utf8towc() has to be in lib.c if strlower() is going to use it, because
scripts/*.c builds against lib.c but not linestack.c.
diff --git a/lib/lib.c b/lib/lib.c
index d011af0..c482dca 100644
--- a/lib/lib.c
+++ b/lib/lib.c
@@ -335,6 +335,38 @@
return off-haystack;
}
+// Convert utf8 sequence to a unicode wide character
+int utf8towc(wchar_t *wc, char *str, unsigned len)
+{
+ unsigned result, mask, first;
+ char *s, c;
+
+ // fast path ASCII
+ if (len && *str<128) return !!(*wc = *str);
+
+ result = first = *(s = str++);
+ for (mask = 6; (first&0xc0)==0xc0; mask += 5, first <<= 1) {
+ if (mask>21) return -1;
+ if (!--len) return -2;
+ c = *(str++);
+ if ((c&0xc0) != 0x80) return -1;
+ result = (result<<6)|(c&0x3f);
+ }
+ result &= (1<<mask)-1;
+ c = str-s;
+ if (mask==6) return -1;
+
+ // Avoid overlong encodings
+ if (mask==6 || mask>21 || result<(unsigned []){0x80,0x800,0x10000}[c-2])
+ return -1;
+
+ // Limit unicode so it can't encode anything UTF-16 can't.
+ if (result>0x10ffff || (result>=0xd800 && result<=0xdfff)) return -1;
+ *wc = result;
+
+ return str-s;
+}
+
char *strlower(char *s)
{
char *try, *new;
@@ -348,7 +380,7 @@
while (*s) {
wchar_t c;
- int len = mbrtowc(&c, s, MB_CUR_MAX, 0);
+ int len = utf8towc(&c, s, MB_CUR_MAX);
if (len < 1) *(new++) = *(s++);
else {
diff --git a/lib/lib.h b/lib/lib.h
index 3756652..9325a89 100644
--- a/lib/lib.h
+++ b/lib/lib.h
@@ -204,6 +204,7 @@
long long atolx(char *c);
long long atolx_range(char *numstr, long long low, long long high);
int stridx(char *haystack, char needle);
+int utf8towc(wchar_t *wc, char *str, unsigned len);
char *strlower(char *s);
char *strafter(char *haystack, char *needle);
char *chomp(char *s);
@@ -257,7 +258,6 @@
void linestack_insert(struct linestack **lls, long pos, char *line, long len);
void linestack_append(struct linestack **lls, char *line);
struct linestack *linestack_load(char *name);
-int utf8towc(wchar_t *wc, char *str, unsigned len);
int crunch_escape(FILE *out, int cols, int wc);
int crunch_rev_escape(FILE *out, int cols, int wc);
int crunch_str(char **str, int width, FILE *out, char *escmore,
diff --git a/lib/linestack.c b/lib/linestack.c
index 4466710..91dec56 100644
--- a/lib/linestack.c
+++ b/lib/linestack.c
@@ -80,37 +80,6 @@
return ls;
}
-// Convert utf8 sequence to a unicode wide character
-int utf8towc(wchar_t *wc, char *str, unsigned len)
-{
- unsigned result, mask, first;
- char *s, c;
-
- // fast path ASCII
- if (len && *str<128) return !!(*wc = *str);
-
- result = first = *(s = str++);
- for (mask = 6; (first&0xc0)==0xc0; mask += 5, first <<= 1) {
- if (!--len) return -2;
- c = *(str++);
- if ((c&0xc0) != 0x80) return -1;
- result = (result<<6)|(c&0x3f);
- }
- result &= (1<<mask)-1;
- c = str-s;
- if (mask==6 || mask>21) return -1;
-
- // Avoid overlong encodings
- if (mask==6 || mask>21 || result<(unsigned []){0x80,0x800,0x10000}[c-2])
- return -1;
-
- // Limit unicode so it can't encode anything UTF-16 can't.
- if (result>0x10ffff || (result>=0xd800 && result<=0xdfff)) return -1;
- *wc = result;
-
- return str-s;
-}
-
// Show width many columns, negative means from right edge, out=0 just measure
// if escout, send it unprintable chars, otherwise pass through raw data.
// Returns width in columns, moves *str to end of data consumed.
@@ -123,7 +92,7 @@
for (end = start = *str; *end; columns += col, end += bytes) {
wchar_t wc;
- if ((bytes = mbrtowc(&wc, end, MB_CUR_MAX, 0))>0 && (col = wcwidth(wc))>=0)
+ if ((bytes = utf8towc(&wc, end, 4))>0 && (col = wcwidth(wc))>=0)
{
if (!escmore || wc>255 || !strchr(escmore, wc)) {
if (width-columns<col) break;
