Category Archives: PHP

Install PHP 7 on Ubuntu 14.04

Update: if you are using Ubuntu 15.04, you can have a look at this script by Maulik Mistry.

I just installed PHP 7.0.0-dev (based on PHPNG) on my GNU/Linux box (Ubuntu 14.04) and I found some errors during the procedure. I decided to write this note to help people using the same Ubuntu environment.

Before to start I would like to remind that PHP 7 is still in development and SHOULD NOT BE USED in production environments. I installed to try and experiment the new features of the language. The first stable release of PHP 7 is scheduled by the end of the year, with a projected release date of November 2015.

To install PHP 7 we need to clone the php-src repository, configure and compile. Let’s create a php7 folder in the home directory and clone the project:

mkdir $HOME/php7
cd $HOME/php7
git clone

After that, we need to prepare and configure the compiler. We need to execute the following commands:

cd php-src
./configure \
    --prefix=$HOME/php7/usr \
    --with-config-file-path=$HOME/php7/usr/etc \
    --enable-mbstring \
    --enable-zip \
    --enable-bcmath \
    --enable-pcntl \
    --enable-ftp \
    --enable-exif \
    --enable-calendar \
    --enable-sysvmsg \
    --enable-sysvsem \
    --enable-sysvshm \
    --enable-wddx \
    --with-curl \
    --with-mcrypt \
    --with-iconv \
    --with-gmp \
    --with-pspell \
    --with-gd \
    --with-jpeg-dir=/usr \
    --with-png-dir=/usr \
    --with-zlib-dir=/usr \
    --with-xpm-dir=/usr \
    --with-freetype-dir=/usr \
    --with-t1lib=/usr \
    --enable-gd-native-ttf \
    --enable-gd-jis-conv \
    --with-openssl \
    --with-mysql=/usr \
    --with-pdo-mysql=/usr \
    --with-gettext=/usr \
    --with-zlib=/usr \
    --with-bz2=/usr \
    --with-recode=/usr \

During the execution of configure command, I found a couple of errors.

Error: Your t1lib distribution is not installed correctly.

I fixed using:

sudo apt-get install libt1-dev

Error: Unable to locate gmp.h

I tried to install the libgmp-dev library:

sudo apt-get install libgmp-dev

but the library was already installed, so I search for it:

locate gm.h

and I found it on /usr/include/x86_64-linux-gnu/gmp.h (I’m using a 64bit version).

I tried to symlink it to /usr/include/gmp.h (the default location for include):

ln -s /usr/include/x86_64-linux-gnu/gmp.h /usr/include/gmp.h 

and finally the configure execution was successful. If you find other errors on your environment I suggest to read this post by Maciej Zgadzaj, where he reported a list of possible errors and solutions for Ubuntu systems.

Now it’s time to compile PHP 7 with the following commands:

make install

The first command compile the PHP. I took about 9 minutes to compile PHP 7 on my computer (Intel i5-2500 at 3.3Ghz). The second command install the PHP modules and configuration in $HOME/php7/usr folder.

We have almost done the installation, we just need to create the php.ini file in the $HOME/php7/usr/etc folder. We can easily create it using vi with the following commands:

mkdir $HOME/php7/usr/etc
vi $HOME/php7/usr/etc/php.ini

We can use the following content for the php.ini:


And finally we can test PHP using the command line interface (CLI):

$HOME/php7/php-src/sapi/cli/php -v

You will get a result like this:

PHP 7.0.0-dev (cli) (built: xxx) 
Copyright (c) 1997-2015 The PHP Group
Zend Engine v3.0.0-dev, Copyright (c) 1998-2015 Zend Technologies

If you want to learn about the new features of PHP 7 I suggest to read the following resources:

Documenting APIs using Apigility

One of the cool feature of Apigility is the ability to generate API documentation using a simple UI. The documentation is generated in HTML format, and optionally in Swagger format. The API documentation is reported in Apigility in the top bar, under the menu “API Docs” (Figure 1, using Apigility 1.0.0beta1).

API documentationFigure 1

In order to generate the API documentation you need to insert some desciptions before. All the information to edit are reported in the Documentation tab on each REST or RPC service (Figure 2).

REST API documentationFigure 2

For each service and for each HTTP method, you can specify a description of the action. In case of RESTful services you can also specify different information for an Entity and a Collection. An interesting feature of the API documentation is the ability to generate the Response Body specification from the configuration, using the “generate from configuration” button (Figure 3).

Generate from documentationFigure 3

This button read the configuration of the API and propose a JSON response based on the fields specified (the fields are documented under the Fields tab of each REST and RPC service). Of course, you can edit the response body changing the output, if you need.

Once you have added some API descriptions, you can go to the “API Docs” menu and show the API documentation (in our case version 1, Figure 4).

