Introduction to Node.js for Grenoble JavaScript User Group

2. Node.js
Grenoble JS - 2011/10/19

3. Who am I?

4. Who am I?
Jeff Mesnil

5. Who am I?
Jeff Mesnil
Software Developer at Bestofmedia

6. Who am I?
Jeff Mesnil
Software Developer at Bestofmedia
http://jmesnil.net/

7. Who am I?
Jeff Mesnil
Software Developer at Bestofmedia
http://jmesnil.net/
jmesnil@gmail.com

8. Who am I?
Jeff Mesnil
Software Developer at Bestofmedia
http://jmesnil.net/
jmesnil@gmail.com
twitter: @jmesnil

9. Previously

10. Previously
Red Hat / JBoss

11. Previously
Red Hat / JBoss
HornetQ messaging server

12. Previously
Red Hat / JBoss
HornetQ messaging server
Written in Java

13. Previously
Red Hat / JBoss
HornetQ messaging server
Written in Java
Concurrency

14. Previously
Red Hat / JBoss
HornetQ messaging server
Written in Java
Concurrency
Thousands of producers/consumers

15. Previously
Red Hat / JBoss
HornetQ messaging server
Written in Java
Concurrency
Thousands of producers/consumers
Hard to write correct multithreaded code

16. Previously
Red Hat / JBoss
HornetQ messaging server
Written in Java
Concurrency
Thousands of producers/consumers
Hard to write correct multithreaded code
Impossible to prove it is correct

18. «I've learned that you can never believe that a
program using shared-memory concurrency (which
is what threading uses) is working correctly - You
can only discover that it's wrong, but you can never
prove that it's right.
...
you can never let yourself become too confident
about your programming abilities when it comes to
shared-memory concurrency. I would not be
surprised if, sometime in the future, someone comes
up with a proof to show that shared-memory
concurrency programming is only possible in
theory, but not in practice.»
-- Bruce Eckel, Thinking in Java

19. Multithreading is Hard
Let’s Go Shopping!

21. • Erlang, Scala

22. • Erlang, Scala
• Actor Model

23. • Erlang, Scala
• Actor Model
• Clojure

24. • Erlang, Scala
• Actor Model
• Clojure
• Software Transaction Model (STM)

25. node.js ?

26. http://nodejs.org/

27. Evented I/O for V8 JavaScript

28. V8

29. V8
From Google

30. V8
From Google
Open Source JavaScript engine

31. V8
From Google
Open Source JavaScript engine
Used in Google Chrome

32. V8
From Google
Open Source JavaScript engine
Used in Google Chrome
Standalone or embedded into any C++
application

33. V8
From Google
Open Source JavaScript engine
Used in Google Chrome
Standalone or embedded into any C++
application
Fast (for a JavaScript Engine...)

34. JavaScript

35. JavaScript
Node applications are written in JavaScript
(or CoffeeScript...)

36. JavaScript
Node applications are written in JavaScript
(or CoffeeScript...)
Functions are 1st-class citizen

37. JavaScript
Node applications are written in JavaScript
(or CoffeeScript...)
Functions are 1st-class citizen
Well suited for event-driven applications

38. JavaScript
Node applications are written in JavaScript
(or CoffeeScript...)
Functions are 1st-class citizen
Well suited for event-driven applications
No standard I/O library...

39. Event Loop
while (true)
ready_channels = select(io_channels)
for channel in ready_channels
performIO(channel)

41. Single thread (on a single core)

42. Single thread (on a single core)
Register interested I/O events and a callback

43. Single thread (on a single core)
Register interested I/O events and a callback
Called back when the event occurs

44. Asynchronous I/O

45. Read a File
var fs = require('fs');
fs.readFile('/usr/share/dict/words', 'utf8', function (err, data) {
if (err) throw err;
console.log(data);
});

46. TCP Echo Server
var net = require('net');
var server = net.createServer(function (socket) {
socket.setEncoding('utf-8');
socket.write("Echo serverrn");
socket.on('data', function(data) {
console.log(data);
socket.write(data.toUpperCase());
});
});
server.listen(1337, "127.0.0.1");

47. $ telnet localhost 1337
Echo server
hello, World!!
HELLO, WORLD!!

