diff --git a/internal/cgen/base/strconv-impl.c b/internal/cgen/base/strconv-impl.c index 640fbc0c..afc5a913 100644 --- a/internal/cgen/base/strconv-impl.c +++ b/internal/cgen/base/strconv-impl.c @@ -299,6 +299,103 @@ fail_out_of_bounds: } while (0); } +// -------- + +// wuffs_base__render_number__first_hundred contains the decimal encodings of +// the first one hundred numbers [0 ..= 99]. +static const uint8_t wuffs_base__render_number__first_hundred[200] = { + '0', '0', '0', '1', '0', '2', '0', '3', '0', '4', // + '0', '5', '0', '6', '0', '7', '0', '8', '0', '9', // + '1', '0', '1', '1', '1', '2', '1', '3', '1', '4', // + '1', '5', '1', '6', '1', '7', '1', '8', '1', '9', // + '2', '0', '2', '1', '2', '2', '2', '3', '2', '4', // + '2', '5', '2', '6', '2', '7', '2', '8', '2', '9', // + '3', '0', '3', '1', '3', '2', '3', '3', '3', '4', // + '3', '5', '3', '6', '3', '7', '3', '8', '3', '9', // + '4', '0', '4', '1', '4', '2', '4', '3', '4', '4', // + '4', '5', '4', '6', '4', '7', '4', '8', '4', '9', // + '5', '0', '5', '1', '5', '2', '5', '3', '5', '4', // + '5', '5', '5', '6', '5', '7', '5', '8', '5', '9', // + '6', '0', '6', '1', '6', '2', '6', '3', '6', '4', // + '6', '5', '6', '6', '6', '7', '6', '8', '6', '9', // + '7', '0', '7', '1', '7', '2', '7', '3', '7', '4', // + '7', '5', '7', '6', '7', '7', '7', '8', '7', '9', // + '8', '0', '8', '1', '8', '2', '8', '3', '8', '4', // + '8', '5', '8', '6', '8', '7', '8', '8', '8', '9', // + '9', '0', '9', '1', '9', '2', '9', '3', '9', '4', // + '9', '5', '9', '6', '9', '7', '9', '8', '9', '9', // +}; + +static size_t // +wuffs_base__private_implementation__render_number_u64(wuffs_base__slice_u8 dst, + uint64_t x, + uint32_t options, + bool neg) { + uint8_t buf[WUFFS_BASE__U64__BYTE_LENGTH__MAX_INCL]; + uint8_t* ptr = &buf[0] + sizeof(buf); + + while (x >= 100) { + size_t index = (x % 100) * 2; + x /= 100; + uint8_t s0 = wuffs_base__render_number__first_hundred[index + 0]; + uint8_t s1 = wuffs_base__render_number__first_hundred[index + 1]; + ptr -= 2; + ptr[0] = s0; + ptr[1] = s1; + } + + if (x < 10) { + ptr -= 1; + ptr[0] = '0' + x; + } else { + size_t index = x * 2; + uint8_t s0 = wuffs_base__render_number__first_hundred[index + 0]; + uint8_t s1 = wuffs_base__render_number__first_hundred[index + 1]; + ptr -= 2; + ptr[0] = s0; + ptr[1] = s1; + } + + if (neg) { + ptr -= 1; + ptr[0] = '-'; + } else if (options & WUFFS_BASE__RENDER_NUMBER__LEADING_PLUS_SIGN) { + ptr -= 1; + ptr[0] = '+'; + } + + size_t n = sizeof(buf) - (ptr - &buf[0]); + if (n > dst.len) { + return 0; + } + memcpy(dst.ptr + ((options & WUFFS_BASE__RENDER_NUMBER__ALIGN_RIGHT) + ? (dst.len - n) + : 0), + ptr, n); + return n; +} + +WUFFS_BASE__MAYBE_STATIC size_t // +wuffs_base__render_number_i64(wuffs_base__slice_u8 dst, + int64_t x, + uint32_t options) { + uint64_t u = x; + bool neg = x < 0; + if (neg) { + u = 1 + ~u; + } + return wuffs_base__private_implementation__render_number_u64(dst, u, options, + neg); +} + +WUFFS_BASE__MAYBE_STATIC size_t // +wuffs_base__render_number_u64(wuffs_base__slice_u8 dst, + uint64_t x, + uint32_t options) { + return wuffs_base__private_implementation__render_number_u64(dst, x, options, + false); +} + // ---------------- Hexadecimal WUFFS_BASE__MAYBE_STATIC size_t // diff --git a/internal/cgen/base/strconv-public.h b/internal/cgen/base/strconv-public.h index 33b4ec3e..098dd05b 100644 --- a/internal/cgen/base/strconv-public.h +++ b/internal/cgen/base/strconv-public.h @@ -95,6 +95,52 @@ wuffs_base__parse_number_u64(wuffs_base__slice_u8 s); WUFFS_BASE__MAYBE_STATIC wuffs_base__result_f64 // wuffs_base__parse_number_f64(wuffs_base__slice_u8 s); +// -------- + +// Options (bitwise or'ed together) for wuffs_base__render_number_etc +// functions. + +#define