API documentation in HTML formatFigure 4

You will see all the API documentation in HTML format, using the Bootstrap 3 template.
You can expand and collapse the information on each HTTP method clicking on the name. All the API documentation are exposed in the /apigility/documentation base URL.

How to install the Swagger adapter

To activate the Swagger adapter for the API documentation, you need to add the following dependency in the composer.json file (in the require field):

"zfcampus/zf-apigility-documentation-swagger": "~1.0-dev"

and execute the composer update commmand.

After the installation of zf-apigility-documentation-swagger you need to enable this module in the config/application.config.php file. You have to edit this configuration file and add the following line after the ‘ZF\Apigility\Documentation’:


Now you can go to the Swagger documentation from the welcome screen, clicking on the Swagger API documentation button, or going directly to the /apigility/swagger URL.
To show the Swagger UI render you have to select the API service version and you will see a web page like the one reported in Figure 5 (using Swagger UI).

API documentation in Swagger formatFigure 5

Customizing the API documentation module

The API documentation feature is offered by Apigility using the zf-apigility-documentation module, written in Zend Framework 2. This module provide an object model of all captured documentation information, including:

  • All APIs available
  • All Services available in each API
  • All Operations available in each API
  • All required/expected Accept and Content-Type request headers, and expected Content-Type response header, for each available API Service Operation.
  • All configured fields for each service

Moreover, it provides a configurable MVC endpoint for returning documentation

  • documentation will be delivered in a serialized JSON structure by default
  • end-users may configure alternate/additional formats via content-negotiation

If you want to customize the format of your API documentation you can have a look at the source code of the zf-apigility-documentation-swagger module. Basically, you need to create a custom route for your format (see the Swagger module.config.php) and use the ZF\Apigility\Documentation\ApiFactory to access the data for the API documentation services. The view model to implement needs to manage a list view and a show view, that’s it.

All the API documentation formats are driven by content negotiation (using the zf-content-negotiation module). For instance, to get the API documentation data in Swagger format you can use the content negotiation “application/vnd.swagger+json”.

For example, if you want to retrieve the API documentation data in JSON format you can use the following request (using HTTPie):

http GET http://localhost:8888/apigility/documentation[/api]/[service] 'Accept:application/json'

where [api] is the name of the API and [service] is the name of the REST or RPC service. To get the same result in Swagger format you can use the following request:

http GET http://localhost:8888/apigility/documentation/[api]/[service] 'Accept:application/vnd.swagger+json'

API authentication in PHP with OAuth2 using Apigility

OAuth2 and Apigility In this post, I show how to build an API authentication system in PHP with OAuth2 using Apigility. In my previous post, I already presented Apigility, the new open source project to create and maintain API for PHP projects.

Apigility offers the following API authentication systems: HTTP Basic, HTTP Digest and OAuth2.

OAuth2 is a well know protocol used worldwide, for instance Facebook, Github, Twitter, etc, use this protocol to authenticate their API. Before start the Apigility functionalities of OAuth2 I would like to introduce briefly the core concepts of this protocol.

In the OAuth2 specification (RFC 6749) we have the following definitions:

  • Resource Owner: the User
  • Resource Server: the API
  • Authorization Server: often the same as the API server
  • Client: the Third-Party Application

In Apigility, the Resource Server and the Authorization Server are delivered from the same API server. The OAuth2 protocol can be considered as a framework for authorization. From the abstract of the RFC 6749 we can read:

The OAuth 2.0 authorization framework enables a third-party application to obtain limited access to an HTTP service, either on behalf of a resource owner by orchestrating an approval interaction between the resource owner and the HTTP service, or by allowing the third-party application to obtain access on its own behalf.

The uses cases covered by the OAuth2 framework are:

In all these uses cases, the goal of the OAuth2 protocol is to exchange a token string between the Client and the Resource Server. This token is used to authenticate all the API calls using the Authorization HTTP header. Below is reported an example of the Bearer token (RFC 7650), the most used token type of OAuth2:

Authorization: Bearer RsT5OjbzRn430zqMLgV3Ia

Security considerations of OAuth2

The OAuth2 protocol doesn’t guarantee confidentiality and integrity of the communications. That means you must protect the HTTP communications using an additional layer. One possible solution is the usage of TLS/SSL (https) to encrypt the communiation channel from the client to the server.

OAuth1 supported an authentication mechanism based on HMAC algorithm to guarantee confidentiality and integrity, OAuth2 doesn’t (there’s a Internet-Draft proposal to support MAC token). That’s one of the main concern about the security of OAuth2 and most developers complain about that (for instance, you can read the blog post of Eran Hammer, the ex-lead author and editor of the OAuth specifications).

That said, use always HTTPS for OAuth2!

