Function removes css class names from element class attribute.
Supports regexp as a class name (thing I wait from jQuery for years).
Preserves order, leaves no duplicate or trailing spaces in className.
Regexp should be matched fully, not partially.
/foo-.*/ will match 'foo-bar' in 'foo-bar baz'
/foo-/ will not.
^ and $ means start and end of class name chunk, not of the whole className string
i (case-independent) modifier search is supported
m (multiline) modifier has no meaning
g (global) modifier is ignored: all search behaves as global
removeClass(o, c)
o required object whose className is changed
c required class name: regexp or string (or anything that coerces to string, excepting null and undefined)
Returns: nothing
Why arrays? You can stay on string level (and still save a few bytes):
function(o,c){o.className=o.className.replace(/(\S+)\s*/g,function(f,x){return((c.exec?(c.exec(x)||0)[0]:c)==x)?'':f})}