How To Install and Configure the Ggcov Web Pages

The ggcov web pages are a set of PHP code which generate a small set of interconnected interactive HTML pages which allow a user to explore test code coverage data using a standard Web browser.

Browser Compatibility

The ggcov PHP pages are compliant HTML4 and (currently) use no Javascript. They have only been tested on the Mozilla Firefox browser, but are designed to be browser agnostic and should work in any standards- compliant browser. I don't know whether they will work in MSIE.

Server Compatibility

The ggcov PHP pages require a web server which

• supports PHP version 4.3 or later, and

• provides the dba extension with the db4 provider (which ggcov uses to read the intermediate database), and

• provides the gd extension (which ggcov uses to generate the bargraphs on the Summary page).

The intermediate database format is designed to be platform- independant so that the web server can be a different platform to the development machine. The pages are intended to be compatible with both PHP 4.3 and PHP 5.

The pages have been tested using PHP 4.3.9 and Apache 2.0 on an i386 Linux machine.

Installing, If You Control the Web Server

If you control the web server, i.e. you have the ability to install software, installation should be easy.

1. install the ggcov-web RPM, e.g. www# rpm -Uvh ggcov-web-fc3-0.4-1.i386.rpm

2. copy the file /usr/share/doc/ggcov-web-VERSION/ggcov.conf into the /etc/httpd/conf.d/ directory (Apache 2) or insert the contents into httpd.conf (Apache 1.3).

3. restart the web server, e.g. www# service httpd restart

4. install one or more tests in subdirectories of /var/cache/ggcov/web/tests/, e.g. dev$ ggcov-webdb *.c dev$ scp ggcov.webdb.tgz root@www:/tmp web# cd /var/cache/ggcov/web/tests web# mkdir foo ; cd foo web# tar -xzf /tmp/ggcov.webdb.tgz

5. point your browser at http://yourserver.com/ggcov/

Note that the web pages do not need write permission for any files in the test directories; the test data is completely read-only.

Installing, If You Don't Control the Web Server

This technique works when you don't control the web server but you do have a shell account and control over some subset of the web site. It assumes that .htaccess files have been enabled. For the purposes of this example let's assume you are a developer on a sourceforge.net project called "foobar"; the prompt "shell$" means commands you run on shell.sourceforge.net and the prompt "dev$" means commands you run on your development machine.

1. from the ggcov build directory, install the PHP files into a directory you own which is outside the directory visible to the web server, e.g.

   shell$ cd software/ggcov-0.4/ shell$ cd php ; make phpdir=/home/groups/f/fo/foobar/php install

2. the subset of those files which need to be seen by the web server are located in the $phpdir/www/ directory, so move, copy or link those into the directory the web server sees, e.g.

   shell$ cd /home/groups/f/fo/foobar/ shell$ mv php/www htdocs/ggcov

3. configure the PHP include path and the ggcov test directory by adding a .htaccess file, e.g.

   shell$ cd /home/groups/f/fo/foobar/ shell$ cat > htdocs/ggcov/.htaccess <EOF DirectoryIndex tests.php php_value include_path .:/home/groups/f/fo/foobar/php/include SetEnv GGCOV_TEST_DIR /home/groups/f/fo/foobar/php/tests EOF

4. install one or more tests in subdirectories of the test directory, e.g.

   dev$ ggcov-webdb *.c dev$ scp ggcov.webdb.tgz shell.sourceforge.net:

   shell$ cd /home/groups/f/fo/foobar/php/tests shell$ mkdir foo ; cd foo shell$ tar -xzf ~/ggcov.webdb.tgz

5. point your browser at http://foobar.sourceforget.net/ggcov/