WUFFS_BASE__RENDER_NUMBER__DEFAULT_OPTIONS ((uint32_t)0x00000000) + +// WUFFS_BASE__RENDER_NUMBER__ALIGN_RIGHT means to render to the right side +// (higher indexes) of the destination slice, leaving any untouched bytes on +// the left side (lower indexes). The default is vice versa: rendering on the +// left with slack on the right. +#define WUFFS_BASE__RENDER_NUMBER__ALIGN_RIGHT ((uint32_t)0x00000001) + +// WUFFS_BASE__RENDER_NUMBER__LEADING_PLUS_SIGN means to render the unnecessary +// leading "+" for non-negative numbers: "+0" and "+123", not "0" and "123". +#define WUFFS_BASE__RENDER_NUMBER__LEADING_PLUS_SIGN ((uint32_t)0x00000002) + +// -------- + +#define WUFFS_BASE__I64__BYTE_LENGTH__MAX_INCL 20 +#define WUFFS_BASE__U64__BYTE_LENGTH__MAX_INCL 21 + +// wuffs_base__render_number_i64 writes the decimal encoding of x to dst and +// returns the number of bytes written. If dst is shorter than the entire +// encoding, it returns 0 (and no bytes are written). +// +// dst will never be too short if its length is at least 20, also known as +// WUFFS_BASE__I64__BYTE_LENGTH__MAX_INCL. +WUFFS_BASE__MAYBE_STATIC size_t // +wuffs_base__render_number_i64(wuffs_base__slice_u8 dst, + int64_t x, + uint32_t options); + +// wuffs_base__render_number_u64 writes the decimal encoding of x to dst and +// returns the number of bytes written. If dst is shorter than the entire +// encoding, it returns 0 (and no bytes are written). +// +// dst will never be too short if its length is at least 21, also known as +// WUFFS_BASE__U64__BYTE_LENGTH__MAX_INCL. +WUFFS_BASE__MAYBE_STATIC size_t // +wuffs_base__render_number_u64(wuffs_base__slice_u8 dst, + uint64_t x, + uint32_t options); + +// -------- + // wuffs_base__ieee_754_bit_representation__etc converts between a double // precision numerical value and its IEEE 754 64-bit representation (1 sign // bit, 11 exponent bits, 52 explicit significand bits). diff --git a/internal/cgen/data/data.go b/internal/cgen/data/data.go index 1242a486..e32eae09 100644 --- a/internal/cgen/data/data.go +++ b/internal/cgen/data/data.go @@ -41,6 +41,11 @@ const BaseStrConvImplC = "" + "} while (0);\n\nhexadecimal:\n do {\n uint64_t v = wuffs_base__parse_number__hexadecimal_digits[*p++];\n if (v == 0) {\n goto fail_bad_argument;\n }\n v &= 0x0F;\n\n for (; p < q; p++) {\n if (*p == '_') {\n continue;\n }\n uint8_t digit = wuffs_base__parse_number__hexadecimal_digits[*p];\n if (digit == 0) {\n goto fail_bad_argument;\n }\n digit &= 0x0F;\n if ((v >> 60) != 0) {\n goto fail_out_of_bounds;\n }\n v = (v << 4) | ((uint64_t)(digit));\n }\n\n wuffs_base__result_u64 ret;\n ret.status.repr = NULL;\n ret.value = v;\n return ret;\n } while (0);\n\nok_zero:\n do {\n wuffs_base__result_u64 ret;\n ret.status.repr = NULL;\n ret.value = 0;\n return ret;\n } while (0);\n\nfail_bad_argument:\n do {\n wuffs_base__result_u64 ret;\n ret.status.repr = wuffs_base__error__bad_argument;\n ret.value = 0;\n return ret;\n } while (0);\n\nfail_out_of_bounds:\n do {\n wuffs_base__result_u64 ret;\n ret.status.repr = wuffs_base__error__" + "out_of_bounds;\n ret.value = 0;\n return ret;\n } while (0);\n}\n\n" + "" + + "// --------\n\n// wuffs_base__render_number__first_hundred contains the decimal encodings of\n// the first one hundred numbers [0 ..= 99].