ESP Pages

ESP Pages are HTML pages with embedded C code that renders what is sent to the client at runtime. Whereas static HTML pages must be fixed at development time, ESP pages can generate HTML and data at runtime and on a per-user basis. ESP provides one of the easiest and most powerful web frameworks without compromising speed or memory footprint.

Embedded C Code

ESP pages are in essence, HTML pages with embedded C code. They are parsed into pure C code and then compiled into native machine binary for exceptionally fast execution.

<h1>Hello World</h1>
<h2>Today is <%= mprGetDate(0); %></h2>

Why C?

Why C code, you may ask? Because C enables ESP to provide the ultimate in speed and efficiency. C is the fastest, closest to the hardware, language available. By using C, invoking device APIs and other system APIs are easy, fast and compact. There is no translation or coupling required getting into and out of another language.

Transparent Compiling

However, ESP is not a traditional low-level environment. When a client request for an ESP page is received, the ESP page is transparently parsed, converted into C code, compiled and saved as a native code shared library. Once loaded, subsequent requests will use the in-memory code without any access to the original ESP page. If the system is rebooted, the on-disk cached module will be loaded without recompilation.


The compilation of ESP pages happens automatically and quickly in the background when in development mode. If the original web page is modified, the page is transparently recompiled and re-cached. For production, pages can be pre-compiled, and locked-down without recompiling in the field. This updating behavior can be controlled by the http.update configuration directive.

Alternatively, ESP pages can be precompiled via the esp generator program. For example:

$ esp compile

In fact, the entire application can be compiled into a single combined module by compiling in combined mode. In this mode, all the ESP pages, controllers and source code can be compiled into a single, manageable flexible library file.

Garbage Collection

ESP uses Garbage Collection for managing memory. This makes working with ESP pages and embedded C code simpler and more secure. Any memory allocated by ESP is automatically managed by the garbage collector. You do not need to free such memory. You can still use the normal malloc and free APIs if you wish. Just don't mix the two kinds of memory.

The use of garbage collection enables the use of decorator patterns and API chaining where the outputs of one function can be used as arguments to others. If you need to manually free memory, this style of programming is often quite difficult, ... but with ESP, it is much easier. For example:

Resource name: <%= mprJoinExt(mprJoinPath(dir, supper(name)), "jpg") %> 

This joins the directory with an upper-case of "name" and appends a "jpg" extension if one is not already in the name. All allocated memory is automatically freed ensuring no memory leaks.

ESP Page Directives

ESP pages support a suite of page directives to control how embedded C code will be processed and at what scope in the generated page the code should be positioned. The following directives are provided:

Directive Purpose
<%^ include "file" %> Include the named file at this location
<%^ global ... Specify that the code from this block must be at a "global" C scope. This is useful for global declarations, extern statements and #include statements.
<%^ start ... Specify that the code from this block will be positioned at the start of the generated function for the template. This is useful for local variable declarations.
<%^ end ... Specify that the code from this block will be positioned at the end of the generated function for the template.
<%= expression %> C expression
<%= [%fmt] expression %> Formatted C expression. The expression is passed through a "printf" style formatter. Use "%S" for HTML-escaped safe strings.
<% statements %> C statements
-%> Trim the trailing newline for the generated block.
%$name Request parameter or session variable. This is equivalent to "renderVar(variable)"
%!variable C string variable
%#field Record field variable. This is equivalent to "ediGetField(getRec(), field))"
%~ Top level application URL. Use this to create relative URLs without having to hard code URI prefixes in your application.

<%^ include "file" %>

This directive can be used to include other HTML or ESP files at this location. The included contents will be parsed as an ESP page and the results substituted at this location. ESP include directives can be nested to an arbitrary depth.

<%^ global ....

This directive specifies that the code from this block must be at a "global" C scope. This is useful for global declarations, extern statements and #include statements.

<%^ start ...

This directive specifies that the code from this block will be positioned at the start of the generated function for the template. Some C compilers do not allow variable declarations to be placed other than at the start of a block. So this directive is useful for local variable declarations.

<%^ end ...

This directive specifies that the code from this block will be positioned at the end of the generated function for the template.

<%= expression %>

Use this directive to embed the result of a C expression. The result must be a null-terminated string. This example below calls the mprGetDate function which returns the local date and time as a string.

Today is <%= mprGetDate(0); %><

<% statements %>

Use this directive to run C code statements. No value is substituted back into the page in place of the directive. If you need to write data inside a statement block, use the render API.

    for (i = 0; i < 10; i++) {
        render("I is %d <br/>\n", i);

Security Note: you should never use render to write back user data that has not been validated. If you must render user input data, use renderSafe which will escape any HTML sequences. This is essential to avoid XSS and other security issues.

You can also iterate over regular HTML code. The following code will display Hello World ten times in the web page.

<% for (i = 0; i < 10; i++) { %>
    Hello World
<% } %>


ESP provides a shorthand for accessing request parameters. Request parameters are a collection of query variables, posted form fields, routing tokens, session variables and any explicitly set parameters. These are stored in a params table and in session data stores and are accessible via the param, espGetParam and espGetSession APIs. However, the %$variable directive provides an easier way. With this directive, the params and session data are examined (in that order) for the given variable name.

In this example below if the URI: was used the following ESP template could be used to access the "name" query value.

<h2>My name is %$name</h2>

This directive is very useful for accessing any request, form, query or session data.


ESP provides a shorthand for accessing C string variables inline.

    char *weather = "fine";
<h2>Current weather is %!fine</h2>

This is equivalent to:

<%= fine %>


ESP provides a shorthand for accessing field data in the current database record. Controllers will often setup the environment for a view by defining a "current" database record. The %#field directive provides an easy way to access any field in the current record.

If a controller defined a current "user" record, this example below would extract the "name" field for display.

<h2>My name is %#name</h2>

Request Context

The full HTTP request context is provided by a set of objects. These include:

The Rx, Tx, Host, Route, EspRoute and EspReq objects are all addressable from the HttpStream object. Here is a collection of especially useful context items:

getStream()Get the current HttpStream object. Can be used inside controllers and template web pages
stream->rxRequest object
stream->txTransmit object
stream->hostHost object
stream->rx->routeRoute object
stream->rx->route->erouteESP Extended Route object
rx->methodHTTP method string
rx->uriRequest URI
rx->pathInfoRequest path information after the scriptName
rx->scriptNameScript name portion of the request URI
tx->filenameResource filename being served
tx->extResource filename extension
tx->finalizedSet if response output has been completed
route->documentsDocument root directory for the route
route->homeRoute home directory

See the Native APIs document for more details about the available APIs. See the ESP Generator documentation for full details.

Common Tasks

Writing Data

ESP provides a set of rendering APIs to send formatted and unformatted data in the response. The most common is render which uses printf style formatting before sending the data to the client. For example:

render("Today, the temp is %d", temp);

Other useful render routines include:


To redirect the client to a new URL, use the redirect API. For example:


For more control over the HTTP status code, use espRedirect API. For example:

espRedirect(getStream(), 302, "checkout.html");

Setting the HTTP Response Code

Use the setStatus API to set the HTTP status response code. For example:

setStatus(404); /* Not found */

Adding a HTTP Header

To add a custom HTTP response header, use the setHeader API. For example:

setHeader("X-ACCESS_CODE", "42");


You can inject debugging code into your views to trace data to the esp log file or back to the client's browser. The mprLog function will print its arguments to esp log file.

To send trace to the browser, use the render function to write data back to the browser.

© Embedthis Software. All rights reserved.