This issue is intended to track progress on supporting non-BMP replacement characters in the `decodeUtf8With` function. See also comment at https://github.com/haskell/text/commit/ab90c65cdb7cd5bcbd739843b98bb5da515c8bce#diff-1c42d239accd16adcae29f591c519ac6R172 Short History: 1. 2189300b09ded9e2637a11edb055da32e11328f7 introduces the first version of `decodeUtf8With` affected by this issue 2. efa87327e5bdb43d5cdc09b5633f7d17b0b0b5d4 `decodeUtf8With` gets rewritten while retaining this this issue 3. 1780bd7940fe5beed86c071536ec38317345f314 `decodeUtf8With` gets changed to ensure that surrogate code-points get remapped to `U+FFFD` 4. ab90c65cdb7cd5bcbd739843b98bb5da515c8bce `decodeUtf8With` gets changed to fail non-silently in case of non-BMP replacement characters