我在文件中有一组行，其中每行可能代表多行注释。原始开发人员选择的行分隔符是 pilcrow （¶），因为他觉得这永远不会出现在某人的评论中。我现在将它们放入数据库中，并希望使用更典型的行分隔符（尽管应用程序安装程序可能设置了分隔符）。
问题是某些行使用 ISO-8859-1 编码（十六进制 b6），而其他行使用 UTF-8 编码（十六进制 c2b6）。我正在寻找一种优雅的方式来解决这个问题，它比我目前正在做的事情有更好的支持。
这就是我到目前为止的处理方式，但我正在寻找一个更优雅的解决方案：
// Due to the way the quote file is stored, line breaks can either be
// in 2-byte or 1-byte characters for the pilcrow. Since we're dealing
// with them on a unix system, it makes more sense to replace these
// funky characters with a newline character as is more standard.
//
// To do this, however, requires a bit of chicanery. We have to do
// 1-byte replacement, but with a 2-byte character.
//
// First, some constants:
define('PILCROW', '¶'); // standard two-byte pilcrow character
define('SHORT_PILCROW', chr(0XB6)); // the one-byte version used in the source data some places
define('NEEDLE', '/['.PILCROW.SHORT_PILCROW.']/'); // this is what is searched for
define('REPLACEMENT', $GLOBALS['linesep']);
function fix_line_breaks($quote)
{
$t0 = preg_replace(NEEDLE,REPLACEMENT,$quote); // convert either long or short pilcrow to a newline.
return $t0;
}
我会这样做：
define('PILCROW', '¶'); // standard two-byte pilcrow character
define('REPLACEMENT', $GLOBALS['linesep']);
function fix_encoding($quote) {
return mb_convert_encoding($quote, 'UTF-8', mb_detect_encoding($quote));
}
function fix_line_breaks($quote) {
// convert UTF-8 pilcrow to a newline.
return str_replace(PILCROW, REPLACEMENT, $quote);
}
对于每行评论，请先调用fix_encoding，然后调用fix_line_breaks
$quote = fix_encoding($quote);
$quote = fix_line_breaks($quote);