48. node.js is good for

49. node.js is good for
• I/O-bound Code

50. node.js is good for
• I/O-bound Code
• DB, Web Services connectivity

51. node.js is good for
• I/O-bound Code
• DB, Web Services connectivity
• Data Aggregation

52. node.js is good for
• I/O-bound Code
• DB, Web Services connectivity
• Data Aggregation
• Non-blocking communication

53. node.js is good for
• I/O-bound Code
• DB, Web Services connectivity
• Data Aggregation
• Non-blocking communication
• Do a bit of I/O and let the others play

54. node.js is not good for

55. node.js is not good for
• CPU-bound applications

56. node.js is not good for
• CPU-bound applications
• Computation (Fibonacci...)

57. node.js is not good for
• CPU-bound applications
• Computation (Fibonacci...)
• Machine-Learning algorithm

58. node.js is not good for
• CPU-bound applications
• Computation (Fibonacci...)
• Machine-Learning algorithm
• ...

59. node.js is not good for
• CPU-bound applications
• Computation (Fibonacci...)
• Machine-Learning algorithm
• ...
• Everybody has to wait its turn

61. Workaround is to start new processes via
child_process.fork()

62. Workaround is to start new processes via
child_process.fork()
Data communication?

63. Workaround is to start new processes via
child_process.fork()
Data communication?
Load?

64. Workaround is to start new processes via
child_process.fork()
Data communication?
Load?
Web Workers API (or Isolates à la Dart)?

65. Google Scraper

66. Google Scraper
• For a query

67. Google Scraper
• For a query
• Connect to Google and fetch a result
page

68. Google Scraper
• For a query
• Connect to Google and fetch a result
page
• Use JQuery to find all the anchors for
the results

69. Google Scraper
• For a query
• Connect to Google and fetch a result
page
• Use JQuery to find all the anchors for
the results
• Print their URLs

70. $ node scraper.js "node javascript"
http://nodejs.org/
http://www.theregister.co.uk/2011/03/01/
the_rise_and_rise_of_node_dot_js/
http://www.howtocreate.co.uk/tutorials/javascript/
dombasics
http://net.tutsplus.com/tutorials/javascript-ajax/learning-
serverside-javascript-with-node-js/
...

71. var request = require('request'),
jsdom = require('jsdom');
var args = process.argv.slice(2);
if (args.length != 1) {
console.log('Usage: node scraper.js <query sentence>');
process.exit(1);
}
var query = escape(args[0]);
var url = 'http://www.google.fr/search?hl=fr&source=hp&ie=ISO-8859-1&q=' + query + '&btnG=Recherche+Google'
request({ uri: url }, function (error, response, body) {
if (error) {
console.log('Error when contacting google.com: ' + error);
process.exit(2);
}
if (response && response.statusCode !== 200) {
console.log('Did not get a correct response from Google: ' + response);
process.exit(2);
}
jsdom.env({
html: body,
scripts: [
'http://code.jquery.com/jquery-1.5.min.js'
]
}, function (err, window) {
var $ = window.jQuery;
$('a.l').each(function() {
console.log($(this).attr('href'));
});
});

72. var request = require('request'),
jsdom = require('jsdom');
var args = process.argv.slice(2);
ica tion
if (args.length != 1) {
a ppl
console.log('Usage: node scraper.js <query sentence>');
p the
u
process.exit(1); set
}
var query = escape(args[0]);
var url = 'http://www.google.fr/search?hl=fr&source=hp&ie=ISO-8859-1&q=' + query + '&btnG=Recherche+Google'
request({ uri: url }, function (error, response, body) {
if (error) {
console.log('Error when contacting google.com: ' + error);
process.exit(2);
}
if (response && response.statusCode !== 200) {
console.log('Did not get a correct response from Google: ' + response);
process.exit(2);
}
jsdom.env({
html: body,
scripts: [
'http://code.jquery.com/jquery-1.5.min.js'
]
}, function (err, window) {
var $ = window.jQuery;
$('a.l').each(function() {
console.log($(this).attr('href'));
});
});

