grapheme_strlen函数的奇怪行为与一些行结尾

Question

Can anyone explain this weird behavior of the Unicode strlen function in PHP's intl extension?

var_dump(grapheme_strlen("a\r\n")); // (ASCII 'a') length: 3
var_dump(grapheme_strlen("の\r\n")); // length: 2
var_dump(grapheme_strlen("\r\n")); // length: 2

Seems like grapheme_strlen is counting "\\r\\n" (CR LF, which are two separate code points used for line separation on Windows) as a single grapheme, which could be quite reasonable considering the name of the function, but it does it only if the line ending is preceded by a non-ASCII character. Why?

Answer 1

This is a bug. grapheme_strlen should work on the Grapheme Cluster Boundaries defined in Unicode Standard Annex #29 (Unicode Text Segmentation) . The standard clearly says not to break between CR and LF.

If you have a look at the PHP source, grapheme_strlen simply returns the number of characters for ASCII strings.