0 Regexes
0 Try-catch blocks
No browser sniffing
Only 3kb (minified and gzipped)
Jaguar is a new standalone CSS selector engine developed for the Shrike JavaScript library, and basically supersedes Puma.
Jaguar was designed from the ground up for speed and extensibility, though I'll admit it's not as extensible as Puma (because of that and the Puma Compiler, I'm still developing Puma). A fast method for matching elements with against a CSS selector -- commonly used for event delegation -- is built-in and often beats the browser's native matchesSelector implementation.
Jaguar supports most of the CSS3 selectors, including the more esoteric selectors such as :root > div[rel|=junkfood] ~ p:nth-last-of-type(5n-7) (if anyone actually uses that, I'd love to know).
It's syntax is just like Sizzle/Slick/Puma/:
Jaguar.search(String selector[, DOMElement|DOMDocument context])
// Or simply:
Jaguar(String selector[, DOMElement|DOMDocument context])
The Jaguar.search() function always returns a true Array.
Calling Jaguar() as a function does exactly the same thing as Jaguar.search().
Jaguar also has a match() method, mentioned above, which returns a boolean:
Jaguar.match(Jaguar('#id')[0], '#id') // true
Jaguar.match(Jaguar('#cake')[0], '.edible') // hopefully true
Full docs will be available in the docs/ folder and at https://alpha123.github.com/Jaguar/docs/.