Functions
bool	c4::base64_valid (const char *encoded, size_t encoded_sz)
	check that the given buffer is a valid base64 encoding
size_t	c4::base64_encode (char encoded, size_t encoded_sz, void const data, size_t data_sz)
	base64-encode binary data.
bool	c4::base64_decode (char const encoded, size_t encoded_sz, void data, size_t data_sz, size_t *data_sz_required)
	decode the base64 encoding in the given buffer.

Detailed Description

See also: https://en.wikipedia.org/wiki/Base64; https://www.base64encode.org/

Function Documentation

◆ base64_valid()

bool c4::base64_valid	(	const char *	encoded,
		size_t	encoded_sz )

check that the given buffer is a valid base64 encoding

See also: https://en.wikipedia.org/wiki/Base64

◆ base64_encode()

size_t c4::base64_encode	(	char *	encoded,
		size_t	encoded_sz,
		void const *	data,
		size_t	data_sz )

base64-encode binary data.

This is a plain implementation with a focus on simplicity and small footprint, such that it runs reasonably well in constrained platforms. On larger platforms it is reasonably fast (reaching 3GB/s and over), but it is not the fastest. If ultimate base64 speed in x64 platforms is your objective, there are faster implementations available. One recommendation is https://github.com/aklomp/base64, which uses a larger Look-Up Table (4096B as compared with 64B in c4core), making it between 1.5x~2x faster than c4core for larger payloads (but also slower for small payloads), and much faster when using AVX2 or AVX512 processing. But this speed comes at a cost in constrained platforms: eg c4core encodes ~2.5x faster in armv4 and armv5.

Parameters

encoded	[out] output buffer for encoded data
encoded_sz	[in] size of the output buffer for encoded data
data	[in] the input buffer with the binary data
data_sz	[in] size of the input buffer with the binary data

Returns: the number of bytes required for the output buffer. No writes occur beyond the end of the output buffer, so it is safe to do a speculative call where the encoded buffer is empty, or maybe too small. The caller should ensure that the returned size is smaller than the size of the encoded buffer.

Note: the result depends on endianness. If transfer between little/big endian systems is desired, the caller should normalize data before encoding.

See also: https://en.wikipedia.org/wiki/Base64

◆ base64_decode()

bool c4::base64_decode	(	char const *	encoded,
		size_t	encoded_sz,
		void *	data,
		size_t	data_sz,
		size_t *	data_sz_required )

decode the base64 encoding in the given buffer.

This is a plain implementation with a focus on simplicity and small footprint, such that it runs reasonably well in constrained platforms. On larger platforms it is reasonably fast, but it is not the fastest. If ultimate base64 speed in x64 platforms is your objective, there are faster implementations available. One recommendation is https://github.com/aklomp/base64, which uses up to 16x larger Look-Up Tables, making it between 1.5x~2x faster than c4core (but also slower for small payloads), and much faster when using AVX2 or AVX512 processing. But this x64 speed comes at a cost in constrained platforms: eg c4core decodes ~4x faster in armv4 and armv5.