PHP stristr假阳性CDATA

Question

以下HTML/CSS是从Hotmail发送的HTML电子邮件......

<style><!-- 
.hmmessage P 
{ 
margin:0px; 
padding:0px 
} 
body.hmmessage 
{ 
font-size: 12pt; 
font-family:Calibri 
} 
--></style> 

我只是想从里面的风格元素让CSS。有些可能包含HTML注释，如上面的或CDATA。由于一些奇怪的原因，PHP函数返回一个假阳性CDATA低于上述字符串...

if (stristr($b,'<style')) 
{ 
    $s = explode('<style',$b,2)[1]; 
    $s = explode('>',$s,2)[1]; 

    if (stristr($s,'<![CDATA[')) 
    { 
    $s = explode('<![CDATA[',$s,2)[1]; 
    $s = explode(']]',$s,2)[0]; 
    } 
    else if (stristr($s,'<!--')) 
    { 
    $s = explode('<!--',$s,2)[1]; 
    $s = explode('-->',$s,2)[0]; 
    } 
    else 
    { 
    $s = explode('</style>',$s,2)[0]; 
    } 

Answer 1

为什么不直接拿DOMDocument？

$html = " 
<style><!-- 
.hmmessage P 
{ 
margin:0px; 
padding:0px 
} 
body.hmmessage 
{ 
font-size: 12pt; 
font-family:Calibri 
} 
--></style>"; 


$dom = new DOMDocument(); 
$dom->loadHTML($html); 
$style = $dom->getElementsByTagName('style'); 

// get the content from first style tag 
$css = $style->item(0)->nodeValue; 
// clear the comments and cdata tags 
$css = str_replace(array('<!--', '-->', '<![CDATA[', ']]>', '//<![CDATA[', '//]]>'), '', $css); 
echo $css;