ariadne
/
smol
2
1
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

chars: speed up the handling of invalid UTF-8 starter bytes 

The first byte of a multi-byte UTF-8 sequence must be in the range
0xC2...0xFF.  Any other byte cannot be a starter byte and can thus
immediately be treated as a single byte.
master
Benno Schulenberg 2021-01-06 12:41:49 +01:00
parent 8c406bc875
commit 5129e718d7
2 changed files with 7 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

14
src/chars.c
View File

@ -180,7 +180,7 @@ char control_mbrep(const char *c, bool isdata)
int mbwidth(const char *c) int mbwidth(const char *c)
{ {
/* Ask for the width only when the character isn't plain ASCII. */ /* Ask for the width only when the character isn't plain ASCII. */
if ((signed char)*c <= 0) { if ((unsigned char)*c > 0xC1) {
wchar_t wc; wchar_t wc;
int width; int width;
@ -227,7 +227,7 @@ int char_length(const char *pointer)
{ {
#ifdef ENABLE_UTF8 #ifdef ENABLE_UTF8
/* If possibly a multibyte character, get its length; otherwise, it's 1. */ /* If possibly a multibyte character, get its length; otherwise, it's 1. */
if ((signed char)*pointer < 0) { if ((unsigned char)*pointer > 0xC1) {
int length = mblen(pointer, MAXCHARLEN); int length = mblen(pointer, MAXCHARLEN);
return (length < 0 ? 1 : length); return (length < 0 ? 1 : length);
@ -243,7 +243,7 @@ size_t mbstrlen(const char *pointer)
while (*pointer != '\0') { while (*pointer != '\0') {
#ifdef ENABLE_UTF8 #ifdef ENABLE_UTF8
if ((signed char)*pointer < 0) { if ((unsigned char)*pointer > 0xC1) {
int length = mblen(pointer, MAXCHARLEN); int length = mblen(pointer, MAXCHARLEN);
pointer += (length < 0 ? 1 : length); pointer += (length < 0 ? 1 : length);
@ -265,7 +265,7 @@ int collect_char(const char *string, char *thechar)
#ifdef ENABLE_UTF8 #ifdef ENABLE_UTF8
/* If this is a UTF-8 starter byte, get the number of bytes of the character. */ /* If this is a UTF-8 starter byte, get the number of bytes of the character. */
if ((signed char)*string < 0) { if ((unsigned char)*string > 0xC1) {
charlen = mblen(string, MAXCHARLEN); charlen = mblen(string, MAXCHARLEN);
/* When the multibyte sequence is invalid, only take the first byte. */ /* When the multibyte sequence is invalid, only take the first byte. */
@ -286,7 +286,7 @@ int collect_char(const char *string, char *thechar)
int advance_over(const char *string, size_t *column) int advance_over(const char *string, size_t *column)
{ {
#ifdef ENABLE_UTF8 #ifdef ENABLE_UTF8
if ((signed char)*string < 0) { if ((unsigned char)*string > 0xC1) {
int charlen = mblen(string, MAXCHARLEN); int charlen = mblen(string, MAXCHARLEN);
if (charlen > 0) { if (charlen > 0) {
@ -310,10 +310,8 @@ int advance_over(const char *string, size_t *column)
*column += 2; *column += 2;
} else if (*string == 0x7F) } else if (*string == 0x7F)
*column += 2; *column += 2;
#ifndef ENABLE_UTF8 else if (!use_utf8 && 0x7F < (unsigned char)*string && (unsigned char)*string < 0xA0)
else if (0x7F < (unsigned char)*string && (unsigned char)*string < 0xA0)
*column += 2; *column += 2;
#endif
else else
*column += 1; *column += 1;

2
src/winio.c
View File

@ -2141,7 +2141,7 @@ void minibar(void)
sprintf(hexadecimal, openfile->current->next ? "U+000A" : "------"); sprintf(hexadecimal, openfile->current->next ? "U+000A" : "------");
else if (*this_position == '\n') else if (*this_position == '\n')
sprintf(hexadecimal, "U+0000"); sprintf(hexadecimal, "U+0000");
else if ((unsigned char)*this_position >= 0x80 && else if ((unsigned char)*this_position > 0xC1 &&
mbtowc(&widecode, this_position, MAXCHARLEN) >= 0) mbtowc(&widecode, this_position, MAXCHARLEN) >= 0)
sprintf(hexadecimal, "U+%04X", widecode); sprintf(hexadecimal, "U+%04X", widecode);
else else