const char *ttodata(const char *src, size_t srclen,
int base, char *dst, size_t dstlen, size_t *lenp);
size_t datatot(const char *src, size_t srclen,
int format, char *dst, size_t dstlen);
Currently supported formats are hexadecimal, base64, and characters.
A hexadecimal text value begins with a 0x (or 0X) prefix and continues with two-digit groups of hexadecimal digits (0-9, and a-f or A-F), each group encoding the value of one binary byte, high-order digit first. A single _ (underscore) between consecutive groups is ignored, permitting punctuation to improve readability; doing this every eight digits seems about right.
A base64 text value begins with a 0s (or 0S) prefix and continues with four-digit groups of base64 digits (A-Z, a-z, 0-9, +, and /), each group encoding the value of three binary bytes as described in section 6.8 of RFC 2045. Note that the last one or two digits of a base64 group can be = to indicate that fewer than three binary bytes are encoded.
A character text value begins with a 0t (or 0T) prefix and continues with text characters, each being the value of one binary byte.
All these functions basically copy data from src (whose size is specified by srclen) to dst (whose size is specified by dstlen), doing the conversion en route. If the result will not fit in dst, it is truncated; under no circumstances are more than dstlen bytes of result written to dst. Dstlen can be zero, in which case dst need not be valid and no result bytes are written at all.
The base parameter of ttodata specifies what format the input is in; normally it should be 0 to signify that this gets figured out from the prefix. Values of 16, 64, and 256 respectively signify hexadecimal, base64, and character-text formats without prefixes.
The format parameter of datatot, a single character used as a type code, specifies which text format is wanted. The value 0 (not ASCII '0', but a zero value) specifies a reasonable default. Other currently-supported values are:
The default format is currently 'h'.
Ttodata returns NULL for success and a pointer to a string-literal error message for failure; see DIAGNOSTICS. On success, if and only if lenp is non-NULL, *lenp is set to the number of bytes required to contain the full untruncated result. It is the caller's responsibility to check this against dstlen to determine whether he has obtained a complete result. The *lenp value is correct even if dstlen is zero, which offers a way to determine how much space would be needed before having to allocate any.
The normal return value of datatot is the number of bytes required to contain the full untruncated result. It is the caller's responsibility to check this against dstlen to determine whether he has obtained a complete result. The return value is correct even if dstlen is zero, which offers a way to determine how much space would be needed before having to allocate any. A return value of 0 signals a fatal error of some kind (see DIAGNOSTICS).
A zero value for srclen in ttodata (but not datatot!) is synonymous with strlen(src). A non-zero srclen in ttodata must not include the terminating NUL.
Unless dstlen is zero, the result supplied by datatot is always NUL-terminated, and its needed-size return value includes space for the terminating NUL.
Several obsolete variants of these functions (atodata, datatoa, atobytes, and bytestoa) are temporarily also supported.
Fatal errors in datatot are: unknown format code; zero-length input.
The 0s and 0t prefixes are the author's inventions and are not a standard of any kind. They have been chosen to avoid collisions with existing practice (some C implementations use 0b for binary) and possible confusion with unprefixed hexadecimal.
The restriction of ttodata error reports to literal strings (so that callers don't need to worry about freeing them or copying them) does limit the precision of error reporting.