Add wuffs_base__utf_8__next_from_end

This commit is contained in:
Nigel Tao committed 2020-07-02 15:08:54 +10:00
1 parent 991e1b2667
commit a6031b6d2e
5 files changed
+178 -75

No files matched your search

+5
View File
@@ -371,6 +371,11 @@ wuffs_base__utf_8__encode(wuffs_base__slice_u8 dst, uint32_t code_point);
WUFFS_BASE__MAYBE_STATIC wuffs_base__utf_8__next__output //
wuffs_base__utf_8__next(wuffs_base__slice_u8 s);
// wuffs_base__utf_8__next_from_end is like wuffs_base__utf_8__next except that
// it looks at the end of s instead of the start.
WUFFS_BASE__MAYBE_STATIC wuffs_base__utf_8__next__output //
wuffs_base__utf_8__next_from_end(wuffs_base__slice_u8 s);
// wuffs_base__utf_8__longest_valid_prefix returns the largest n such that the
// sub-slice s[..n] is valid UTF-8.
//
+33
View File
@@ -167,6 +167,39 @@ wuffs_base__utf_8__next(wuffs_base__slice_u8 s) {
WUFFS_BASE__UNICODE_REPLACEMENT_CHARACTER, 1);
}
WUFFS_BASE__MAYBE_STATIC wuffs_base__utf_8__next__output //
wuffs_base__utf_8__next_from_end(wuffs_base__slice_u8 s) {
if (s.len == 0) {
return wuffs_base__make_utf_8__next__output(0, 0);
}
uint8_t* ptr = &s.ptr[s.len - 1];
if (*ptr < 0x80) {
return wuffs_base__make_utf_8__next__output(*ptr, 1);
} else if (*ptr < 0xC0) {
uint8_t* too_far = &s.ptr[(s.len > 4) ? (s.len - 4) : 0];
uint32_t n = 1;
while (ptr != too_far) {
ptr--;
n++;
if (*ptr < 0x80) {
break;
} else if (*ptr < 0xC0) {
continue;
}
wuffs_base__utf_8__next__output o =
wuffs_base__utf_8__next(wuffs_base__make_slice_u8(ptr, n));
if (o.byte_length != n) {
break;
}
return o;
}
}
return wuffs_base__make_utf_8__next__output(
WUFFS_BASE__UNICODE_REPLACEMENT_CHARACTER, 1);
}
WUFFS_BASE__MAYBE_STATIC size_t //
wuffs_base__utf_8__longest_valid_prefix(wuffs_base__slice_u8 s) {
// TODO: possibly optimize the all-ASCII case (4 or 8 bytes at a time).
+5 -4
View File
@@ -209,8 +209,9 @@ const BaseUTF8SubmoduleC = "" +
" 6 7\n // 8 9 A B C D E F\n 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, // 0x00 ..= 0x07.\n 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, // 0x08 ..= 0x0F.\n 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, // 0x10 ..= 0x17.\n 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, // 0x18 ..= 0x1F.\n 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, // 0x20 ..= 0x27.\n 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, // 0x28 ..= 0x2F.\n 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, // 0x30 ..= 0x37.\n 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, // 0x38 ..= 0x3F.\n\n 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, // 0x40 ..= 0x47.\n 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, // 0x48 ..= 0x4F.\n 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, // 0x50 ..= 0x57.\n 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, // 0x58 ..= 0x5F.\n 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, // 0x60 ..= 0x67.\n 0x00, 0x00, 0x00, 0x00, 0x00, 0x00" +
", 0x00, 0x00, // 0x68 ..= 0x6F.\n 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, // 0x70 ..= 0x77.\n 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, // 0x78 ..= 0x7F.\n\n 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, // 0x80 ..= 0x87.\n 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, // 0x88 ..= 0x8F.\n 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, // 0x90 ..= 0x97.\n 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, // 0x98 ..= 0x9F.\n 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, // 0xA0 ..= 0xA7.\n 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, // 0xA8 ..= 0xAF.\n 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, // 0xB0 ..= 0xB7.\n 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, 0x40, // 0xB8 ..= 0xBF.\n\n 0x80, 0x80, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, // 0xC0 ..= 0xC7.\n 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, // 0xC8 ..= 0xCF.\n 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, // 0xD0 ..= 0xD7.\n 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, // 0xD8 ..= 0" +
