
Downloading the example code for this book
You can download the example code files for all Packt books you have purchased from your account at http://www.PacktPub.com. If you purchased this book elsewhere, you can visit http://www.PacktPub.com/support and register to have the files e-mailed directly to you.
Errata
Although we have taken every care to ensure the accuracy of our content, mistakes do happen. If you find a mistake in one of our books—maybe a mistake in the text or the code—we would be grateful if you would report this to us. By doing so, you can save other readers from frustration and help us improve subsequent versions of this book. If you find any errata, please report them by visiting http://www.packtpub.com/support, selecting your book, clicking on the errata submission form link, and entering the details of your errata. Once your errata are verified, your submission will be accepted and the errata will be uploaded on our website, or added to any list of existing errata, under the Errata section of that title. Any existing errata can be viewed by selecting your title from http://www.packtpub.com/support.
Piracy
Piracy of copyright material on the Internet is an ongoing problem across all media. At Packt, we take the protection of our copyright and licenses very seriously. If you come across any illegal copies of our works, in any form, on the Internet, please provide us with the location address or website name immediately so that we can pursue a remedy.
Please contact us at<[email protected]> with a link to the suspected pirated material.
We appreciate your help in protecting our authors, and our ability to bring you valuable content.
Questions
You can contact us at<[email protected]> if you are having a problem with any aspect of the book, and we will do our best to address it.
Chapter 1. Compiling and Using MySQL Plugins
As you progress through this book you will see several examples of how to use MySQL plugins. This chapter is designed to help you compile and install the UDFs (User Defined Functions) and MySQL plugins that will be created in the following chapters. Do not miss it; you will need this knowledge in every single chapter later on.
UDF libraries
MySQL comes with a small utility called mysql_config, which aids the supply of some of the required options to your compiler. In most cases you need:
This will print something such as the following:
Both MySQL plugins and UDFs need to be compiled as shared libraries. How this is done depends on the platform.
Linux
Under Linux, UDFs should be compiled as follows:
The mysql_config in backticks will apply the results for the command as switches to gcc, so the include directories as well as other required build options are automatically inserted. The -shared option tells the compiler that we are creating a shared library and ‑fPIC enables Position Independent Code, which is required for dynamic linking of this shared library.
Mac OS X
Compiling on Mac OS X is very much like compiling on Linux, but the way shared libraries are defined is slightly different:
A bundle is the Mac OS X equivalent of a shared library. If the UDF needs to call functions in the server binary (for example, if it uses the DBUG debugging facility) the command line will need to be:
Windows
Setting up for compiling UDFs in Windows is generally more involved than in other operating systems.
As everywhere, we need to have the required libraries and include files installed. To do this we run the MySQL installer. If you already have MySQL installed, you can use this tool to modify your installation. The following screenshot shows that we have selected Custom to do this, but a complete install will also give the required files:
Now we need to select Developer Components and then C Include Files / Lib Files to have them included in the installation. Once this is done the installer should look similar to this:
Also, you need to have Microsoft Visual Studio installed. There are free express editions available from the Microsoft website, which we can use.
In Visual Studio we need to create a new empty project to put our source code into and set up the build environment:
Then we need to add a source file to this project. We can either create a new .cpp file or add an existing one to a project:
Now we need to modify the project properties to set up everything required to compile the UDF. To start with, inside the General configuration section, we need to set the Configuration Type to a .dll file (a Windows dynamic link library):
Then in the C/C++ section we need to add the MySQL include path to Additional Include Directories:
Finally, we need to create a definitions file that lists the functions from this library which we wish to export for MySQL to use. It may look as follows:
This is then added to the Linker configuration in the Input section under Module Definition File. This gives a hand-typed dialog, so we need to type in the full path to the definitions file we just created:
We can then compile our UDF and, if successful, we will have a brand new .dll file:
Installing a UDF
Now that we have our UDF, we need to install it in the MySQL server. For security reasons MySQL will only load plugins and UDFs from the location defined in the plugin_dir system variable. This variable can only be set during the startup of the MySQL server. By default it is in the lib/mysql/plugin subdirectory inside the directory where MySQL is installed. So we need to put our UDF library there.
We can then tell MySQL to load the library using:
More details on how to use this syntax and how to solve UDF loading errors are in the UDF chapter of this book.
Plugin libraries
Building and installing plugin libraries is very much like building and installing UDFs. The include and library paths are the same but some further build options are needed. This is slightly complicated by the fact that some plugin types (namely Information Schema and Storage Engine plugins) require the MySQL source to be downloaded for the version of the MySQL server you have installed. This is so that the plugin can have access to data and functions that are only "half-public" and are not declared in the installed C header files.
Linux
When compiling on Linux and using just the normal plugin API we can compile in the same way as with UDFs:
Notice that the main difference here is -DMYSQL_DYNAMIC_PLUGIN. This sets up the necessary environment for the plugin at compile time.
For plugins that require access to the MySQL server source, compiling is slightly different (suppose, the MySQL source tree is in /Sources/mysql‑5.1.35):
Typically, such a plugin will be in C++, not C. It is compiled exactly the same way the main server is—without exceptions or runtime type identification. Technically, it could use exceptions, but then it may need to use g++ instead of gcc as a C++ compiler. Either way, it needs extra include paths that point to the include/, regex/, and sql/ directories of the MySQL source tree.
Mac OS X
Just as in the UDF case, compiling plugins on Mac OS X is almost the same as on Linux. You can use the same command line and only replace ‑shared ‑fPIC with ‑bundle or bundle ‑Wl, ‑undefined ‑Wl,dynamic_lookup as explained before.
Windows
In Windows we can compile MySQL plugins that do not require the inclusion of the MySQL source code (everything except Information Schema and Storage Engine plugins) using a process very similar to compiling UDFs.
First, we need to create an empty project file to contain the source and build environment:
We can then add or create a .cpp file containing the source for our plugin:
This project needs to be a .dll, not an executable one. We can set this in the project's Property Pages dialog:
We now need to set up the C/C++ include paths so that the MySQL include path is in them:
This final step is different to compiling the UDFs. We need to add a C/C++ preprocessor definition so that the include files set up everything we need for a MySQL plugin. To do this we simply add MYSQL_DYNAMIC_PLUGIN to the definitions list:
Installing a plugin
Just as with UDFs, our MySQL plugin needs to be in plugin_dir before it can be added to MySQL. Once it is located there the syntax is very simple. All of the details about how to use the plugin are in the plugin itself. So we simply need:
Summary
Using the information in this chapter we should be able to compile all of the UDFs and plugins for this book as well as any others. We should be able to prepare all of the auxiliary files for plugins to be built with configure && make, as a standalone project or as a part of MySQL, either dynamically or statically. We can package them for distributing in the source form that allows the user to build and install the plugin easily.
Chapter 2. User Defined Functions
Way back in 1998, MySQL 3.21 introduced a framework that allowed users to create new SQL functions easily. It made it possible for developers to write their own functions in C/C++ and load them dynamically into the running server. The functions loaded within this framework were called User Defined Functions or UDFs.
Today not much has changed with UDFs, they are more stable and slightly more secure than they used to be, and they can be declared aggregate for use together with GROUP BY queries. However, many UDFs that worked in 1998 with MySQL 3.21.24 would still work at the time of writing in 2010, with MySQL 5.1.47.