Setup OAuth2 in Apigility

Before we jump into the different use cases for OAuth2 authentication we need to configure Apigility to use OAuth2. If you don’t have Apigility installed on your computer you can go to the website and follow the instruction reported here.

To use OAuth2 with Apigility you need to go to the dashboard page admin UI and click on the OAuth2 button, you will see a form like that:

OAuth2 configuration

The OAuth2 implementation in Apigility uses the oauth2-server-php library by Brent Shaffer.
Apigility uses a PDO database to store all the information related to the OAuth2 protocol. You can create your database using the schema reported in the file /vendor/zfcampus/zf-oauth2/data/db_oauth2.sql under your Apigility installation folder (zf-apigility-skeleton).

For testing purposes, you can use the example SQLite database that we shipped in the zf-oauth2 module, /vendor/zfcampus/zf-oauth2/data/dbtest.sqlite. In this case, you need to specify the absolute path of the dbtest.sqlite database in the PDO DSN field, using the syntax sqlite://dbtest.sqlite. Moreover, you need to choose the URI for the authentication API in the OAuth2 route field, for instance /oauth.

In the example database we created a client with client_id testclient and client_secret testpass, and a user with username testuser and password testpass. We will use this example data in the following use cases.

All the sensitive data such as client_secret (in the oauth_clients table) and password (in the oauth_users table), are encrypted by Apigility using the bcrypt algorithm. If you want to generate the bcrypt hash value of a plaintext password in PHP, you can use the Zend\Crypt\Password\Bcrypt component of Zend Framework 2.

In order to facilitate the usage of the PDO database for OAuth2 we included a simple PHP script that can be executed from the command line to generate bcrypt hash values. This command line tool is available in the /vendor/zfcampus/zf-oauth2/bin folder. For instance, you can execute the following command to generate the hash of the string “test” with a cost value of 10 (the cost is a parameter of the bcrypt algorithm that indicates the computational time to consume):

php bcrypt.php test 10

You will see an output like that:


The output of the bcrypt algorithm is a string of 60 bytes.

Web-server applications

The Web-server applications scenario is used to authenticate a web application with a third-party service (e.g. imagine you built a web application that needs to consume the API of Facebook). You can authenticate your application using the third-party server with a 3 steps flow as reported in the diagram below:

OAuth2 web server application

The web application send a request (including the client_id and the redirect_uri) to the third-party service asking for an Authorization code (1).
The third-party server show an Allow/Deny page to request the authorization for the access. If the user click on Allow the server send the Authorization Code to the web application using the redirect_uri (2). The web application can now perform a token request passing the client_id, the redirect_uri and the client_secret, to proof that is authorized to perform this request (3). The third-party server send the token as response if the request is valid (4).

Using Apigility we can request an access code using the following 3 steps:

1) Request the authorization code

Using a browser you can request the authorization approval from this page:

http://<apigility URL>/oauth/authorize?response_type=code

You will see a web page like the follow:

OAuth2 authorization access

You can customize this web page using the view file /vendor/zfcampus/zf-oauth2/view/zf/auth/authorize.phtml.

2) Approve the authorization access

If you Approve the authorization access clicking the Yes button, Apigility will redirect you to the URI specified in the redirect_uri passing the authorization code in the query string (code). In our example you will be redirect to the page /oauth/receive as reported below:

OAuth2 access code

3) Request the Bearer token

Now that we have the authorization code we can request an access token with a POST to /oauth passing the authorization code, the client_id, the client_secret and the redirect_uri as reported in the following HTTPie command:

http -f POST http://<apigility URL>/oauth grant_type=authorization_code
redirect_uri=/oauth/receivecode client_id=testclient client_secret=testpass

The OAuth2 server will reply with the token using a JSON structure like that:

    "access_token": "907c762e069589c2cd2a229cdae7b8778caa9f07", 
    "expires_in": 3600, 
    "refresh_token": "43018382188f462f6b0e5784dd44c36f476ccce6", 
    "scope": null, 
    "token_type": "Bearer"

You have 30 seconds to request the access token starting from the time that you get the authorization code.

Finally, we can access the API using the Bearer token in the HTTP header request. For instance, we provided a test resource in zf-oauth2 module, you can access it using the following HTTPie command:

http http://<Apigility URL>/oauth/resource
"Authorization:Bearer 907c762e069589c2cd2a229cdae7b8778caa9f07"

Browser-based applications

This scenario is quite common when you have a Javascript client (e.g. a Single Page Application) that requests access to the API of a third-party server.
In a browser-based application you cannot store the client_secret in a secure way, that means you cannot use the previous workflow. We need to use an implicit grant. This is similar to the authorization code, but rather than an authorization code being returned from the authorization request, a token is returned.

