Daniel Lemire

There are many instances in our code base where we could use the C++17 construct `if constexpr`. Starting with PR `match_system ` is a constexpr function. Thus many checks against...

Currently, we basically rely on the programmer to guard the code with feature macros in a consistent manner. Sadly, we do not appear to have automated ways to check that...

We have new functions to compute the UTF-8 length from UTF-16 content with replacement. We should add a RISC-V (RVV) implementation. Reference : https://github.com/simdutf/simdutf/pull/851

We have new functions to compute the UTF-8 length from UTF-16 content with replacement. We should add a RISC-V (RVV) implementation. Reference : https://github.com/simdutf/simdutf/pull/851

We have a new set of functions to compute the UTF-8 length from broken UTF 16 content. We should add it to our fuzzers. Currently, we can't yet transcode with...

Currently functions like `convert_utf16_to_utf8_safe` return how much was written, but not how much was read. It would be easy (meaning: no new complicated algorithm) to create a richer variant. References:...

By default, the `base64_to_binary_safe` quits on error, returning an error code and not necessarily decoding up to the point where there is an error in the stream. This flag exists...

We have our own specialized implementation of `std::text_encoding`. We should add interoperability with with `std::text_encoding`. https://www.en.cppreference.com/w/cpp/locale/text_encoding.html

implement conversion of binary data to base64 with lines for RISC-V processors

3

Currently we use a scalar routine, but we could optimize it. See https://github.com/simdutf/simdutf/pull/840 for x64 and ARM implementations.

Currently we use a scalar routine, but we could optimize it. See https://github.com/simdutf/simdutf/pull/840 for x64 and ARM implementations.