"xDF.\n 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, // 0xE0 ..= 0xE7.\n 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, // 0xE8 ..= 0xEF.\n 0x03, 0x03, 0x03, 0x03, 0x03, 0x80, 0x80, 0x80, // 0xF0 ..= 0xF7.\n 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, // 0xF8 ..= 0xFF.\n // 0 1 2 3 4 5 6 7\n // 8 9 A B C D E F\n};\n\nWUFFS_BASE__MAYBE_STATIC wuffs_base__utf_8__next__output //\nwuffs_base__utf_8__next(wuffs_base__slice_u8 s) {\n if (s.len == 0) {\n return wuffs_base__make_utf_8__next__output(0, 0);\n }\n uint32_t c = s.ptr[0];\n switch (wuffs_base__utf_8__byte_length_minus_1[c & 0xFF]) {\n case 0:\n return wuffs_base__make_utf_8__next__output(c, 1);\n\n case 1:\n if (s.len < 2) {\n break;\n }\n c = wuffs_base__load_u16le__no_bounds_check(s.ptr);\n if ((c & 0xC000) != 0x8000) {\n break;\n }\n c = (0x0007C0 & (c << 6)) | (0x00003F & (c >> 8));\n return wuffs_base__make_utf_8__next__output(" +
"c, 2);\n\n case 2:\n if (s.len < 3) {\n break;\n }\n c = wuffs_base__load_u24le__no_bounds_check(s.ptr);\n if ((c & 0xC0C000) != 0x808000) {\n break;\n }\n c = (0x00F000 & (c << 12)) | (0x000FC0 & (c >> 2)) |\n (0x00003F & (c >> 16));\n if ((c <= 0x07FF) || ((0xD800 <= c) && (c <= 0xDFFF))) {\n break;\n }\n return wuffs_base__make_utf_8__next__output(c, 3);\n\n case 3:\n if (s.len < 4) {\n break;\n }\n c = wuffs_base__load_u32le__no_bounds_check(s.ptr);\n if ((c & 0xC0C0C000) != 0x80808000) {\n break;\n }\n c = (0x1C0000 & (c << 18)) | (0x03F000 & (c << 4)) |\n (0x000FC0 & (c >> 10)) | (0x00003F & (c >> 24));\n if ((c <= 0xFFFF) || (0x110000 <= c)) {\n break;\n }\n return wuffs_base__make_utf_8__next__output(c, 4);\n }\n\n return wuffs_base__make_utf_8__next__output(\n WUFFS_BASE__UNICODE_REPLACEMENT_CHARACTER, 1);\n}\n\nWUFFS_BASE__MAYBE_STATIC size_t //\nwuffs_base__utf_8__longest_va" +
"lid_prefix(wuffs_base__slice_u8 s) {\n // TODO: possibly optimize the all-ASCII case (4 or 8 bytes at a time).\n //\n // TODO: possibly optimize this by manually inlining the\n // wuffs_base__utf_8__next calls.\n size_t original_len = s.len;\n while (s.len > 0) {\n wuffs_base__utf_8__next__output o = wuffs_base__utf_8__next(s);\n if ((o.code_point > 0x7F) && (o.byte_length == 1)) {\n break;\n }\n s.ptr += o.byte_length;\n s.len -= o.byte_length;\n }\n return original_len - s.len;\n}\n\nWUFFS_BASE__MAYBE_STATIC size_t //\nwuffs_base__ascii__longest_valid_prefix(wuffs_base__slice_u8 s) {\n // TODO: possibly optimize this by checking 4 or 8 bytes at a time.\n uint8_t* original_ptr = s.ptr;\n uint8_t* p = s.ptr;\n uint8_t* q = s.ptr + s.len;\n for (; (p != q) && ((*p & 0x80) == 0); p++) {\n }\n return (size_t)(p - original_ptr);\n}\n" +
