Building the Examples
~~~~~~~~~~~~~~~~~~~~~
If you're installing MySQL++ from the source tarball, the example
programs get built when you build the library. If you change
any example code, just say 'make' to rebuild the examples.
The examples are built against the headers and library in the
lib subdirectory, not against the ones you may have installed
elsewhere on the system.
If these example files were installed on your system as part of
the -devel RPM, copy all the files to a directory you can write
to, then say 'make' in that directory. This uses a simplified
Makefile, which builds the examples against the headers and
libraries installed in the system directories.
Getting Started with the Examples
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
MySQL++ is built as a shared library on most systems, and a DLL
on Windows. Since it isn't built in the same directory as the
examples, this means that your system won't be able to find the
library without help until you install it. Since you generally
want to run the examples _before_ installing the library, to test
that the library actually works, we need a workaround.
That workaround is the exrun script. There are two versions,
a Bourne shell script called just exrun for POSIX systems, and
exrun.bat for Windows.
Before running the other examples, you must first create the
sample database. On POSIX systems, you do that like so:
$ ./exrun resetdb [-s server_addr] [-u user] [-p password]
On Windows, that would instead be:
C:\mysql++\> exrun.bat resetdb [-s server] [-u user] [-p pass]
You don't have to give any of these options. If you don't pass -s,
it assumes the database server is running on the same machine,
and so tries to contact the server over some form of local IPC.
If you don't pass -u, it uses your own user name when logging
into to the database server. If you don't pass -p, it assumes the
database user has an empty password, which hopefully is not true.
The -s option accepts many different forms of address. The main
one is some sort of TCP/IP address, with an optional port number
or service name. On Unixy systems, you can give a Unix domain
socket name. On Windows, you can give just a period to use named
pipes, if the server supports it. All of these are legal:
.
localhost
172.20.0.252:12345
/var/run/mysqld.sock
my.server.name.com:mysql
If you give -s but don't give a port number or service name with
it, it assumes the default, port 3306.
Running the Other Command Line Examples
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The following examples use the database set up by resetdb, and
have the same command line format as resetdb:
simple1: Shows the item names for all records in the sample
stock table. Try this one if you do nothing else.
simple2: Similar to simple1, but it displays all columns
from the sample stock table, not just the item name.
simple3: Same as simple2, except that it retrieves the data
with a "use" query instead of a "store" query. See the
user manual for the difference between these methods.
ssqls1-5: These demonstrate the SSQLS features. Read the
SSQLS sections in the user manual for details about
these examples.
multiquery: MySQL++ allows you to issue multiple queries at
once, and get the results as separate sets. This shows
that, and also how to use stored procedures, which return
their results in the same way as a multiquery.
tquery1-3: Shows how to use the template query facility.
transaction: Shows how to use the Transaction class to create
transaction sets which automatically roll back if not
explicitly committed.
deadlock: Shows how to handle errors that happen when the
database server detects a deadlock in a transaction set.
Also demonstrates the need for BadQuery::errnum()
store_if: Demonstrates the Query::store_if() method, which
allows you to store the results of a query in an STL
container conditionally. Think of it as a way to express
rules for selecting records in C++, to be used when SQL's
WHERE clause isn't powerful enough.
for_each: Demonstrates the Query::for_each() method, which
allows you to execute a query and call a functor on each
returned row. This example uses this to gather statistics
on the sample table's contents.
cpoolp, cpoolw: Respectively, the POSIX and Windows threads
variants of the ConnectionPool class demonstration.
load_jpeg: Inserts a JPEG file into the sample database,
for use by the cgi_jpeg example. (See below.) Unlike the
other examples, this one takes anything given on the
command line that isn't a switch to be a JPEG file name.
We've included examples/logo.jpg as a sample, if you want
to use that.
fieldinf: Shows how to get information about the fields in
a result set. (Types, etc.)
dbinfo: Dumps a bunch of information about the database
server and some of the data it's managing.
If you run the load_jpeg example, you should consider also
playing with the other half of the demonstration, cgi_jpeg.
To run it, you'll need to install MySQL++ on a machine with
a web server, then copy the cgi_jpeg program to the server's
CGI directory. For example, on a stock Red Hat type box,
that would be /var/www/cgi-bin. At that point, a request like
http://my.server.com/cgi-bin/cgi_jpeg?id=1 should show the JPEG you
loaded. The ID value to use will be that reported by load_jpeg.
Dedicated Windows Examples
~~~~~~~~~~~~~~~~~~~~~~~~~~
If you're a Visual C++ user, there are two GUI examples, too:
examples\vstudio\mfc.vcproj: MFC-based replacement for
the simple2 example.
examples\vstudio\wforms.vcproj: The same thing, only written in
C++/CLI and using Windows Forms. Please follow the
instructions in README-Visual-C++.txt before trying to
build and run this example. The MySQL++ library build
settings must be modified and the library rebuilt first.
Although these examples show use of MySQL++ in a GUI rather than
command line program, that's not the main point of these examples.
What we're really showing here is how to deal with Unicode.
The MySQL server prefers the UTF-8 encoding for Unicode, which
works naturally with most non-Windows systems. Windows, on the
other hand, uses a different Unicode character encoding, UCS-2.
These examples show how to do the necessary conversions. (See the
Unicode chapter in the user manual for more on this topic.)
We need two different examples because Unicode conversions and
string handling are so wildly different under .NET than with the
native Win32 API. .NET makes these tasks much easier.
These examples build and run as-is under Visual C++ 2005. To make
them work with VC++ 2008, you will have to change several paths
in both project's settings to reference the "vc2008" subdirectory
instead of "vc2005":
o Configuration Properties > Debugging > Working Directory
o Configuration Properties > Linker > General >
Additional Library Directories
If you want to backport these examples to VC++ 2003, it's probably
not hard. The main difficulty is that VS 2003 supports Managed
C++, which isn't the same thing as C++/CLI.
Special exrun Capabilities
~~~~~~~~~~~~~~~~~~~~~~~~~~
The Bourne shell version of the exrun script has a few features
not avaiable in the Windows batch file version. These features
let you run the examples under various debugging tools.
You can get simple gdb debugging if you run an example like this:
$ ./exrun gdb simple1 foo bar qux
The script also supports valgrind, in memory leak testing mode:
$ ./exrun valgrind simple1 foo bar qux