\nstatic const uint8_t wuffs_base__render_number__first_hundred[200] = {\n '0', '0', '0', '1', '0', '2', '0', '3', '0', '4', //\n '0', '5', '0', '6', '0', '7', '0', '8', '0', '9', //\n '1', '0', '1', '1', '1', '2', '1', '3', '1', '4', //\n '1', '5', '1', '6', '1', '7', '1', '8', '1', '9', //\n '2', '0', '2', '1', '2', '2', '2', '3', '2', '4', //\n '2', '5', '2', '6', '2', '7', '2', '8', '2', '9', //\n '3', '0', '3', '1', '3', '2', '3', '3', '3', '4', //\n '3', '5', '3', '6', '3', '7', '3', '8', '3', '9', //\n '4', '0', '4', '1', '4', '2', '4', '3', '4', '4', //\n '4', '5', '4', '6', '4', '7', '4', '8', '4', '9', //\n '5', '0', '5', '1', '5', '2', '5', '3', '5', '4', //\n '5', '5', '5', '6', '5', '7', '5', '8', '5', '9', //\n '6', '0', '6', '1', '6', '2', '6', '3', '6', '4', //\n '6', '5', '6', '6', '6', '7', '6', '8', '6', '9', //\n '" + + "7', '0', '7', '1', '7', '2', '7', '3', '7', '4', //\n '7', '5', '7', '6', '7', '7', '7', '8', '7', '9', //\n '8', '0', '8', '1', '8', '2', '8', '3', '8', '4', //\n '8', '5', '8', '6', '8', '7', '8', '8', '8', '9', //\n '9', '0', '9', '1', '9', '2', '9', '3', '9', '4', //\n '9', '5', '9', '6', '9', '7', '9', '8', '9', '9', //\n};\n\nstatic size_t //\nwuffs_base__private_implementation__render_number_u64(wuffs_base__slice_u8 dst,\n uint64_t x,\n uint32_t options,\n bool neg) {\n uint8_t buf[WUFFS_BASE__U64__BYTE_LENGTH__MAX_INCL];\n uint8_t* ptr = &buf[0] + sizeof(buf);\n\n while (x >= 100) {\n size_t index = (x % 100) * 2;\n x /= 100;\n uint8_t s0 = wuffs_base__render_number__first_hundred[index + 0];\n uint8_t s1 = wuffs_base__render_number__first_hundred[index + 1];\n ptr -= 2;\n ptr[0] = s0;\n ptr[1] = s1;\n }\n\n if (x < 10) {\n " + + " ptr -= 1;\n ptr[0] = '0' + x;\n } else {\n size_t index = x * 2;\n uint8_t s0 = wuffs_base__render_number__first_hundred[index + 0];\n uint8_t s1 = wuffs_base__render_number__first_hundred[index + 1];\n ptr -= 2;\n ptr[0] = s0;\n ptr[1] = s1;\n }\n\n if (neg) {\n ptr -= 1;\n ptr[0] = '-';\n } else if (options & WUFFS_BASE__RENDER_NUMBER__LEADING_PLUS_SIGN) {\n ptr -= 1;\n ptr[0] = '+';\n }\n\n size_t n = sizeof(buf) - (ptr - &buf[0]);\n if (n > dst.len) {\n return 0;\n }\n memcpy(dst.ptr + ((options & WUFFS_BASE__RENDER_NUMBER__ALIGN_RIGHT)\n ? (dst.len - n)\n : 0),\n ptr, n);\n return n;\n}\n\nWUFFS_BASE__MAYBE_STATIC size_t //\nwuffs_base__render_number_i64(wuffs_base__slice_u8 dst,\n int64_t x,\n uint32_t options) {\n uint64_t u = x;\n bool neg = x < 0;\n if (neg) {\n u = 1 + ~u;\n }\n return wuffs_base__private_implementation__render_number_u64(dst, u, options,\n " + + " neg);\n}\n\nWUFFS_BASE__MAYBE_STATIC size_t //\nwuffs_base__render_number_u64(wuffs_base__slice_u8 dst,\n uint64_t x,\n uint32_t options) {\n return wuffs_base__private_implementation__render_number_u64(dst, x, options,\n false);\n}\n\n" + + "" + "// ---------------- Hexadecimal\n\nWUFFS_BASE__MAYBE_STATIC size_t //\nwuffs_base__hexadecimal__decode2(wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 src) {\n size_t src_len2 = src.len / 2;\n size_t len = dst.len < src_len2 ? dst.len : src_len2;\n uint8_t* d = dst.ptr;\n uint8_t* s = src.ptr;\n size_t n = len;\n\n while (n--) {\n *d = (uint8_t)((wuffs_base__parse_number__hexadecimal_digits[s[0]] << 4) |\n (wuffs_base__parse_number__hexadecimal_digits[s[1]] & 0x0F));\n d += 1;\n s += 2;\n }\n\n return len;\n}\n\nWUFFS_BASE__MAYBE_STATIC size_t //\nwuffs_base__hexadecimal__decode4(wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 src) {\n size_t src_len4 = src.len / 4;\n size_t len = dst.len < src_len4 ? dst.len : src_len4;\n uint8_t* d = dst.ptr;\n uint8_t* s = src.ptr;\n size_t n = len;\n\n while (n--) {\n *d = (uint8_t)((wuffs_base__parse_number__hexadecimal_digits[s[2]] << 4) |\n (wuffs_base__parse_number__hexa" + "decimal_digits[s[3]] & 0x0F));\n d += 1;\n s += 4;\n }\n\n return len;\n}\n\n" + "" + @@ -414,8 +419,14 @@ const BaseStrConvPublicH = "" + "" + "// ---------------- IEEE 754 Floating Point\n\n// wuffs_base__parse_number_f64 parses the floating point number in s. For\n// example, if s contains the bytes \"1.5\" then it will return the double 1.5.\n//\n// It returns an error if s does not contain a floating point number.