"c, 2);\n\n case 2:\n if (s.len < 3) {\n break;\n }\n c = wuffs_base__load_u24le__no_bounds_check(s.ptr);\n if ((c & 0xC0C000) != 0x808000) {\n break;\n }\n c = (0x00F000 & (c << 12)) | (0x000FC0 & (c >> 2)) |\n (0x00003F & (c >> 16));\n if ((c <= 0x07FF) || ((0xD800 <= c) && (c <= 0xDFFF))) {\n break;\n }\n return wuffs_base__make_utf_8__next__output(c, 3);\n\n case 3:\n if (s.len < 4) {\n break;\n }\n c = wuffs_base__load_u32le__no_bounds_check(s.ptr);\n if ((c & 0xC0C0C000) != 0x80808000) {\n break;\n }\n c = (0x1C0000 & (c << 18)) | (0x03F000 & (c << 4)) |\n (0x000FC0 & (c >> 10)) | (0x00003F & (c >> 24));\n if ((c <= 0xFFFF) || (0x110000 <= c)) {\n break;\n }\n return wuffs_base__make_utf_8__next__output(c, 4);\n }\n\n return wuffs_base__make_utf_8__next__output(\n WUFFS_BASE__UNICODE_REPLACEMENT_CHARACTER, 1);\n}\n\nWUFFS_BASE__MAYBE_STATIC wuffs_base__utf_8__next__output //\nwuff" +
"s_base__utf_8__next_from_end(wuffs_base__slice_u8 s) {\n if (s.len == 0) {\n return wuffs_base__make_utf_8__next__output(0, 0);\n }\n uint8_t* ptr = &s.ptr[s.len - 1];\n if (*ptr < 0x80) {\n return wuffs_base__make_utf_8__next__output(*ptr, 1);\n\n } else if (*ptr < 0xC0) {\n uint8_t* too_far = &s.ptr[(s.len > 4) ? (s.len - 4) : 0];\n uint32_t n = 1;\n while (ptr != too_far) {\n ptr--;\n n++;\n if (*ptr < 0x80) {\n break;\n } else if (*ptr < 0xC0) {\n continue;\n }\n wuffs_base__utf_8__next__output o =\n wuffs_base__utf_8__next(wuffs_base__make_slice_u8(ptr, n));\n if (o.byte_length != n) {\n break;\n }\n return o;\n }\n }\n\n return wuffs_base__make_utf_8__next__output(\n WUFFS_BASE__UNICODE_REPLACEMENT_CHARACTER, 1);\n}\n\nWUFFS_BASE__MAYBE_STATIC size_t //\nwuffs_base__utf_8__longest_valid_prefix(wuffs_base__slice_u8 s) {\n // TODO: possibly optimize the all-ASCII case (4 or 8 bytes at a time).\n //\n // TODO: possibly optimize this b" +
"y manually inlining the\n // wuffs_base__utf_8__next calls.\n size_t original_len = s.len;\n while (s.len > 0) {\n wuffs_base__utf_8__next__output o = wuffs_base__utf_8__next(s);\n if ((o.code_point > 0x7F) && (o.byte_length == 1)) {\n break;\n }\n s.ptr += o.byte_length;\n s.len -= o.byte_length;\n }\n return original_len - s.len;\n}\n\nWUFFS_BASE__MAYBE_STATIC size_t //\nwuffs_base__ascii__longest_valid_prefix(wuffs_base__slice_u8 s) {\n // TODO: possibly optimize this by checking 4 or 8 bytes at a time.\n uint8_t* original_ptr = s.ptr;\n uint8_t* p = s.ptr;\n uint8_t* q = s.ptr + s.len;\n for (; (p != q) && ((*p & 0x80) == 0); p++) {\n }\n return (size_t)(p - original_ptr);\n}\n" +
""
const BaseFundamentalPrivateH = "" +
@@ -468,8 +469,8 @@ const BaseStrConvPublicH = "" +
"FFF));\n case 4:\n return (0x00010000 <= cp) && (cp <= 0x0010FFFF);\n }\n }\n return false;\n}\n\n#ifdef __cplusplus\n\ninline bool //\nwuffs_base__utf_8__next__output::is_valid() const {\n return wuffs_base__utf_8__next__output__is_valid(this);\n}\n\n#endif // __cplusplus\n\n" +
"" +
"// --------\n\n// wuffs_base__utf_8__encode writes the UTF-8 encoding of code_point to s and\n// returns the number of bytes written. If code_point is invalid, or if s is\n// shorter than the entire encoding, it returns 0 (and no bytes are written).\n//\n// s will never be too short if its length is at least 4, also known as\n// WUFFS_BASE__UTF_8__BYTE_LENGTH__MAX_INCL.\nWUFFS_BASE__MAYBE_STATIC size_t //\nwuffs_base__utf_8__encode(wuffs_base__slice_u8 dst, uint32_t code_point);\n\n// wuffs_base__utf_8__next returns the next UTF-8 code point (and that code\n// point's byte length) at the start of s.\n//\n// There are exactly two cases in which this function returns something where\n// wuffs_base__utf_8__next__output__is_valid is false:\n// - If s is empty then it returns {.code_point=0, .byte_length=0}.\n// - If s is non-empty and starts with invalid UTF-8 then it returns\n// {.code_point=WUFFS_BASE__UNICODE_REPLACEMENT_CHARACTER, .byte_length=1}.\n//\n// Otherwise, it returns something where\n// wuffs_base__utf_8__next__ou" +
