[Feature Request] stable interface for base64 encoding/decoding

[bgemmill]

This came up tangentially on Stack Overflow, would it be possible to promote the base64 encoding/decoding functions from the detail namespace?

They're really useful, and far and away cleaner than the other pure-boost approaches.

[vinniefalco]

If I do that, I will have to improve the algorithms to hold them to a higher standard, and they will need documentation and more robust tests. I will consider it!

[yost-jacob]

since you already rely on openssl wouldn't it make sense to just use the openssl tools? Something like this on Stack Overflow?

certainly solves the algorithm problem.

[vinniefalco]

You can just copy the code, it is small enough and I think copying it creates the least headache.

[sehe]

@vinniefalco Can I refine this story to the point where we can make it happen? Perhaps it should become part of a more general purpose library. Could it be part of Boost Strings (algorithms) or even Core?

Reason I think it's worth it is because of continuous support on SO. And in part the fact that the "other" Boost facility is ... terrible. Truly terrible (see https://stackoverflow.com/a/71933410/85371 (boostorg/serialization#254), https://stackoverflow.com/a/72034755/85371, and more recently this thread https://stackoverflow.com/questions/70624519/boost-gzip-how-to-output-compressed-string-as-text/70625599#comment129071097_70625599)

[vinniefalco]

First check if Boost doesn't already have something. Second, yes I think there is room for prescription-strength radix conversions (think gigabytes per second streaming speed). However this doesn't belong in beast, and I don't have the bandwidth to think through a proper solution at the moment. Perhaps you should ask on the list?