diff options
Diffstat (limited to 'Source/SPSQLTokenizer.l')
-rw-r--r-- | Source/SPSQLTokenizer.l | 59 |
1 files changed, 3 insertions, 56 deletions
diff --git a/Source/SPSQLTokenizer.l b/Source/SPSQLTokenizer.l index 24dac938..b9170edb 100644 --- a/Source/SPSQLTokenizer.l +++ b/Source/SPSQLTokenizer.l @@ -31,12 +31,12 @@ // More info at <https://github.com/sequelpro/sequelpro> #import "SPSQLTokenizer.h" +#include "SPParserUtils.h" -int utf8strlenfortoken(const char * _s); -int yyuoffset, yyuleng; +size_t yyuoffset, yyuleng; //keep track of the current utf-8 character (not byte) offset and token length -#define YY_USER_ACTION { yyuoffset += yyuleng; yyuleng = utf8strlenfortoken(yytext); } +#define YY_USER_ACTION { yyuoffset += yyuleng; yyuleng = utf8strlen(yytext); } //ignore the output of unmatched characters #define ECHO {} %} @@ -90,56 +90,3 @@ compend {s}"end" return 0; } %% -#define ONEMASK ((size_t)(-1) / 0xFF) -// adapted from http://www.daemonology.net/blog/2008-06-05-faster-utf8-strlen.html -int utf8strlenfortoken(const char * _s) -{ - const char * s; - size_t count = 0; - size_t u; - unsigned char b; - - /* Handle any initial misaligned bytes. */ - for (s = _s; (uintptr_t)(s) & (sizeof(size_t) - 1); s++) { - b = *s; - - /* Exit if we hit a zero byte. */ - if (b == '\0') - goto done; - - /* Is this byte NOT the first byte of a character? */ - count += (b >> 7) & ((~b) >> 6); - } - - /* Handle complete blocks. */ - for (; ; s += sizeof(size_t)) { - /* Prefetch 256 bytes ahead. */ - __builtin_prefetch(&s[256], 0, 0); - - /* Grab 4 or 8 bytes of UTF-8 data. */ - u = *(size_t *)(s); - - /* Exit the loop if there are any zero bytes. */ - if ((u - ONEMASK) & (~u) & (ONEMASK * 0x80)) - break; - - /* Count bytes which are NOT the first byte of a character. */ - u = ((u & (ONEMASK * 0x80)) >> 7) & ((~u) >> 6); - count += (u * ONEMASK) >> ((sizeof(size_t) - 1) * 8); - } - - /* Take care of any left-over bytes. */ - for (; ; s++) { - b = *s; - - /* Exit if we hit a zero byte. */ - if (b == '\0') - break; - - /* Is this byte NOT the first byte of a character? */ - count += (b >> 7) & ((~b) >> 6); - } - -done: - return (int)((s - _s) - count); -} |