73. var request = require('request'),
jsdom = require('jsdom');
var args = process.argv.slice(2);
ica tion
if (args.length != 1) {
a ppl
console.log('Usage: node scraper.js <query sentence>');
p the
u
process.exit(1); set
}
var query = escape(args[0]);
var url = 'http://www.google.fr/search?hl=fr&source=hp&ie=ISO-8859-1&q=' + query + '&btnG=Recherche+Google'
request({ uri: url }, function (error, response, body) {
if (error) {
console.log('Error when contacting google.com: ' + error);
process.exit(2);
}
if (response && response.statusCode !== 200) {
console.log('Did not get a correct response from Google: ' + response);
es ults
process.exit(2);
og le r
}
h Go
jsdom.env({ fetc
html: body,
scripts: [
'http://code.jquery.com/jquery-1.5.min.js'
]
}, function (err, window) {
var $ = window.jQuery;
$('a.l').each(function() {
console.log($(this).attr('href'));
});
});

74. var request = require('request'),
jsdom = require('jsdom');
var args = process.argv.slice(2);
ica tion
if (args.length != 1) {
a ppl
console.log('Usage: node scraper.js <query sentence>');
p the
u
process.exit(1); set
}
var query = escape(args[0]);
var url = 'http://www.google.fr/search?hl=fr&source=hp&ie=ISO-8859-1&q=' + query + '&btnG=Recherche+Google'
request({ uri: url }, function (error, response, body) {
if (error) {
console.log('Error when contacting google.com: ' + error);
process.exit(2);
}
if (response && response.statusCode !== 200) {
console.log('Did not get a correct response from Google: ' + response);
es ults
process.exit(2);
og le r
}
h Go
jsdom.env({ fetc
y
html: body,
JQ uer
h
wit
scripts: [
rs
cho
'http://code.jquery.com/jquery-1.5.min.js'
an
the
]
}, function (err, window) {
find
var $ = window.jQuery;
$('a.l').each(function() {
console.log($(this).attr('href'));
});
});

75. var request = require('request'),
jsdom = require('jsdom');
var args = process.argv.slice(2);
ica tion
if (args.length != 1) {
a ppl
console.log('Usage: node scraper.js <query sentence>');
p the
u
process.exit(1); set
}
var query = escape(args[0]);
var url = 'http://www.google.fr/search?hl=fr&source=hp&ie=ISO-8859-1&q=' + query + '&btnG=Recherche+Google'
request({ uri: url }, function (error, response, body) {
if (error) {
console.log('Error when contacting google.com: ' + error);
process.exit(2);
}
if (response && response.statusCode !== 200) {
console.log('Did not get a correct response from Google: ' + response);
es ults
process.exit(2);
og le r
}
h Go
jsdom.env({ fetc
y
html: body,
JQ uer
h
wit
scripts: [
rs
cho
'http://code.jquery.com/jquery-1.5.min.js'
an s
the URL
]
}, function (err, window) {
find eir
t th
prin
var $ = window.jQuery;
$('a.l').each(function() {
console.log($(this).attr('href'));
});
});

76. // Node.js modules required to run the scraper
// use npm to install them:
// $ `curl http://npmjs.org/install.sh | sh`
var request = require('request'),
jsdom = require('jsdom');

77. // Node.js modules required to run the scraper
// use npm to install them:
// $ `curl http://npmjs.org/install.sh | sh`
ap per
wr
u est
P req
var request = require('request'), HTT
jsdom = require('jsdom');

