Fano Validation Example Application

Web application skeleton using Fano Framework, Pascal web application framework that demonstrate how to use Fano Framework data validation feature.

When you visit url /guestbook, the application middleware will validate if required query string parameter id is supplied and contains alpha numeric character. If failed, validation error will be reported. If success, string nice will be printed.

This project is generated using Fano CLI command line tools to help scaffolding web application using Fano Framework.

Requirement

• Free Pascal >= 3.0
• libcurl development
• Web Server (Apache, nginx)
• Fano Web Framework

Installation

Build

libcurl development package installation

Check if libcurl package for development is installed by running curl-config.

$ curl-config --version

If libcurl installed you will get something like libcurl x.xx.x where x.xx.x is version. For example libcurl 7.47.0 otherwise you get

The program 'curl-config' can be found in the following packages:
 * libcurl4-gnutls-dev
 * libcurl4-nss-dev
 * libcurl4-openssl-dev
Try: sudo apt install <selected package>

In case libcurl not installed, run

$ sudo apt install libcurl4-gnutls-dev

Free Pascal installation

Make sure Free Pascal is installed. Run

$ fpc -i

If you see something like Free Pascal Compiler version 3.0.4, you are good to go.

Clone this repository

$ git clone [email protected]:fanofamework/fano-app.git --recursive

--recursive is needed so git also pull Fano repository.

If you are missing --recursive when you clone, you may find that vendor/fano directory is empty. In this case run

$ git submodule update --init

To update Fano to its latest commit, run

$ git checkout master && git submodule foreach --recursive git pull origin master

Above command will checkout to master branch of this repository and pull latest update from master branch of Fano repository.

Copy *.cfg.sample to *.cfg. Make adjustment as you need in build.cfg, build.prod.cfg, build.dev.cfg and run build.sh shell script (if you are on Windows, then build.cmd).

These *.cfg files contain some Free Pascal compiler switches that you can turn on/off to change how executable is compiled and generated. For complete explanation on available compiler switches, consult Free Pascal documentation.

Also copy src/config/config.json.sample to src/config/config.json and edit configuration as needed. For example, you may need to change baseUrl to match your own base url so JavaScript or CSS stylesheets point to correct URL.

$ cp config/config.json.sample config/config.json
$ cp build.prod.cfg.sample build.prod.cfg
$ cp build.dev.cfg.sample build.dev.cfg
$ cp build.cfg.sample build.cfg
$ ./build.sh

tools/config.setup.sh shell script is provided to simplify copying those configuration files. Following shell command is similar to command above.

$ ./tools/config.setup.sh
$ ./build.sh

By default, it will output binary executable in public directory.

Build for different environment

To build for different environment, set BUILD_TYPE environment variable.

Build for production environment

$ BUILD_TYPE=prod ./build.sh

Build process will use compiler configuration defined in vendor/fano/fano.cfg, build.cfg and build.prod.cfg. By default, build.prod.cfg contains some compiler switches that will aggressively optimize executable both in speed and size.

Build for development environment

$ BUILD_TYPE=dev ./build.sh

Build process will use compiler configuration defined in vendor/fano/fano.cfg, build.cfg and build.dev.cfg.

If BUILD_TYPE environment variable is not set, production environment will be assumed.

Change executable output directory

Compilation will output executable to directory defined in EXEC_OUTPUT_DIR environment variable. By default is public directory.

$ EXEC_OUTPUT_DIR=/path/to/public/dir ./build.sh

Change executable name

Compilation will use executable filename as defined in EXEC_OUTPUT_NAME environment variable. By default is app.cgi filename.

$ EXEC_OUTPUT_NAME=index.cgi ./build.sh

Run

Run with a webserver

Setup a virtual host. Please consult documentation of web server you use.

For example on Apache,

<VirtualHost *:80>
     ServerName www.example.com
     DocumentRoot /home/example/public

     <Directory "/home/example/public">
         Options +ExecCGI
         AllowOverride FileInfo
         Require all granted
         DirectoryIndex app.cgi
         AddHandler cgi-script .cgi
     </Directory>
</VirtualHost>

On Apache, you will need to enable CGI module, such as mod_cgi or mod_cgid. If CGI module not loaded, above virtual host will cause app.cgi is downloaded instead of executed.

For example, on Debian, this will enable mod_cgi module.

$ sudo a2enmod cgi
$ sudo systemctl restart apache2

Depending on your server setup, for example, if you use .htaccess, add following code:

<IfModule mod_rewrite.c>
    RewriteEngine On
    RewriteCond %{REQUEST_FILENAME} !-d
    RewriteCond %{REQUEST_FILENAME} !-f
    RewriteRule ^(.*)$ app.cgi [L]
</IfModule>

and put .htaccess file in same directory as app.cgi file (i.e., in public directory).

Content of .htaccess basically tells Apache to serve existing files/directories directly. For any non-existing files/directories, pass them to our application.

Simulate run on command line

$ cd public
$ REQUEST_METHOD=GET \
  REQUEST_URI=/test/test \
  SERVER_NAME=juhara.com \
  ./app.cgi

tools/simulate.run.sh is bash script that can be used to simplify simulating run application in shell.

$ ./tools/simulate.run.sh

or to change route to access, set REQUEST_URI variable.

$ REQUEST_URI=/test/test ./tools/simulate.run.sh

This is similar to simulating browser requesting this page,for example,

$ wget -O- http://[your fano app hostname]/test/test

However, running using tools/simulate.run.sh allows you to view output of heaptrc unit for detecting memory leak (if you enable -gh switch in build.dev.cfg).

Deployment

You need to deploy only executable binary and any supporting files such as HTML templates, images, css stylesheets, application config. Any pas or inc files or shell scripts is not needed in deployment machine in order application to run.

So for this repository, you will need to copy public, Templates, config and storages directories to your deployment machine. make sure that storages directory is writable by web server.

Known Issues

Issue with GNU Linker

When running build.sh script, you may encounter following warning:

/usr/bin/ld: warning: public/link.res contains output sections; did you forget -T?

This is known issue between Free Pascal and GNU Linker. See FAQ: link.res syntax error, or "did you forget -T?"

However, this warning is minor and can be ignored. It does not affect output executable.

Issue with unsynchronized compiled unit with unit source

Sometime Free Pascal can not compile your code because, for example, you deleted a unit source code (.pas) but old generated unit (.ppu, .o, .a files) still there or when you switch between git branches. Solution is to remove those files.

By default, generated compiled units are in bin/unit directory. But do not delete README.md file inside this directory, as it is not being ignored by git.

$ rm bin/unit/*.ppu
$ rm bin/unit/*.o
$ rm bin/unit/*.rsj
$ rm bin/unit/*.a

Following shell command will remove all files inside bin/unit directory except README.md file.

$ find bin/unit ! -name 'README.md' -type f -exec rm -f {} +

tools/clean.sh script is provided to simplify this task.

Windows user

Free Pascal supports Windows as target operating system, however, this repository is not yet tested on Windows. To target Windows, in build.cfg replace compiler switch -Tlinux with -Twin64 and uncomment line #-WC to become -WC.

Lazarus user

While you can use Lazarus IDE, it is not mandatory tool. Any text editor for code editing (Atom, Visual Studio Code, Sublime, Vim etc) should suffice.