I have been looking for a CSS selector function other than Sizzle and I have come across this function.
function SparkEn(xpath,root) {
xpath = xpath
.replace(/((^|\|)\s*)([^/|\s]+)/g,'$2.//$3')
.replace(/\.([\w-]+)(?!([^\]]*]))/g, '[@class="$1" or @class$=" $1" or @class^="$1 " or @class~=" $1 "]')
.replace(/#([\w-]+)/g, '[@id="$1"]')
.replace(/\/\[/g,'/*[');
str = '(@\\w+|"[^"]*"|\'[^\']*\')';
xpath = xpath
.replace(new RegExp(str+'\\s*~=\\s*'+str,'g'), 'contains($1,$2)')
.replace(new RegExp(str+'\\s*\\^=\\s*'+str,'g'), 'starts-with($1,$2)')
.replace(new RegExp(str+'\\s*\\$=\\s*'+str,'g'), 'substring($1,string-length($1)-string-length($2)+1)=$2');
var got = document.evaluate(xpath, root||document, null, 5, null);
var result=[];
while (next = got.iterateNext())
result.push(next);
return result;
}
I just feel like it is too good to be true, is this a firefox only function (xpath?) or is it slow? Basically why would I use Sizzle over this?
I believe no stable version of IE supports document.evaluate
, so you're limited to every other browser. It's not slow since it's a native implementation of XPath.
Sizzle is useful because it uses the native support browsers offer when available (such as document.getElementsByClassName), but falls back to doing it itself when unavailable (IE). It's also used by jQuery and Prototype, so it's heavily, heavily tested and is unlikely to give you any trouble. Sizzle is also heavily speed-tested and optimized (they have a whole speed test suite), which is more work you don't have to do.
I'd say go with jQuery, Prototype, or just Sizzle by itself unless you are doing something incredibly performance-sensitive (which, honestly, is probably an indicator that you've structured your application poorly).