Match highlight is offset after Unicode character

Created by: LinqLover

• Sourcegraph version: Sourcegraph.com
• Platform information: Chrome 93.0.4577.82, Android 12

Steps to reproduce:

1. Go here: https://sourcegraph.com/search?q=context:global+repo:%5Egithub%5C.com/LinqLover/LinqLover%24+contributor+rev:f0308e7b2bbfb12d44e573a3e20bc3ef0755a173&patternType=literal

Expected behavior:

The word contributor in the match should be highlighted.

Actual behavior:

The highlight is contributo (offset by -1). This might be caused by a 📧 Unicode character earlier in this line.

The entire line looks like this:

- 📧 I’m an active contributor to the [Squeak developers mailing list](http://forum.world.st/Squeak-Dev-f45488.html): [Christoph Thiede on squeak-dev]