\n//\n// It does not necessarily return an error if the conversion is lossy, e.g. if\n// s is \"0.3\", which double-precision floating point cannot represent exactly.\n//\n// Similarly, the returned value may be infinite (and no error returned) even\n// if s was not \"inf\", when the input is nominally finite but sufficiently\n// larger than DBL_MAX, about 1.8e+308.\n//\n// It is similar to the C standard library's strtod function, but:\n// - Errors are returned in-band (in a result type), not out-of-band (errno).\n// - It takes a slice (a pointer and length), not a NUL-terminated C string.\n// - It does not take an optional endptr argument. It does not allow a partial\n// parse: it returns an error unless all of s is consumed.\n// - It does not allow whi" + "tespace, leading or otherwise.\n// - It does not allow unnecessary leading zeroes (\"0\" is valid and its sole\n// zero is necessary). All of \"00\", \"0644\" and \"00.7\" are invalid.\n// - It is not affected by i18n / l10n settings such as environment variables.\n// - Conversely, it always accepts either ',' or '.' as a decimal separator.\n// In particular, \"3,141,592\" is always invalid but \"3,141\" is always valid\n// (and approximately π). The caller is responsible for e.g. previously\n// rejecting or filtering s if it contains a comma, if that is unacceptable\n// to the caller. For example, JSON numbers always use a dot '.' and never a\n// comma ',', regardless of the LOCALE environment variable.\n// - It does allow arbitrary underscores. For example, \"_3.141_592\" would\n// successfully parse, again approximately π.\n// - It does allow \"inf\", \"+Infinity\" and \"-NAN\", case insensitive, but it\n// does not permit \"nan\" to be followed by an integer mantissa.\n// - It does not allow hexadecimal float" + - "ing point numbers.\n//\n// For modular builds that divide the base module into sub-modules, using this\n// function requires the WUFFS_CONFIG__MODULE__BASE__F64CONV sub-module, not\n// just WUFFS_CONFIG__MODULE__BASE__CORE.\nWUFFS_BASE__MAYBE_STATIC wuffs_base__result_f64 //\nwuffs_base__parse_number_f64(wuffs_base__slice_u8 s);\n\n// wuffs_base__ieee_754_bit_representation__etc converts between a double\n// precision numerical value and its IEEE 754 64-bit representation (1 sign\n// bit, 11 exponent bits, 52 explicit significand bits).\n//\n// For example, it converts between:\n// - +1.0 and 0x3FF0_0000_0000_0000.\n// - +5.5 and 0x4016_0000_0000_0000.\n// - -inf and 0xFFF0_0000_0000_0000.\n//\n// See https://en.wikipedia.org/wiki/Double-precision_floating-point_format\n\nstatic inline uint64_t //\nwuffs_base__ieee_754_bit_representation__from_f64(double f) {\n uint64_t u = 0;\n if (sizeof(uint64_t) == sizeof(double)) {\n memcpy(&u, &f, sizeof(uint64_t));\n }\n return u;\n}\n\nstatic inline double //\nwuffs_base__ieee_754_bi" + - "t_representation__to_f64(uint64_t u) {\n double f = 0;\n if (sizeof(uint64_t) == sizeof(double)) {\n memcpy(&f, &u, sizeof(uint64_t));\n }\n return f;\n}\n\n" + + "ing point numbers.\n//\n// For modular builds that divide the base module into sub-modules, using this\n// function requires the WUFFS_CONFIG__MODULE__BASE__F64CONV sub-module, not\n// just WUFFS_CONFIG__MODULE__BASE__CORE.\nWUFFS_BASE__MAYBE_STATIC wuffs_base__result_f64 //\nwuffs_base__parse_number_f64(wuffs_base__slice_u8 s);\n\n" + + "" + + "// --------\n\n// Options (bitwise or'ed together) for wuffs_base__render_number_etc\n// functions.\n\n#define WUFFS_BASE__RENDER_NUMBER__DEFAULT_OPTIONS ((uint32_t)0x00000000)\n\n// WUFFS_BASE__RENDER_NUMBER__ALIGN_RIGHT means to render to the right side\n// (higher indexes) of the destination slice, leaving any untouched bytes on\n// the left side (lower indexes). The default is vice versa: rendering on the\n// left with slack on the right.