Emscripten Tutorial

Using Emscripten is, at a base level, fairly simple. This tutorial takes you through the steps needed to compile your first Emscripten examples from the command line. It also shows how to work with files and set the main compiler optimization flags.

Tip

Check out this topic if you want to use Emscripten with Microsoft Visual Studio 2010.

First things first

Make sure you have downloaded and installed Emscripten (the exact approach for doing this will depend your operating system: Linux, Windows, or Mac).

Emscripten is accessed using the Emscripten Compiler Frontend (emcc). This script invokes all the other tools needed to build your code, and can act as a drop-in replacement for a standard compiler like gcc or clang. It is called on the command line using ./emcc or ./em++.

Note

On Windows the tool is called using the slightly different syntax: emcc or em++. The remainder of this tutorial uses the Linux approach (./emcc).

For the next section you will need to open a command prompt:

  • On Linux or Mac OS X, open a Terminal.
  • On Windows open the Emscripten Command Prompt, a command prompt that has been pre-configured with the correct system paths and settings to point to the active Emscripten tools. To access this prompt, type Emscripten in the Windows 8 start screen, and then select the Emscripten Command Prompt option.

Navigate with the command prompt to the SDK root directory for your target SDK. This is an SDK-version-specific folder below the emsdk root directory, for example <emsdk root directory>/emscripten/1.20.0/.

Note

The tests should be compiled from the “SDK Root” directory. This is required because some tests load files, and the locations of these files within Emscripten’s virtual file system root is relative to the current directory at build time.

Verifying Emscripten

If you haven’t run Emscripten before, run it now with:

./emcc -v

If the output contains warnings about missing tools, see Verifying the Emscripten Development Environment for debugging help. Otherwise continue to the next sections where we’ll build some code.

Running Emscripten

You can now compile your first C/C++ file to JavaScript.

First, lets have a look at the file to be compiled: hello_world.c. This is the simplest test code in the SDK, and as you can see, all it does is print “hello, world!” to the console and then exit.

#include<stdio.h>

int main() {
  printf("hello, world!\n");
  return 0;
}

To build the JavaScript version of this code, simply specify the C/C++ file after emcc (use em++ to force compilation as C++):

./emcc tests/hello_world.c

There should now be an a.out.js file in the current directory. Run it using node.js:

node a.out.js

This prints “hello, world!” to the console, as expected.

Tip

If an error occurs when calling emcc, run it with the -v option to print out a lot of useful debug information.

Note

In this section, and later on, we run some files from the tests/ folder. That folder contains files for the Emscripten test suite. Some can be run standalone, but others must be run through the test harness itself, see Emscripten Test Suite for more information.

Generating HTML

Emscripten can also generate HTML for testing embedded JavaScript. To generate HTML, use the -o (output) command and specify an html file as the target file:

./emcc tests/hello_world.c -o hello.html

Open the web page in a web browser. As you can see, the framework defines a text area for displaying the output of the printf() calls in the native code.

The HTML output isn’t limited just to just displaying text. You can also use the SDL API to show a colored cube in a <canvas> element (on browsers that support it). For an example, build the hello_world_sdl.cpp test code and then refresh the browser:

./emcc tests/hello_world_sdl.cpp -o hello.html

The source code for the second example is given below:

#include <stdio.h>
#include <SDL/SDL.h>

#ifdef __EMSCRIPTEN__
#include <emscripten.h>
#endif

extern "C" int main(int argc, char** argv) {
  printf("hello, world!\n");

  SDL_Init(SDL_INIT_VIDEO);
  SDL_Surface *screen = SDL_SetVideoMode(256, 256, 32, SDL_SWSURFACE);

