JavaScript Regular Expression - grouping, one-or-more characters, excluding set character strings

Question

I'm trying to match and replace broken HTML using a regex, but I've done a couple of full circles with grouping and lookbacks and quantifiers. I'm struggling to match every scenario.

JavaScript, because the issue is triggered in a Web client browser HTML editor.

The broken HTML is specific - any text between a closing LI and the closing list UL or OL, that is not properly formed as a list item.

For instance, this piece here, from the greater example underneath:

    </li>
        bbb<strong>bbbb</strong><strong>bbb&nbsp;&nbsp;&nbsp; <span style="text-decoration: underline;"><em>bbbbb</em></span></strong>=0==
</ul>

Here is the full example of where the issue could exist:

<ul>
    <li>1111</li>
    <li>Could be anything here</li>
    <li>aaaa</li>
        bbb<strong>bbbb</strong><strong>bbb&nbsp;&nbsp;&nbsp; <span style="text-decoration: underline;"><em>bbbbb</em></span></strong>=0==
</ul>
<ol>
    <li>more?<li>
    <li>echo</li>
</ol>

This is what I intend the HTML to look like using a match + replace.

<ul>
    <li>1111</li>
    <li>Could be anything here</li>
    <li>aaaabbb<strong>bbbb</strong><strong>bbb&nbsp;&nbsp;&nbsp; <span style="text-decoration: underline;"><em>bbbbb</em></span></strong>=0==
</ul>
<ol>
    <li>more?<li>
    <li>echo</li>
</ol>

A few expressions I've tried are the following, but depending on these (or slight variations), I'm matching too much or not correctly or something:

/<\/li>.*?<\/[ou]l>/mig
/<\/li>([\s\n]*[\w!\.?;,<:>&\\\-\{\}\[\]\(\)~#'"=/]+[\s\n]*)+<\/[ou]l>/mig
/<\/li>([\s\n]*[^\s\n]+[\s\n]*)+<\/[ou]l>/i

Searched for a couple of days on and off, no luck.. I realise I'm probably asking something answered hundreds of times before.

I left the closing tag off the "This is what I intend the HTML to look like" bit.. the last list item should have the
argh! do NOT use **regex** to parse **HTML** bro, use **HTML parser**. Well this might be different case as you are trying to repair **broken HTML** but really does the same issue occure that many times to use **regex**? It also seems like you wanted to remove the closing tag ``, which can be done using simple text editor and `ctrl+f` with replace, which doesn't require any patterns and it shouldn't mess up your HTML. — Tafari, Dec 03 '13 at 08:25

score 0 · Answer 1 · edited May 23 '17 at 12:28

0

it's recommended to use a dom based approach to pocessing html

using jQuery:

$('ul>:not(li)').wrapAll('<li></li>');

edited May 23 '17 at 12:28

Community

1
1

answered Dec 03 '13 at 08:29

actual_kangaroo

5,971
2
31
45

Problem might be that you can not be sure what kind of DOM structure you’ll actually be getting when letting the browser parse invalid HTML. (Although for HTML5 this should be specified explicitly.) – CBroe Dec 03 '13 at 08:39

JavaScript Regular Expression - grouping, one-or-more characters, excluding set character strings

1 Answers1