\n#define WUFFS_BASE__RENDER_NUMBER__ALIGN_RIGHT ((uint32_t)0x00000001)\n\n// WUFFS_BASE__RENDER_NUMBER__LEADING_PLUS_SIGN means to render the unnecessary\n// leading \"+\" for non-negative numbers: \"+0\" and \"+123\", not \"0\" and \"123\".\n#define WUFFS_BASE__RENDER_NUMBER__LEADING_PLUS_SIGN ((uint32_t)0x00000002)\n\n" + + "" + + "// --------\n\n#define WUFFS_BASE__I64__BYTE_LENGTH__MAX_INCL 20\n#define WUFFS_BASE__U64__BYTE_LENGTH__MAX_INCL 21\n\n// wuffs_base__render_number_i64 writes the decimal encoding of x to dst and\n// returns the number of bytes written. If dst is shorter than the entire\n// encoding, it returns 0 (and no bytes are written).\n//\n// dst will never be too short if its length is at least 20, also known as\n// WUFFS_BASE__I64__BYTE_LENGTH__MAX_INCL.\nWUFFS_BASE__MAYBE_STATIC size_t //\nwuffs_base__render_number_i64(wuffs_base__slice_u8 dst,\n int64_t x,\n uint32_t options);\n\n// wuffs_base__render_number_u64 writes the decimal encoding of x to dst and\n// returns the number of bytes written. If dst is shorter than the entire\n// encoding, it returns 0 (and no bytes are written).\n//\n// dst will never be too short if its length is at least 21, also known as\n// WUFFS_BASE__U64__BYTE_LENGTH__MAX_INCL.\nWUFFS_BASE__MAYBE_STATIC size_t //\nwuffs_base__render_number_u64(wuffs_bas" + + "e__slice_u8 dst,\n uint64_t x,\n uint32_t options);\n\n" + + "" + + "// --------\n\n// wuffs_base__ieee_754_bit_representation__etc converts between a double\n// precision numerical value and its IEEE 754 64-bit representation (1 sign\n// bit, 11 exponent bits, 52 explicit significand bits).\n//\n// For example, it converts between:\n// - +1.0 and 0x3FF0_0000_0000_0000.\n// - +5.5 and 0x4016_0000_0000_0000.\n// - -inf and 0xFFF0_0000_0000_0000.\n//\n// See https://en.wikipedia.org/wiki/Double-precision_floating-point_format\n\nstatic inline uint64_t //\nwuffs_base__ieee_754_bit_representation__from_f64(double f) {\n uint64_t u = 0;\n if (sizeof(uint64_t) == sizeof(double)) {\n memcpy(&u, &f, sizeof(uint64_t));\n }\n return u;\n}\n\nstatic inline double //\nwuffs_base__ieee_754_bit_representation__to_f64(uint64_t u) {\n double f = 0;\n if (sizeof(uint64_t) == sizeof(double)) {\n memcpy(&f, &u, sizeof(uint64_t));\n }\n return f;\n}\n\n" + "" + "// ---------------- Hexadecimal\n\n// wuffs_base__hexadecimal__decode2 converts \"6A6b\" to \"jk\", where e.g. 'j' is\n// U+006A. There are 2 source bytes for every destination byte.\n//\n// It returns the number of dst bytes written: the minimum of dst.len and\n// (src.len / 2). Excess source bytes are ignored.\n//\n// It assumes that the src bytes are two hexadecimal digits (0-9, A-F, a-f),\n// repeated. It may write nonsense bytes if not, although it will not read or\n// write out of bounds.\nWUFFS_BASE__MAYBE_STATIC size_t //\nwuffs_base__hexadecimal__decode2(wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 src);\n\n// wuffs_base__hexadecimal__decode4 converts \"\\\\x6A\\\\x6b\" to \"jk\", where e.g.\n// 'j' is U+006A. There are 4 source bytes for every destination byte.\n//\n// It returns the number of dst bytes written: the minimum of dst.len and\n// (src.len / 4). Excess source bytes are ignored.\n//\n// It assumes that the src bytes are two ignored bytes and then two hexadecimal\n// digits (0-9, A-F, a" + "-f), repeated. It may write nonsense bytes if not,\n// although it will not read or write out of bounds.