commit 51e32d49b56c86cd288c64fccf6cd765547781b9
from: NRK <nrk@disroot.org>
via: Hiltjo Posthuma <hiltjo@codemadness.org>
date: Sun Jul 14 09:42:55 2024 UTC

overhaul utf8decode()

this changes the utf8decode function to:

* report when an error occurs
* report how many bytes to advance on error

these will be useful in the next commit to render invalid utf8
sequences.

the new implementation is also shorter and more direct.

commit - 7be720cc88ed2294338f7182600df10f21c575ce
commit + 51e32d49b56c86cd288c64fccf6cd765547781b9
blob - 78a2b27211ab59ad312efbb11e1a0c5146e09555
blob + eb71da7f08e52722b11c70faf1a486bed2e59bb1
--- drw.c
+++ drw.c
@@ -9,54 +9,40 @@
 #include "util.h"
 
 #define UTF_INVALID 0xFFFD
-#define UTF_SIZ     4
 
-static const unsigned char utfbyte[UTF_SIZ + 1] = {0x80,    0, 0xC0, 0xE0, 0xF0};
-static const unsigned char utfmask[UTF_SIZ + 1] = {0xC0, 0x80, 0xE0, 0xF0, 0xF8};
-static const long utfmin[UTF_SIZ + 1] = {       0,    0,  0x80,  0x800,  0x10000};
-static const long utfmax[UTF_SIZ + 1] = {0x10FFFF, 0x7F, 0x7FF, 0xFFFF, 0x10FFFF};
-
-static long
-utf8decodebyte(const char c, size_t *i)
+static int
+utf8decode(const char *s_in, long *u, int *err)
 {
-	for (*i = 0; *i < (UTF_SIZ + 1); ++(*i))
-		if (((unsigned char)c & utfmask[*i]) == utfbyte[*i])
-			return (unsigned char)c & ~utfmask[*i];
-	return 0;
-}
+	static const unsigned char lens[] = {
+		/* 0XXXX */ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
+		/* 10XXX */ 0, 0, 0, 0, 0, 0, 0, 0,  /* invalid */
+		/* 110XX */ 2, 2, 2, 2,
+		/* 1110X */ 3, 3,
+		/* 11110 */ 4,
+		/* 11111 */ 0,  /* invalid */
+	};
+	static const unsigned char leading_mask[] = { 0x7F, 0x1F, 0x0F, 0x07 };
+	static const unsigned int overlong[] = { 0x0, 0x80, 0x0800, 0x10000 };
 
-static size_t
-utf8validate(long *u, size_t i)
-{
-	if (!BETWEEN(*u, utfmin[i], utfmax[i]) || BETWEEN(*u, 0xD800, 0xDFFF))
-		*u = UTF_INVALID;
-	for (i = 1; *u > utfmax[i]; ++i)
-		;
-	return i;
-}
-
-static size_t
-utf8decode(const char *c, long *u, size_t clen)
-{
-	size_t i, j, len, type;
-	long udecoded;
-
+	const unsigned char *s = (const unsigned char *)s_in;
+	int len = lens[*s >> 3];
 	*u = UTF_INVALID;
-	if (!clen)
-		return 0;
-	udecoded = utf8decodebyte(c[0], &len);
-	if (!BETWEEN(len, 1, UTF_SIZ))
+	*err = 1;
+	if (len == 0)
 		return 1;
-	for (i = 1, j = 1; i < clen && j < len; ++i, ++j) {
-		udecoded = (udecoded << 6) | utf8decodebyte(c[i], &type);
-		if (type)
-			return j;
+
+	long cp = s[0] & leading_mask[len - 1];
+	for (int i = 1; i < len; ++i) {
+		if (s[i] == '\0' || (s[i] & 0xC0) != 0x80)
+			return i;
+		cp = (cp << 6) | (s[i] & 0x3F);
 	}
-	if (j < len)
-		return 0;
-	*u = udecoded;
-	utf8validate(u, len);
+	/* out of range, surrogate, overlong encoding */
+	if (cp > 0x10FFFF || (cp >> 11) == 0x1B || cp < overlong[len - 1])
+		return len;
 
+	*err = 0;
+	*u = cp;
 	return len;
 }
 
@@ -242,7 +228,7 @@ drw_text(Drw *drw, int x, int y, unsigned int w, unsig
 	unsigned int tmpw, ew, ellipsis_w = 0, ellipsis_len, hash, h0, h1;
 	XftDraw *d = NULL;
 	Fnt *usedfont, *curfont, *nextfont;
-	int utf8strlen, utf8charlen, render = x || y || w || h;
+	int utf8strlen, utf8charlen, utf8err, render = x || y || w || h;
 	long utf8codepoint = 0;
 	const char *utf8str;
 	FcCharSet *fccharset;
@@ -272,11 +258,11 @@ drw_text(Drw *drw, int x, int y, unsigned int w, unsig
 	if (!ellipsis_width && render)
 		ellipsis_width = drw_fontset_getwidth(drw, "...");
 	while (1) {
-		ew = ellipsis_len = utf8strlen = 0;
+		ew = ellipsis_len = utf8err = utf8charlen = utf8strlen = 0;
 		utf8str = text;
 		nextfont = NULL;
 		while (*text) {
-			utf8charlen = utf8decode(text, &utf8codepoint, UTF_SIZ);
+			utf8charlen = utf8decode(text, &utf8codepoint, &utf8err);
 			for (curfont = drw->fonts; curfont; curfont = curfont->next) {
 				charexists = charexists || XftCharExists(drw->dpy, curfont->xfont, utf8codepoint);
 				if (charexists) {