This component had nothing to do with being inline.

Also accessing the ID outside the class through a method.

- Added missing semicolons.
- Added and improved comments.
- Removed extra whitespace.
- Removed self-variables, which were of minimal use.

Also change the CSS class name to be more CSS-ish.

Same thing as with tree in previous commit.

The tree nodes are now actual links, so middle-click is handled
by the browser itself, and it's even possible to right-click and
choose "Open in new tab" from menu.

Added small "<" and ">" buttons below the search results dropdown
that switch from one page to another.

Fixes newlines problem in IE7/8.

The resulting documentation may otherwise contain files like
JSON.html and Json.html, which are different in UNIX, but will
cause a conflict when transferred to Windows.

Now doing a case-insensitive file_exists? check even on UNIX.

Moved HTML surrounding footer text to app.rb

The --body-html replaces --append-html.

The canonical URL and description are now added through --head-html -
so they aren't hardcoded and can be changed when JSDuck is used for
something else than ExtJS 4.

The other possibility is pretty useless.

Plus the code is now re-using the @export property for STDOUT export.

That's equivalent to using Ext.define().

See this issue: https://github.com/senchalabs/jsduck/issues/40

As suggested in: https://github.com/senchalabs/jsduck/issues/38

I feel it's a bit better without the scrolling.  Especially when
clicking on a method in search results, in which case the scrolling
was really only a distraction.

A hash (#) was missing at the beginning of URL, so double-clicking
or ctrl-clicking a link was opening a 404-page.

This way when generating the docs, the source files for guides
will also be included, allowing users of the docs to use these
to generate the guides by themselves.

As suggested by Aaron Conran.

The main repo is now moved to senchalabs.

Instead attempt to auto-detect if we are parsing Ext4 code.