\nWUFFS_BASE__MAYBE_STATIC size_t //\nwuffs_base__hexadecimal__decode4(wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 src);\n\n" + diff --git a/release/c/wuffs-unsupported-snapshot.c b/release/c/wuffs-unsupported-snapshot.c index 580e4eb4..7588b06f 100644 --- a/release/c/wuffs-unsupported-snapshot.c +++ b/release/c/wuffs-unsupported-snapshot.c @@ -3943,6 +3943,52 @@ wuffs_base__parse_number_u64(wuffs_base__slice_u8 s); WUFFS_BASE__MAYBE_STATIC wuffs_base__result_f64 // wuffs_base__parse_number_f64(wuffs_base__slice_u8 s); +// -------- + +// Options (bitwise or'ed together) for wuffs_base__render_number_etc +// functions. + +#define WUFFS_BASE__RENDER_NUMBER__DEFAULT_OPTIONS ((uint32_t)0x00000000) + +// WUFFS_BASE__RENDER_NUMBER__ALIGN_RIGHT means to render to the right side +// (higher indexes) of the destination slice, leaving any untouched bytes on +// the left side (lower indexes). The default is vice versa: rendering on the +// left with slack on the right. +#define WUFFS_BASE__RENDER_NUMBER__ALIGN_RIGHT ((uint32_t)0x00000001) + +// WUFFS_BASE__RENDER_NUMBER__LEADING_PLUS_SIGN means to render the unnecessary +// leading "+" for non-negative numbers: "+0" and "+123", not "0" and "123". +#define WUFFS_BASE__RENDER_NUMBER__LEADING_PLUS_SIGN ((uint32_t)0x00000002) + +// -------- + +#define WUFFS_BASE__I64__BYTE_LENGTH__MAX_INCL 20 +#define WUFFS_BASE__U64__BYTE_LENGTH__MAX_INCL 21 + +// wuffs_base__render_number_i64 writes the decimal encoding of x to dst and +// returns the number of bytes written. If dst is shorter than the entire +// encoding, it returns 0 (and no bytes are written). +// +// dst will never be too short if its length is at least 20, also known as +// WUFFS_BASE__I64__BYTE_LENGTH__MAX_INCL. +WUFFS_BASE__MAYBE_STATIC size_t // +wuffs_base__render_number_i64(wuffs_base__slice_u8 dst, + int64_t x, + uint32_t options); + +// wuffs_base__render_number_u64 writes the decimal encoding of x to dst and +// returns the number of bytes written. If dst is shorter than the entire +// encoding, it returns 0 (and no bytes are written). +// +// dst will never be too short if its length is at least 21, also known as +// WUFFS_BASE__U64__BYTE_LENGTH__MAX_INCL. +WUFFS_BASE__MAYBE_STATIC size_t // +wuffs_base__render_number_u64(wuffs_base__slice_u8 dst, + uint64_t x, + uint32_t options); + +// -------- + // wuffs_base__ieee_754_bit_representation__etc converts between a double // precision numerical value and its IEEE 754 64-bit representation (1 sign // bit, 11 exponent bits, 52 explicit significand bits). @@ -8474,6 +8520,103 @@ fail_out_of_bounds: } while (0); } +// -------- + +// wuffs_base__render_number__first_hundred contains the decimal encodings of +// the first one hundred numbers [0 ..= 99]. +static const uint8_t wuffs_base__render_number__first_hundred[200] = { + '0', '0', '0', '1', '0', '2', '0', '3', '0', '4', // + '0', '5', '0', '6', '0', '7', '0', '8', '0', '9', // + '1', '0', '1', '1', '1', '2', '1', '3', '1', '4', // + '1', '5', '1', '6', '1', '7', '1', '8', '1', '9', // + '2', '0', '2', '1', '2', '2', '2', '3', '2', '4', // + '2', '5', '2', '6', '2', '7', '2', '8', '2', '9', // + '3', '0', '3', '1', '3', '2', '3', '3', '3', '4', // + '3', '5', '3', '6', '3', '7', '3', '8', '3', '9', // + '4', '0', '4', '1', '4', '2', '4', '3', '4', '4', // + '4', '5', '4', '6', '4', '7', '4', '8', '4', '9', // + '5', '0', '5', '1', '5', '2', '5', '3', '5', '4', // + '5', '5', '5', '6', '5', '7', '5', '8', '5', '9', // + '6', '0', '6', '1', '6', '2', '6', '3', '6', '4', // + '6', '5', '6', '6', '6', '7', '6', '8', '6', '9', // + '7', '0', '7', '1', '7', '2', '7', '3', '7', '4', // + '7', '5', '7', '6', '7', '7', '7', '8', '7', '9', // + '8', '0', '8', '1', '8', '2', '8', '3', '8', '4', // + '8', '5', '8', '6', '8', '7', '8', '8', '8', '9', // + '9', '0', '9', '1', '9', '2', '9', '3', '9', '4', // + '9', '5', '9', '6', '9', '7', '9', '8', '9', '9', // +}; + +static size_t // +wuffs_base__private_implementation__render_number_u64(wuffs_base__slice_u8 dst, + uint64_t x, + uint32_t options, + bool neg) { + uint8_t