In the following diagram we reported the 2 steps needed for the authentication of browser-based application scenarios:

OAuth2 Browser-based applications

The browser-based application request the authorization page to third-party service (Step 1). This page contains the Allow/Deny buttons to authorize the API access to the application. If the user click on the Allow button the third-party server send the access token using the URI fragment identifier (#access_token in Step 2). The usage of the fragment identifier for the access_token is important here, from a security point of view, because the token is not passed to the server, the scope of the token is only in the client side (the browser).

The browser-based applications scenario is supported by Apigility using the implicit grant type. This grant type is disabled by default and you need to enable it by hand, changing the configuration of allow_implicit to true in the /config/autoload/local.php file:

return array(
    'zf-oauth2' => array(
        // ...
        'allow_implicit' => true,
        // ...

After this change, we can request the access token using the browser-based application 2 steps:

1) Request the authorization token

We need to request the same URL used in step 1 of Web-server application scenario:

http://<Apigility URL>/oauth/authorize?response_type=token

We will see the same web page of the Web-server application scenario asking for the authorization approval.

2) Approve the authorization access

If we approve the authorization access, clicking on Yes, Apigility will send the access token to the redirect_uri using a URI fragment identifier (#access_token).

In our example, we redirect the access token to the /oauth/receive page, reported below:

OAuth2 access token

If you click on the “Click here to read…” you will see the access token appear on the page. This action is performed by a simple javascript code that parse the URL to extract the access_token value. An example of this Javascript code is reported below:

// function to parse fragment parameters
var parseQueryString = function( queryString ) {
    var params = {}, queries, temp, i, l;

    // Split into key/value pairs
    queries = queryString.split("&");

    // Convert the array of strings into an object
    for ( i = 0, l = queries.length; i < l; i++ ) {
        temp = queries[i].split('=');
        params[temp[0]] = temp[1];
    return params;

// get token params from URL fragment
var tokenParams = parseQueryString(window.location.hash.substr(1));

Mobile apps

This OAuth2 scenario is similar to previous for browser-based applications. The only difference is the redirect_uri that in the mobile world can be a custom URI scheme. This allow native mobile apps to interact with a web browser application, opening a URL from a native app and going back to the app with a custom URI.
For instance, iPhone apps can register a custom URI protocol such as “facebook://”. On Android, apps can register URL matching patterns which will launch the native app if a URL matching the pattern is visited.

Below is reported the diagram for the OAuth2 authentication with Mobile apps:

OAuth2 Mobile app

As you can see the flow is a 2 steps authentication mechanism as for the browser-based applications.

Username and password access

This use case can be used to authenticate an API with a user based grants (password grant). The typical scenario includes a Login web page with username and password that is used to authenticate against a first-party API. Password grant is only appropriate for trusted clients. If you build your own website as a client of your API, thenk this is a great way to handle loggin in.

The authentication mechanism is very simple and it is just 1 step (see diagram below).

OAuth2 username and password

The client application send a POST to the OAuth2 server with the username and password values. The OAuth2 server gives the token access as response in JSON format.

Application access

This use case can be used to authenticate against application access, mosty likely in machine to machine scenarios. The OAuth2 grant type for this use case is the client_credential. The usage is similar to the username and password access reported above, the application send a POST request to the OAuth2 server passing the client_id, the client_secret, that acts like the user’s password. The server reply with the token if the client credentials are valid.

Refresh OAuth2 token

The OAuth2 protocol gives you the possibility to refresh the access token generating a new one, with a new life time. This action can be performed using the refresh_token that the OAuth2 server gives as response during the authentication step.

In Apigility you can refresh the access token with a POST to the OAuth2 server endpoint. In our OAuth2 database example we can perform a refresh token using the following command:

http -f POST http://<Apigility URL>/oauth grant_type=refresh_token 
refresh_token=<here the refresh_token> client_id=testclient

The response will be something like that:

    "access_token": "470d9f3c6b0371ff2a88d0c554cbee9cad495e8d", 
    "expires_in": 3600, 
    "scope": null, 
    "token_type": "Bearer"

Revoke OAuth2 token

Recently (August 2013) the IETF published the RFC 7009 about the OAuth2 token revocation. The actual version 0.8 of Apigility doesn’t support the token revocation, we will support this feature soon. Anyway, is still possible to revoke specific access token using the PDO database. All the tokens are stored in the oauth_access_tokens table, if you want to revoke a token you can delete it from the table, with a SQL query like that:

DELETE FROM oauth_access_tokens WHERE access_token="<token to remove>";


In Apigility we started to support OAuth2 authentication from version 0.8. We are still working on some of the feature of this protocol, like the token revocation. We are collecting feedbacks from the community and if you want to share your comments you are more than welcome to join our apigility-user or apigility-dev mailing list. More information are available on the official web site of the project,