#ifdef TEST_SDL_LOCK_OPTS
  EM_ASM("SDL.defaults.copyOnLock = false; SDL.defaults.discardOnLock = true; SDL.defaults.opaqueFrontBuffer = false;");
#endif

  if (SDL_MUSTLOCK(screen)) SDL_LockSurface(screen);
  for (int i = 0; i < 256; i++) {
    for (int j = 0; j < 256; j++) {
#ifdef TEST_SDL_LOCK_OPTS
      // Alpha behaves like in the browser, so write proper opaque pixels.
      int alpha = 255;
#else
      // To emulate native behavior with blitting to screen, alpha component is ignored. Test that it is so by outputting
      // data (and testing that it does get discarded)
      int alpha = (i+j) % 255;
#endif
      *((Uint32*)screen->pixels + i * 256 + j) = SDL_MapRGBA(screen->format, i, j, 255-i, alpha);
    }
  }
  if (SDL_MUSTLOCK(screen)) SDL_UnlockSurface(screen);
  SDL_Flip(screen); 

  printf("you should see a smoothly-colored square - no sharp lines but the square borders!\n");
  printf("and here is some text that should be HTML-friendly: amp: |&| double-quote: |\"| quote: |'| less-than, greater-than, html-like tags: |<cheez></cheez>|\nanother line.\n");

  SDL_Quit();

  return 0;
}

Using files

Note

Your C/C++ code can access files using the normal libc stdio API (fopen, fclose, etc.)

JavaScript is usually run in the sandboxed environment of a web browser, without direct access to the local file system. Emscripten simulates a file system that you can access from your compiled C/C++ code using the normal libc stdio API.

Files that you want to access should be preloaded or embedded into the virtual file system. Preloading (or embedding) generates a virtual file system that corresponds to the file system structure at compile time, relative to the current directory.

The hello_world_file.cpp example shows how to load a file (both the test code and the file to be loaded shown below):

#include <stdio.h>
int main() {
  FILE *file = fopen("tests/hello_world_file.txt", "rb");
  if (!file) {
    printf("cannot open file\n");
    return 1;
  }
  while (!feof(file)) {
    char c = fgetc(file);
    if (c != EOF) {
      putchar(c);
    }
  }
  fclose (file);
  return 0;
}

==
This data has been read from a file.
The file is readable as if it were at the same location in the filesystem, including directories, as in the local filesystem where you compiled the source.
==

Note

The example expects to be able to load a file located at tests/hello_world_file.txt:

FILE *file = fopen("tests/hello_world_file.txt", "rb");

We compile the example from the directory “above” tests to ensure that virtual filesystem is created with the correct structure relative to the compile-time directory.

The following command is used to specify a data file to preload into Emscripten’s virtual file system — before running any compiled code. This approach is useful because Browsers can only load data from the network asynchronously (except in Web Workers) while a lot of native code uses synchronous file system access. Preloading ensures that the asynchronous download of data files is complete (and the file is available) before compiled code has the opportunity to access the Emscripten file system.

./emcc tests/hello_world_file.cpp -o hello.html --preload-file tests/hello_world_file.txt

Run the above command, then open hello.html in the Firefox web browser to see the data from hello_world_file.txt being displayed.

Note

Unfortunately Chrome and Internet Explorer do not support file:// XHR requests, and can’t directly load the local file in which preloaded data is stored. For these browsers you’ll need to serve the files using a webserver. The easiest way to do this is to use the python SimpleHTTPServer (in the current directory do python -m SimpleHTTPServer 8080 and then open http://localhost:8080/hello.html).

For more information about working with the file system see the File System Overview, File System API and Synchronous Virtual XHR Backed File System Usage.

Optimizing code

Emscripten, like gcc and clang, generates unoptimized code by default. You can generate slightly-optimized code with the -O1 command line argument (run the test code from the SDK root directory):

./emcc -O1 tests/hello_world.cpp

The “hello world” code created in a.out.js doesn’t really need to be optimized, so you won’t see a difference in speed when compared to the unoptimized version.

However, you can compare the generated code to see the differences. -O1 applies several minor optimizations and removes some runtime assertions. For example, printf will have been replaced by puts in the generated code.

The optimizations provided by -O2 (see here) are much more aggressive. If you run the following command and inspect the generated code (a.out.js) you will see that it looks very different:

./emcc -O2 tests/hello_world.cpp

For more information about compiler optimization options see Optimizing Code and the emcc tool reference.

Emscripten Test Suite and Benchmarks

Emscripten has a comprehensive test suite, which covers virtually all Emscripten functionality. These tests are an excellent resource for developers as they provide practical examples of most features, and are known to build successfully on the master branch.

See Emscripten Test Suite for more information.

General tips and next steps

This tutorial walked you through your first steps in calling Emscripten from the command line. There is, of course, far more you can do with the tool. Below are other general tips for using Emscripten: