Is mb_* necessary to replace single-byte characters from a multibyte string?

Question

Let's say I have an UTF-8 text like this:

âàêíóôõ <br> âàêíóôõ <br> âàêíóôõ

I want to replace <br> with <br />. Do I need to use mb_str_replace or I can use str_replace ?

Consindering < b r / > are all single byte char?

Answer 1

Since str_replace is binary-safe and UTF-8 is a bijective encoding, you can use str_replace, even if search string or replacement contains multi-byte characters, as long as all three parameters are encoded as UTF-8.

That's why there isn't an mb_str_replace function in the first place.

If your encoding is not bijective - i.e. there are multiple representations of the same string, for example < in UTF-7, which can be expressed both as '+ADw-' and '<', you should convert all strings to the same (bijective) encoding, apply str_replace, and then convert the strings to the target encoding.

Answer 2

Reference for manipulating UTF-8 strings safely in PHP. There is no hard-and-fast rule. Some native PHP string functions functions can operate safely on utf-8, some can with care, and some cannot.

There is no mb_str_replace(). Notice the section "UTF-8 Safe Functionality": explode() and str_replace() are safe as long as all three arguments to it are valid UTF-8 strings.