buf[WUFFS_BASE__U64__BYTE_LENGTH__MAX_INCL]; + uint8_t* ptr = &buf[0] + sizeof(buf); + + while (x >= 100) { + size_t index = (x % 100) * 2; + x /= 100; + uint8_t s0 = wuffs_base__render_number__first_hundred[index + 0]; + uint8_t s1 = wuffs_base__render_number__first_hundred[index + 1]; + ptr -= 2; + ptr[0] = s0; + ptr[1] = s1; + } + + if (x < 10) { + ptr -= 1; + ptr[0] = '0' + x; + } else { + size_t index = x * 2; + uint8_t s0 = wuffs_base__render_number__first_hundred[index + 0]; + uint8_t s1 = wuffs_base__render_number__first_hundred[index + 1]; + ptr -= 2; + ptr[0] = s0; + ptr[1] = s1; + } + + if (neg) { + ptr -= 1; + ptr[0] = '-'; + } else if (options & WUFFS_BASE__RENDER_NUMBER__LEADING_PLUS_SIGN) { + ptr -= 1; + ptr[0] = '+'; + } + + size_t n = sizeof(buf) - (ptr - &buf[0]); + if (n > dst.len) { + return 0; + } + memcpy(dst.ptr + ((options & WUFFS_BASE__RENDER_NUMBER__ALIGN_RIGHT) + ? (dst.len - n) + : 0), + ptr, n); + return n; +} + +WUFFS_BASE__MAYBE_STATIC size_t // +wuffs_base__render_number_i64(wuffs_base__slice_u8 dst, + int64_t x, + uint32_t options) { + uint64_t u = x; + bool neg = x < 0; + if (neg) { + u = 1 + ~u; + } + return wuffs_base__private_implementation__render_number_u64(dst, u, options, + neg); +} + +WUFFS_BASE__MAYBE_STATIC size_t // +wuffs_base__render_number_u64(wuffs_base__slice_u8 dst, + uint64_t x, + uint32_t options) { + return wuffs_base__private_implementation__render_number_u64(dst, x, options, + false); +} + // ---------------- Hexadecimal WUFFS_BASE__MAYBE_STATIC size_t // diff --git a/test/c/std/json.c b/test/c/std/json.c index 6dcef26e..bb5d9674 100644 --- a/test/c/std/json.c +++ b/test/c/std/json.c @@ -858,6 +858,159 @@ test_wuffs_strconv_parse_number_u64() { return NULL; } +const char* // +test_wuffs_strconv_render_number_i64() { + CHECK_FOCUS(__func__); + + struct { + int64_t x; + const char* want; + } test_cases[] = { + {.x = +0x0000000000000000l, .want = "0"}, + {.x = +0x0000000000000009l, .want = "9"}, + {.x = +0x000000000000000Al, .want = "10"}, + {.x = +0x000000000000004Al, .want = "74"}, + {.x = +0x0000000000000063l, .want = "99"}, + {.x = +0x0000000000000064l, .want = "100"}, + {.x = +0x000000000000007Cl, .want = "124"}, + {.x = +0x00000000000001F4l, .want = "500"}, + {.x = +0x000000000000036Cl, .want = "876"}, + {.x = +0x000000000000036Fl, .want = "879"}, + {.x = +0x0000000000000929l, .want = "2345"}, + {.x = +0x0000000000010932l, .want = "67890"}, + {.x = +0x00000000FFFFFFFFl, .want = "4294967295"}, + {.x = +0x0000000100000000l, .want = "4294967296"}, + {.x = +0x0123456789ABCDEFl, .want = "81985529216486895"}, + {.x = +0x7FFFFFFFFFFFFFFFl, .want = "9223372036854775807"}, + + {.x = -0x0000000000000009l, .want = "-9"}, + {.x = -0x000000000000000Al, .want = "-10"}, + {.x = -0x000000000000004Al, .want = "-74"}, + {.x = -0x0000000000000063l, .want = "-99"}, + {.x = -0x0000000000000064l, .want = "-100"}, + {.x = -0x000000000000007Cl, .want = "-124"}, + {.x = -0x00000000000001F4l, .want = "-500"}, + {.x = -0x000000000000036Cl, .want = "-876"}, + {.x = -0x000000000000036Fl, .want = "-879"}, + {.x = -0x0000000000000929l, .want = "-2345"}, + {.x = -0x0000000000010932l, .want = "-67890"}, + {.x = -0x00000000FFFFFFFFl, .want = "-4294967295"}, + {.x = -0x0000000100000000l, .want = "-4294967296"}, + {.x = -0x0123456789ABCDEFl, .want = "-81985529216486895"}, + {.x = -0x7FFFFFFFFFFFFFFFl, .want = "-9223372036854775807"}, + {.x = -0x8000000000000000l, .want = "-9223372036854775808"}, + }; + + if (g_have_slice_u8.len < WUFFS_BASE__I64__BYTE_LENGTH__MAX_INCL) { + return "g_have_slice_u8.len is too short"; + } + + int tc; + for (tc = 0; tc < WUFFS_TESTLIB_ARRAY_SIZE(test_cases); tc++) { + size_t n = wuffs_base__render_number_i64( + g_have_slice_u8, test_cases[tc].x, + WUFFS_BASE__RENDER_NUMBER__DEFAULT_OPTIONS); + if (n == 0) { + RETURN_FAIL("%" PRId64 ": could not render", test_cases[tc].x); + } else if (n >= g_have_slice_u8.len) { + RETURN_FAIL("%" PRId64 ": n is too large", test_cases[tc].x); + } + g_have_slice_u8.ptr[n] = 0x00; + if (strcmp((const char*)(g_have_slice_u8.ptr), test_cases[tc].want) != 0) { + RETURN_FAIL("%" PRId64 ": have \"%s\", want \"%s\"", test_cases[tc].x, + g_have_slice_u8.ptr, test_cases[tc].want); + } + } + + return NULL; +} + +const char* // +test_wuffs_strconv_render_number_u64() { + CHECK_FOCUS(__func__); + + struct { + uint64_t x; + const char* want; + } test_cases[] = { + {.x = 0x0000000000000000l, .want = "0"}, + {.x = 0x0000000000000009l, .want = "9"}, + {.x = 0x000000000000000Al, .want = "10"}, + {.x = 0x000000000000004Al, .want = "74"}, + {.x = 0x0000000000000063l, .want = "99"}, + {.x = 0x0000000000000064l, .want = "100"}, + {.x = 0x000000000000007Cl, .want = "124"}, + {.x = 0x00000000000001F4l, .want = "500"}, + {.x = 0x000000000000036Cl, .want = "876"}, + {.x = 0x000000000000036Fl, .want = "879"}, + {.x = 0x0000000000000929l, .want = "2345"}, + {.x = 0x0000000000010932l, .want = "67890"}, + {.x = 0x00000000FFFFFFFFl, .want = "4294967295"}, + {.x = 0x0000000100000000l, .want = "4294967296"}, + {.x = 0x0123456789ABCDEFl, .want = "81985529216486895"}, + {.x = 0x7FFFFFFFFFFFFFFFl, .want = "9223372036854775807"}, + {.x = 0x8000000000000000l, .want = "9223372036854775808"}, + {.x = 0xFFFFFFFFFFFFFFF9l, .want = "18446744073709551609"}, + {.x = 0xFFFFFFFFFFFFFFFAl, .want = "18446744073709551610"}, + {.x = 0xFFFFFFFFFFFFFFFEl, .want = "18446744073709551614"}, + {.x = 0xFFFFFFFFFFFFFFFFl, .want = "18446744073709551615"}, + }; + + if (g_have_slice_u8.len < WUFFS_BASE__U64__BYTE_LENGTH__MAX_INCL) { + return "g_have_slice_u8.len is too short"; + } + + int tc; + for (tc = 0; tc < WUFFS_TESTLIB_ARRAY_SIZE(test_cases); tc++) { + size_t n = wuffs_base__render_number_u64( + g_have_slice_u8, test_cases[tc].x, + WUFFS_BASE__RENDER_NUMBER__DEFAULT_OPTIONS); + if (n == 0) { + RETURN_FAIL("%" PRIu64 ": could not render", test_cases[tc].x); + } else if (n >= g_have_slice_u8.len) { + RETURN_FAIL("%" PRIu64 ": n is too large", test_cases[tc].x); + } + g_have_slice_u8.ptr[n] = 0x00; + if (strcmp((const char*)(g_have_slice_u8.ptr), test_cases[tc].want) != 0) { + RETURN_FAIL("%" PRIu64 ": have \"%s\", want \"%s\"", test_cases[tc].x, + g_have_slice_u8.ptr, test_cases[tc].want); + } + } + + // Test dst slice is too short. + { + uint8_t dst[5]; + size_t n = wuffs_base__render_number_u64( + wuffs_base__make_slice_u8(&dst[0], sizeof(dst)), 123456, + WUFFS_BASE__RENDER_NUMBER__DEFAULT_OPTIONS); + if (n != 0) { + RETURN_FAIL("dst too short: have %zu, want 0", n); + } + } + + // Test options. + { + uint8_t dst[8]; + memcpy(&dst[0], "ABCDEFG\x00", 8); + size_t n = wuffs_base__render_number_u64( + wuffs_base__make_slice_u8(&dst[0], sizeof(dst) - 1), 1234, + WUFFS_BASE__RENDER_NUMBER__ALIGN_RIGHT | + WUFFS_BASE__RENDER_NUMBER__LEADING_PLUS_SIGN); + if (n != 5) { + RETURN_FAIL("ALIGN_RIGHT | LEADING_PLUS_SIGN: have %zu, want 5", n); + } + uint64_t have = wuffs_base__load_u64be__no_bounds_check(&dst[0]); + uint64_t want = 0x41422B3132333400; // "AB+1234\x00". + if (have != want) { + RETURN_FAIL("ALIGN_RIGHT | LEADING_PLUS_SIGN: have 0x%" PRIX64 + ", want 0x%" PRIX64, + have, want); + } + } + + return NULL; +} + const char* // test_wuffs_strconv_utf_8_next() { CHECK_FOCUS(__func__); @@ -2481,6 +2634,8 @@ proc g_tests[] = { test_wuffs_strconv_parse_number_f64, test_wuffs_strconv_parse_number_i64, test_wuffs_strconv_parse_number_u64, + test_wuffs_strconv_render_number_i64, + test_wuffs_strconv_render_number_u64, test_wuffs_strconv_utf_8_next, test_wuffs_json_decode_end_of_data,