pike.git/ src/ builtin.cmod

2010-03-02

2010-03-02 20:14:02 by Martin Nilsson <mani@lysator.liu.se>

• 4d3288e0b9d25448136a14f9391828aada5ddd32 (45 lines) (+33/-12) ^{[ Show | Annotate ]}
  Branch: 7.9

String.trim_all_whites 10% faster and includes all Unicode.

Rev: CHANGES:1.208
Rev: src/builtin.cmod:1.254

2:   || This file is part of Pike. For copyright information see COPYRIGHT.   || Pike is distributed under GPL, LGPL and MPL. See the file COPYING   || for more information. - || $Id: builtin.cmod,v 1.253 2010/01/11 16:30:48 srb Exp $ + || $Id: builtin.cmod,v 1.254 2010/03/02 20:12:49 nilsson Exp $   */      #include "global.h"
907:    *! @belongs String    *!    *! Trim leading and trailing white spaces characters (space, tab, -  *! newline and carriage return) from the string @[s]. +  *! newline, carriage return, form feed, vertical tab and all the +  *! white spaces defined in Unicode) from the string @[s].    */   PMOD_EXPORT   PIKEFUN string string_trim_all_whites (string s)
917:    ptrdiff_t start = 0, end = s->len;    int chr;    switch (s->size_shift) { - #define DO_IT(TYPE) \ + #define SPACECASE8 \ +  case ' ':case '\t':case '\r':case '\n':case '\v':case '\f': \ +  case 0x85:case 0xa0: + #include "whitespace.h" +  + #define DO_IT(TYPE,CASE) \    { \ -  for (; start < s->len; start++) { \ +  for (; start < end; start++) { \    chr = ((TYPE *) s->str)[start]; \ -  if (chr != ' ' && chr != '\t' && chr != '\n' && chr != '\r') \ -  break; \ +  switch(chr) { \ +  CASE \ +  continue; \ +  } \ +  break; \    } \    while (--end > start) { \    chr = ((TYPE *) s->str)[end]; \ -  if (chr != ' ' && chr != '\t' && chr != '\n' && chr != '\r') \ -  break; \ +  switch(chr) { \ +  CASE \ +  continue; \ +  } \ +  break; \    } \    } -  case 0: DO_IT (p_wchar0); break; -  case 1: DO_IT (p_wchar1); break; -  case 2: DO_IT (p_wchar2); break; +  case 0: DO_IT (p_wchar0,SPACECASE8); break; +  case 1: DO_IT (p_wchar1,SPACECASE16); break; +  case 2: DO_IT (p_wchar2,SPACECASE16); break;   #undef DO_IT -  + #undef SPACECASE8 + #undef SPACECASE16    }    RETURN string_slice (s, start, end + 1 - start);   }
2366:    struct Buffer_struct *str = THIS;    if(!str->str.s)    init_string_builder_alloc(&str->str, str->initial, 0); +  + #if 0 +  if( c>=0 && c<256 && str->str.s->len+1<str->str.malloced ) +  STR0(str->str.s)[ str->str.s->len++ ] = c; +  else + #endif    string_builder_putchar(&str->str, c);    }