使用UTF-8字符串的strpos是否安全？

Question

I have a bunch of strings with different charsets. The $charset variable contains the charset of the current string.

$content = iconv($charset, 'UTF-8', $content);

With this done, is it safe to use strpos, strlen, substr etcetera and not their multibyte equivalent? I'm asking this because I use preg_match a lot as well. So if I use PREG_OFFSET_CAPTURE to get the position of a word in the string I can't use that value with mb_substr to remove everything before the word.

Answer 1

That entirely depends on what you want to do. The core strlen and similar functions work on bytes. Every number they accept and return is a byte count or byte offset. The mb_* functions work encoding-aware on characters. All numbers they accept and return are character counts or offsets.

If you have a safe way of getting a byte offset in a string ("safe" meaning the offset is not in the middle of a multi-byte character) and then, for example, crop everything before that offset using substr, that'll work just fine. For instance:

$str     = '漢字';
$offset  = strpos($str, '字');
$cropped = substr($str, $offset);

Works fine.

However, this won't work:

$cropped = substr($str, $offset, 1);

You can't safely cut out a single byte without running the risk of cutting into a multi-byte character.