"tput__is_valid is true.\n//\n// In any case, it always returns an output that satisfies both of:\n// - (output.code_point <= WUFFS_BASE__UNICODE_CODE_POINT__MAX_INCL).\n// - (output.byte_length <= s.len).\n//\n// If s is a sub-slice of a larger slice of valid UTF-8, but that sub-slice\n// boundary occurs in the middle of a multi-byte UTF-8 encoding of a single\n// code point, then this function may return something invalid. It is the\n// caller's responsibility to split on or otherwise manage UTF-8 boundaries.\nWUFFS_BASE__MAYBE_STATIC wuffs_base__utf_8__next__output //\nwuffs_base__utf_8__next(wuffs_base__slice_u8 s);\n\n// wuffs_base__utf_8__longest_valid_prefix returns the largest n such that the\n// sub-slice s[..n] is valid UTF-8.\n//\n// In particular, it returns s.len if and only if all of s is valid UTF-8.\n//\n// If s is a sub-slice of a larger slice of valid UTF-8, but that sub-slice\n// boundary occurs in the middle of a multi-byte UTF-8 encoding of a single\n// code point, then this function will return less than" +
" s.len. It is the\n// caller's responsibility to split on or otherwise manage UTF-8 boundaries.\nWUFFS_BASE__MAYBE_STATIC size_t //\nwuffs_base__utf_8__longest_valid_prefix(wuffs_base__slice_u8 s);\n\n// wuffs_base__ascii__longest_valid_prefix returns the largest n such that the\n// sub-slice s[..n] is valid ASCII.\n//\n// In particular, it returns s.len if and only if all of s is valid ASCII.\n// Equivalently, when none of the bytes in s have the 0x80 high bit set.\nWUFFS_BASE__MAYBE_STATIC size_t //\nwuffs_base__ascii__longest_valid_prefix(wuffs_base__slice_u8 s);\n" +
"tput__is_valid is true.\n//\n// In any case, it always returns an output that satisfies both of:\n// - (output.code_point <= WUFFS_BASE__UNICODE_CODE_POINT__MAX_INCL).\n// - (output.byte_length <= s.len).\n//\n// If s is a sub-slice of a larger slice of valid UTF-8, but that sub-slice\n// boundary occurs in the middle of a multi-byte UTF-8 encoding of a single\n// code point, then this function may return something invalid. It is the\n// caller's responsibility to split on or otherwise manage UTF-8 boundaries.\nWUFFS_BASE__MAYBE_STATIC wuffs_base__utf_8__next__output //\nwuffs_base__utf_8__next(wuffs_base__slice_u8 s);\n\n// wuffs_base__utf_8__next_from_end is like wuffs_base__utf_8__next except that\n// it looks at the end of s instead of the start.\nWUFFS_BASE__MAYBE_STATIC wuffs_base__utf_8__next__output //\nwuffs_base__utf_8__next_from_end(wuffs_base__slice_u8 s);\n\n// wuffs_base__utf_8__longest_valid_prefix returns the largest n such that the\n// sub-slice s[..n] is valid UTF-8.\n//\n// In particular, it returns s.len " +
"if and only if all of s is valid UTF-8.\n//\n// If s is a sub-slice of a larger slice of valid UTF-8, but that sub-slice\n// boundary occurs in the middle of a multi-byte UTF-8 encoding of a single\n// code point, then this function will return less than s.len. It is the\n// caller's responsibility to split on or otherwise manage UTF-8 boundaries.\nWUFFS_BASE__MAYBE_STATIC size_t //\nwuffs_base__utf_8__longest_valid_prefix(wuffs_base__slice_u8 s);\n\n// wuffs_base__ascii__longest_valid_prefix returns the largest n such that the\n// sub-slice s[..n] is valid ASCII.\n//\n// In particular, it returns s.len if and only if all of s is valid ASCII.\n// Equivalently, when none of the bytes in s have the 0x80 high bit set.\nWUFFS_BASE__MAYBE_STATIC size_t //\nwuffs_base__ascii__longest_valid_prefix(wuffs_base__slice_u8 s);\n" +
""
const BaseTokenPrivateH = "" +