Conquering a large JavaScript codebase

JavaScript was originally conceived to write little scripts that could add some interactivity to a web page. Today, it is used for increasingly large applications — in the browser, but also on the server, e.g. using node.js.
Cloud9 IDE is an Integrated Development Environment developed exclusively in JavaScript, both on the client and server-side, counting over 240,000 lines of JavaScript code. Managing a code base this large requires a good architecture, good solid code foundation as well as solid tools. During this talk I will describe the Cloud9 infrastructure as well as the techniques, libraries and tools we are developing to make it scale, and also how language analysis is helping us make sure that our code is rock-solid thoughout all our application.

Transcript

1. a large JavaScript codebase Conquering
2. @sergimansilla
6. Create Share Test Deploy Run/Debug
7. The$easiest$way$to$create$Node.js$apps
8. Ftp WebDAV Heroku Git Language Analysis Vim mode Run/Debug Cloud storage Revisions Search Collaboration Cloudfoundry Azure Testing CLI Bitbucket Plugins
9. Pretty big app
10. More than 1M lines (with semicolons)
11. Of JavaScript
12. Pure madness
13. Why sparta madness?
14. Javascript/Node.js make you more productive
15. But they won’t help you through your growing pains
16. What about some examples
18. Why Node.JS?
20. V8 JavaScript engine
21. Bindings to libuv
22. Best cross-platform I/O library
23. API accessible from JavaScript
24. Async I/O
25. Async I/O is a big deal
26. You can do things like this without blocking
27. You can do things like this without blocking Loop$monkey,$loop! Write$a$‘.’$for$each$read$file Read$this$very$file
29. You probably want something like this Read$this$very$file Write$a$‘.’$for$each$read$file Loop$monkey,$loop! Only$100$items$at$a$ Dme,$thanks
30. Never, ever, ever, EVER block
32. In node, all APIs are asynchronous
33. In node, all APIs are asynchronous?
34. Getting a hash
35. Getting a hash
36. Getting a hash
37. Getting a hash
38. Profile before going live
39. DTrace + node-stackvis http://blog.nodejs.org/2012/04/25/profiling-node-js/
40. Embrace the paradigm
41. Push processing to the client
42. Use workers
43. Use workers serialization < computation BUT only if
44. Use workers worker.postMessage(...) Small objects only!
45. Modularize your code
46. Your application will fail
47. Be prepared
48. Architect github.com/c9/architect
49. “A simple yet powerful plugin system for large-scale node applications”
50. Isolation
51. Security
52. Providers / Consumers
53. Advanced IPC between processes
54. Sourcemint github.com/sourcemint/sm-npm
55. Package manager that wraps package managers
56. Understands any kind of dependency (npm, pip, git, zip...)
57. Easy edit/publish
58. Tooling matters
59. Normal developers
60. JavaScript Developer
62. Tooling matters
63. Tooling matters especially for JavaScript
65. Use the power of...
66. tatic
67. program analysis
69. Iterating using undeclared variable
70. Warning: you are in an anonymous inner function with its own “this” pointer
71. Did you mean length?
72. “The most important thing I have done as a programmer in recent years is to aggressively pursue static code analysis.” John Carmack
73. treehugger.js github.com/ajaxorg/treehugger
74. “The jQuery of AST analysis.”
75. What can you do with it?
76. Low-level tooling
77. Intelligent code completion
78. Complex refactoring
79. Automate everything
80. Human error happens
81. Human error happens ALL THE TIME
82. Spend time on automation scripts
83. Spend time on automation scripts It really pays off
84. Happy coding!
85. Questions?
86. http://github.com/sergi @sergimansilla http://c9.io