这几天有在用simple_html_dom抓一些文章。不同网站的编码在国内基本上是gbk gb2312 utf-8。而以gb2312和utf-8居多。
我这一版的simple_html_dom有一个方法 convert_text 是这个样子的。
// PaperG - Function to convert the text from one character set to another if the two sets are not the same.
function convert_text($text)
{
global $debug_object;
if (is_object($debug_object)) {$debug_object->debug_log_entry(1);}
$converted_text = $text;
$sourceCharset = "";
$targetCharset = "";
if ($this->dom)
{
$sourceCharset = strtoupper($this->dom->_charset);
$targetCharset = strtoupper($this->dom->_target_charset);
}
if (is_object($debug_object)) {$debug_object->debug_log(3, "source charset: " . $sourceCharset . " target charaset: " . $targetCharset);}
if (!empty($sourceCharset) && !empty($targetCharset) && (strcasecmp($sourceCharset, $targetCharset) != 0))
{
// Check if the reported encoding could have been incorrect and the text is actually already UTF-8
if ((strcasecmp($targetCharset, 'UTF-8') == 0) && ($this->is_utf8($text)))
{
$converted_text = $text;
}
else
{
$converted_text = iconv($sourceCharset, $targetCharset, $text);
}
}
// Lets make sure that we don't have that silly BOM issue with any of the utf-8 text we output.
if ($targetCharset == 'UTF-8')
{
if (substr($converted_text, 0, 3) == "\xef\xbb\xbf")
{
$converted_text = substr($converted_text, 3);
}
if (substr($converted_text, -3) == "\xef\xbb\xbf")
{
$converted_text = substr($converted_text, 0, -3);
}
}
return $converted_text;
}
来看这一行:
$converted_text = iconv($sourceCharset, $targetCharset, $text);
会引起转码不正确。比如会把gb2312的文字转成:
4月26日在<span style="color:#C03">????</span>公园马术场举行的2014浪琴国际马联场地障碍世界杯中国联赛资格赛上,24岁的韩壮壮不仅拿到零罚分的成绩 ...第7个出场的<span style="color:#C03">??浜?</span>奥运骑手赵志文第一个收获零罚分,用时77秒07 ...
既成的事实了,证明里头的转码功能没有处理好。由于我使用这个simple_html_dom只是想要用来构建dom。我并没有打算花时间去很好地处理这个bug。而是简单地把
$converted_text = iconv($sourceCharset, $targetCharset, $text);
改成
$converted_text = $text;
就行了。思路就是取消它的转码。好吧工作不必纠结,可以继续了。