padStart/padEnd in current safari is buggy and shouldnt be used

[trashhalo]

The currently released version of safari has a bug in padStart padEnd that causes it to return null.

Fixed a bug where String.prototype.padStart() and padEnd() would insert null in rare cases

https://developer.apple.com/safari/technology-preview/release-notes/ fix in preview 15

While it says its rare we are hitting it about 40% of the time. As far as I can tell there isn't a externally facing way to force a polyfill to override native. Can you guys mark padStart/end as $export.F until this fix lands?

[benjie]

I was hitting this also on iOS (weirdly it doesn't happen so much when the Developer Tools are connected, but much more when they are not); lost over an hour to finding out what the issue in my code was (not least because I couldn't debug it with the devtools and it was Mobile Safari so I had to resort to alert() based debugging like some kind of Neanderthal).

A buggyfill for this would have saved me a lot of time.

[zloirock]

Is forced replacement for /Version\/10\.\d+(\.\d+)? Safari\//.test(userAgent) enough, I hope?

[zloirock]

Fixed in core-js@2.5.3.

[sunderls]

safari: Mozilla/5.0 (iPhone; CPU iPhone OS 10_2 like Mac OS X) AppleWebKit/602.3.12 (KHTML, like Gecko) Version/10.0 Mobile/14C89 Safari/602.1

webview: Mozilla/5.0 (iPhone; CPU iPhone OS 10_2 like Mac OS X) AppleWebKit/602.3.12 (KHTML, like Gecko) Mobile/14C89 Safari

since there is no version string, maybe we better check /iphone os 10/i?

[zloirock]

@sunderls it will not work with desktop Safari. Feel free to propose better detection.