preg_match_all - 正则表达式的贪婪部分，但最大化匹配数

Question

I have the following html to parse:

<h1 class="x">test</h1>
<p>some text <img src="x" /></p>

<h1 class="x1">test2</h1>
<p>some text </p>

<h1 class="2">test3</h1>
<p>some text <img src="x" /></p>

Can I parse this into an array with a single regular expression?

I tried

preg_match_all('#(<h1[^>]*?>)(.*?)(</h1>)(.*)#ism',$html,$arr);

which gives me only one entry, because the last part of the regex is greedy, and

preg_match_all('#(<h1[^>]*?>)(.*?)(</h1>)(.*?)#ism',$html,$arr);

which gives me nothing of the HTML between the <h1>, because the expression is not greedy.

How can I make the part after the be matched greedy, while at the same time matching as many occurences as possible?

Additional comments:

• the question is fairly academical, I have resolved the problem using pre_split and a variety of other methods would work, but may also have downsides (for example DOM may not work on invalid HTML that I cannot control). However it is a recurring problem that I'd be interested to know more about.

Answer 1

You need some form of end maker. The regex can not guess until which part you want to match.

Possible in this case might be a lookahead assertion after the (.*?) at the end:

(?=<h1|</body>|\z)#ims