How to convert from mixed HTML-string/DOM-elements to DOM-elements in Javascript? -

i wish implement following javascript function:

function alltodom(partsarray) {     // magic! }  // called like: var rowobject = alltodom(['<tr>', tdelem1, tdelem2, '<td>xxx</td>', '<td>',      divcontents,'</td></tr>']);  // tdelem1, tdelem2, divcontents dom node objects. 

the thing want work on kinds of combinations of dom nodes , html fragments. long produces valid html of course. unclosed html tags , disallowed element combinations (like <table><div></div>) allowed have undefined behavior.

my first idea concatenate in html string, except in place of dom elements add placeholder comment <!--snoopy-->. above result in following string:

<tr><!--snoopy--><!--snoopy--><td>xxx</td><td><!--snoopy--></td></tr> 

this valid piece of html, next create <div>, assign innerhtml, gather produced dom nodes, , iterate through them , replace <!--snoopy--> respective dom element.

there 2 flaws approach however:

1. adding <tr> child element <div> invalid. don't know if might not break on condition.
2. internet explorer 8 (the least version need support) strips comments when assigning innerhtml.

are there workarounds? possible @ all?

finally found answer: jquery has done dirty work in parsehtml() method. , happen using jquery anyway, me! :)

i checked magic behind scenes, , it's pretty gruesome. first, inspect html (with regexs...) see parent tag need use, , have workaround ie8, apparently preserve comment nodes - if come after text node. comments before first text node lost. , tags affected way too, had no idea about. , there's half dozen other workarounds ie & webkit problems i've never heard of.

so, i'm going leave them right thing, because trying reproduce stuff madness.