78. // Node.js modules required to run the scraper
// use npm to install them:
// $ `curl http://npmjs.org/install.sh | sh`
per
w rap
u est ver
req Ser
TTP the
var request = require('request'), H yo
n
Q uer
u se j
jsdom = require('jsdom');

79. // remove node and the file name from the arguments list
var args = process.argv.slice(2);
if (args.length != 1) {
console.log('Usage: node scraper.js <query sentence>');
process.exit(1);
}
// the query to search on Google
var query = escape(args[0]);
// Google URL that will be request and whose content will be scraped
var url = 'http://www.google.com/search?q=' + query;

80. request({ uri: url }, function (error, response, body) {
// nothing to do if we get an error from Google
if (error) {
console.log('Error when contacting google.com: ' + error);
process.exit(2);
}
// idem if the response is not OK
if (response && response.statusCode !== 200) {
console.log('Did not get a correct response from Google: ' + response);
process.exit(2);
}
...
});

81. po nse
request({ uri: url }, function (error, response, body) { a res
e get
w
// nothing to do if we get an error from Google
w hen
b ack
if (error) { call
console.log('Error when contacting google.com: ' + error);
process.exit(2);
}
// idem if the response is not OK
if (response && response.statusCode !== 200) {
console.log('Did not get a correct response from Google: ' + response);
process.exit(2);
}
...
});

82. po nse
request({ uri: url }, function (error, response, body) { a res
e get
w
// nothing to do if we get an error from Google
w hen
b ack
if (error) { call
console.log('Error when contacting google.com: ' + error);
process.exit(2);
}
// idem if the response is not OK
if (response && response.statusCode !== 200) {
console.log('Did not get a correct response from Google: ' + response);
process.exit(2);
} b ody
in the
li nks
the
find
et’s
...
l
});

83. // use jsdom and jQuery to manipulate the DOM corresponding to the response's body
jsdom.env({
html: body,
scripts: [
// jQuery is loaded directly from its web site.
// We could instead cache it locally for efficiency
'http://code.jquery.com/jquery-1.5.min.js'
]
}, function (err, window) {
...
});

84. // use jsdom and jQuery to manipulate the DOM corresponding to the response's body
y
jsdom.env({
jQ uer
up
o set
html: body,
do mt
s
scripts: [ use j
// jQuery is loaded directly from its web site.
// We could instead cache it locally for efficiency
'http://code.jquery.com/jquery-1.5.min.js'
]
}, function (err, window) {
...
});

85. // use jsdom and jQuery to manipulate the DOM corresponding to the response's body
y
jsdom.env({
jQ uer
up
o set
html: body,
do mt
s
scripts: [ use j
// jQuery is loaded directly from its web site.
// We could instead cache it locally for efficiency
'http://code.jquery.com/jquery-1.5.min.js'
]
}, function (err, window) {
yc ode
uer
... iliar jQ
m
oe s fa
});
he re g

86. // alias jQuery object to the more familiar '$'
var $ = window.jQuery;
// print the URLS of the results listed by Google...
$('a.l').each(function() {
console.log($(this).attr('href'));
});

88. Russian Dolls

89. Spaghetti Code?

90. No Silver Bullet...

91. Another arrow in your quiver!

92. «Ruby, Javascript, and Clojure are all general-purpose
languages, but they each excel at certain use cases. Ruby's
highly dynamic nature and emphasis on beauty makes it a
natural fit for user-facing web apps. Node.js's evented
concurrency makes it a great fit for the realtime
web. Clojure covers a new use case on the Heroku platform:
components which demand correctness, performance,
composability; and optionally, access to the Java ecosystem.»
-- http://blog.heroku.com/archives/2011/7/5/
clojure_on_heroku

93. https://github.com/jmesnil/node-samples
https://groups.google.com/forum/?hl=fr#!
forum/grenoblejs

94. Questions?

95. Thanks!