Bug 205477 – Unicode Regex with Emojis corrupts string replacement

• Status: RESOLVED
• Resolution: DUPLICATE
• Component: JavaScriptCore

Created attachment 386160 [details]
Reproduction example

Given a surrogate-pair emoji: https://emojipedia.org/blue-heart/

And a string containing it: foobar💙

And a Regular Expression such as this:

const re = new RegExp('[^a-zA-Z💙]', 'uig')

A replacement such as this on iOS:

"foobar💙".replace(re, '')

Yields a corrupted string:

"foobar�"

It seems that the replacement only preserves the first half of the surrogate pair, even though the regular expression has the unicode flag enabled.

On MacOS Safari, the result is completely missing emojis.

On an alternative browser like Chrome, the result still contains the 💙 emoji. This reproduction case has been known to work before on Safari, but now no longer seems to be the case.

[Radar WebKit Bug Importer]

<rdar://problem/58103878>

[Alexey Proskuryakov]

*** Bug 256948 has been marked as a duplicate of this bug. ***

Here is a simple reproduction demo. This bug is very annoying, so I'd like to know if there is any progress?

https://stackblitz.com/fork/github/wcshds/safari-regex-replace-bug?file=src%2FApp.vue&terminal=dev

[Alexey Shvayka]

This was fixed in https://commits.webkit.org/272928@main.

I've checked that both test cases (https://bugs.webkit.org/show_bug.cgi?id=205477#c0, https://bugs.webkit.org/show_bug.cgi?id=205477#c3) work correctly.

*** This bug has been marked as a duplicate of bug 267011 ***