I'm using a search-function for a documentation site which upon selection of search hit shows page with text highlighted (just as a pdf-reader or netbeans would do).
To achive the highlight i use javascript with:
function searchHighlight(searchTxt) {
var target = $('#page').html();
var re = new RegExp(searchTxt, 'gi');
target = target.replace(
re,
'<span class="high">' + searchTxt + '</span>'
);
$('#page').html(target);
}
Problem / Question:
Since page incudes images with filenames based on md5, some searches messes up the image src.
Searching on "1000" will distort the
<img src="53451000abababababa---.jpg"
to
<img src="5334<span class="hl">1000</span>abababab--.jpg">
Is it possible to solve this with regexp, somehow excluding anything anjcent to ".jpg"?
Or would it be possible to, before highligting replace the images with placeholders, and after replace revert back to src?
Example:
- replace all <img *> with {{I-01}}, {{I-02}} etc and keep the real src in a var.
- Do the replace above.
- Revert back from {{I-01}} to the <img src=".."/>
DOM-manipulation is of course an option, but I figure this could be done with regexp somehow, however, my regexp skills are lacking badly.
UPDATE
This code works for me now:
function searchHighlight(searchTxt) {
var stack = new Array();
var stackPtr = 0;
var target = $('#page').html();
//pre
target = target.replace(/<img.+?>/gi,function(match) {
stack[stackPtr] = match;
return '{{im' + (stackPtr++) + '}}';
});
//replace
var re = new RegExp(searchTxt, 'gi');
target = target.replace(re,'<span class="high">' + searchTxt + '</span>');
//post
stackPtr = 0;
target = target.replace(/{{im.+?}}/gi,function(match) {
return stack[stackPtr++];
});
$('#page').html(target);
}
See Question&Answers more detail:
os 与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…