Changeset f444633 in mainline for common/stdc/uchar.c

Timestamp:

2025-04-15T22:42:20Z (3 days ago)

Author:

GitHub <noreply@…>

Children:

1d3ae66

Parents:

b6061f8c (diff), 65bf084 (diff)
Note: this is a merge changeset, the changes displayed below correspond to the merge itself.
Use the (diff) links above to see all the changes relative to each parent.

git-author:

Wayne Thornton <wmthornton-dev@…> (2025-04-15 22:42:20)

git-committer:

GitHub <noreply@…> (2025-04-15 22:42:20)

Message:

Merge branch 'HelenOS:master' into master

File:

• common/stdc/uchar.c (modified) (9 diffs)

common/stdc/uchar.c

@@ -59 +59 @@
 }
 
-static bool _is_continuation(uint8_t c)
-{
-        return (c & 0xC0) == 0x80;
-}
-
-static bool _is_1_byte(uint8_t c)
-{
-        return (c & 0x80) == 0;
-}
-
-static bool _is_2_byte(uint8_t c)
-{
-        return (c & 0xE0) == 0xC0;
-}
-
-static bool _is_3_byte(uint8_t c)
-{
-        return (c & 0xF0) == 0xE0;
-}
-
-static bool _is_4_byte(uint8_t c)
-{
-        return (c & 0xF8) == 0xF0;
-}
-
-static bool _is_non_shortest(unsigned short cont, uint8_t b)
-{
-        return (cont == 0b1111110000000000 && !(b & 0b00100000)) ||
-            (cont == 0b1111111111110000 && !(b & 0b00110000));
-}
-
-size_t mbrtoc32(char32_t *c, const char *s, size_t n, mbstate_t *mb)
-{
-#if __STDC_HOSTED__
-        static fibril_local mbstate_t global_state = { };
-
-        if (!mb)
-                mb = &global_state;
-#else
-        assert(mb);
-#endif
-
-        if (n == 0)
-                return UCHAR_INCOMPLETE;
-
-        char32_t dummy;
-
-        if (!c)
-                c = &dummy;
-
-        if (!s) {
-                // Equivalent to mbrtoc32(NULL, "", 1, mb).
-                if (mb->continuation) {
-                        _set_ilseq();
-                        return UCHAR_ILSEQ;
-                } else {
-                        return 0;
-                }
-        }
-
-        size_t i = 0;
-
-        if (!mb->continuation) {
-                /* Clean slate, read initial byte. */
-
-                uint8_t b = s[i++];
-
-                if (_is_1_byte(b)) {
-                        *c = b;
-                        return b == 0 ? 0 : 1;
-                }
-
-                if (_is_continuation(b)) {
-                        /* unexpected continuation byte */
-                        _set_ilseq();
-                        return UCHAR_ILSEQ;
-                }
-
-                /*
-                 * The value stored into `continuation` is designed to have
-                 * just enough leading ones that after shifting in one less than
-                 * the expected number of continuation bytes, the most significant
-                 * bit becomes zero. (The field is 16b wide.)
-                 */
-
-                if (_is_2_byte(b)) {
-                        /* Reject non-shortest form. */
-                        if (!(b & 0b00011110)) {
-                                _set_ilseq();
-                                return UCHAR_ILSEQ;
-                        }
-
-                        /* 2 byte encoding               110xxxxx */
-                        mb->continuation = b ^ 0b0000000011000000;
-
-                } else if (_is_3_byte(b)) {
-                        /* 3 byte encoding               1110xxxx */
-                        mb->continuation = b ^ 0b1111110011100000;
-
-                } else if (_is_4_byte(b)) {
-                        /* 4 byte encoding               11110xxx */
-                        mb->continuation = b ^ 0b1111111100000000;
-                }
-        }
-
-        for (; i < n; i++) {
-                /* Read continuation bytes. */
-                uint8_t b = s[i];
-
-                if (!_is_continuation(b) || _is_non_shortest(mb->continuation, b)) {
-                        _set_ilseq();
-                        return UCHAR_ILSEQ;
-                }
-
-                /* Top bit becomes zero just before the last byte is shifted in. */
-                if (!(mb->continuation & 0x8000)) {
-                        *c = ((char32_t) mb->continuation) << 6 | (b & 0x3f);
-                        mb->continuation = 0;
-                        return ++i;
-                }
-
-                mb->continuation = mb->continuation << 6 | (b & 0x3f);
-        }
-
-        return UCHAR_INCOMPLETE;
-}
-
 #define UTF8_CONT(c, shift) (0x80 | (((c) >> (shift)) & 0x3F))
 
@@ -253 +126 @@
         if (!s) {
                 /* Equivalent to mbrtoc16(NULL, "", 1, mb). */
-                if (mb->continuation) {
+                if (mb->state) {
                         _set_ilseq();
                         return UCHAR_ILSEQ;
@@ -261 +134 @@
         }
 
-        if ((mb->continuation & 0xD000) == 0xD000) {
+        if ((mb->state & 0xD000) == 0xD000) {
                 /* mbstate_t contains the second surrogate character. */
                 /* mbrtoc32() will never set it to such value.        */
-                *c = mb->continuation;
-                mb->continuation = 0;
+                *c = mb->state;
+                mb->state = 0;
                 return UCHAR_CONTINUED;
         }
@@ -276 +149 @@
                 } else {
                         /* Encode UTF-16 surrogates. */
-                        mb->continuation = (c32 & 0x3FF) + 0xDC00;
+                        mb->state = (c32 & 0x3FF) + 0xDC00;
                         *c = (c32 >> 10) + 0xD7C0;
                 }
@@ -298 +171 @@
         if (!s) {
                 // Equivalent to c16rtomb(buf, L’\0’, mb).
-                if (mb->continuation) {
+                if (mb->state) {
                         _set_ilseq();
                         return UCHAR_ILSEQ;
@@ -307 +180 @@
 
         if (!_is_surrogate(c)) {
-                if (mb->continuation) {
+                if (mb->state) {
                         _set_ilseq();
                         return UCHAR_ILSEQ;
@@ -315 +188 @@
         }
 
-        if (!mb->continuation) {
-                mb->continuation = c;
+        if (!mb->state) {
+                mb->state = c;
                 return 0;
         }
@@ -323 +196 @@
 
         /* Decode UTF-16 surrogates. */
-        if (_is_low_surrogate(mb->continuation) && _is_high_surrogate(c)) {
-                c32 = ((c - 0xD7C0) << 10) | (mb->continuation - 0xDC00);
-        } else if (_is_high_surrogate(mb->continuation) && _is_low_surrogate(c)) {
-                c32 = ((mb->continuation - 0xD7C0) << 10) | (c - 0xDC00);
+        if (_is_low_surrogate(mb->state) && _is_high_surrogate(c)) {
+                c32 = ((c - 0xD7C0) << 10) | (mb->state - 0xDC00);
+        } else if (_is_high_surrogate(mb->state) && _is_low_surrogate(c)) {
+                c32 = ((mb->state - 0xD7C0) << 10) | (c - 0xDC00);
         } else {
                 _set_ilseq();
@@ -332 +205 @@
         }
 
-        mb->continuation = 0;
+        mb->state = 0;
         return c32rtomb(s, c32, mb);
 }