| 6 |
and semantics are as close as possible to those of the Perl 5 language. |
and semantics are as close as possible to those of the Perl 5 language. |
| 7 |
|
|
| 8 |
Written by Philip Hazel |
Written by Philip Hazel |
| 9 |
Copyright (c) 1997-2009 University of Cambridge |
Copyright (c) 1997-2012 University of Cambridge |
| 10 |
|
|
| 11 |
----------------------------------------------------------------------------- |
----------------------------------------------------------------------------- |
| 12 |
Redistribution and use in source and binary forms, with or without |
Redistribution and use in source and binary forms, with or without |
| 67 |
type the newline type |
type the newline type |
| 68 |
endptr pointer to the end of the string |
endptr pointer to the end of the string |
| 69 |
lenptr where to return the length |
lenptr where to return the length |
| 70 |
utf8 TRUE if in utf8 mode |
utf TRUE if in utf mode |
| 71 |
|
|
| 72 |
Returns: TRUE or FALSE |
Returns: TRUE or FALSE |
| 73 |
*/ |
*/ |
| 74 |
|
|
| 75 |
BOOL |
BOOL |
| 76 |
_pcre_is_newline(USPTR ptr, int type, USPTR endptr, int *lenptr, BOOL utf8) |
PRIV(is_newline)(PCRE_PUCHAR ptr, int type, PCRE_PUCHAR endptr, int *lenptr, |
| 77 |
|
BOOL utf) |
| 78 |
{ |
{ |
| 79 |
int c; |
int c; |
| 80 |
if (utf8) { GETCHAR(c, ptr); } else c = *ptr; |
(void)utf; |
| 81 |
|
#ifdef SUPPORT_UTF |
| 82 |
|
if (utf) |
| 83 |
|
{ |
| 84 |
|
GETCHAR(c, ptr); |
| 85 |
|
} |
| 86 |
|
else |
| 87 |
|
#endif /* SUPPORT_UTF */ |
| 88 |
|
c = *ptr; |
| 89 |
|
|
| 90 |
if (type == NLTYPE_ANYCRLF) switch(c) |
if (type == NLTYPE_ANYCRLF) switch(c) |
| 91 |
{ |
{ |
| 104 |
case 0x000c: *lenptr = 1; return TRUE; /* FF */ |
case 0x000c: *lenptr = 1; return TRUE; /* FF */ |
| 105 |
case 0x000d: *lenptr = (ptr < endptr - 1 && ptr[1] == 0x0a)? 2 : 1; |
case 0x000d: *lenptr = (ptr < endptr - 1 && ptr[1] == 0x0a)? 2 : 1; |
| 106 |
return TRUE; /* CR */ |
return TRUE; /* CR */ |
| 107 |
case 0x0085: *lenptr = utf8? 2 : 1; return TRUE; /* NEL */ |
#ifdef COMPILE_PCRE8 |
| 108 |
|
case 0x0085: *lenptr = utf? 2 : 1; return TRUE; /* NEL */ |
| 109 |
case 0x2028: /* LS */ |
case 0x2028: /* LS */ |
| 110 |
case 0x2029: *lenptr = 3; return TRUE; /* PS */ |
case 0x2029: *lenptr = 3; return TRUE; /* PS */ |
| 111 |
|
#else |
| 112 |
|
case 0x0085: /* NEL */ |
| 113 |
|
case 0x2028: /* LS */ |
| 114 |
|
case 0x2029: *lenptr = 1; return TRUE; /* PS */ |
| 115 |
|
#endif /* COMPILE_PCRE8 */ |
| 116 |
default: return FALSE; |
default: return FALSE; |
| 117 |
} |
} |
| 118 |
} |
} |
| 131 |
type the newline type |
type the newline type |
| 132 |
startptr pointer to the start of the string |
startptr pointer to the start of the string |
| 133 |
lenptr where to return the length |
lenptr where to return the length |
| 134 |
utf8 TRUE if in utf8 mode |
utf TRUE if in utf mode |
| 135 |
|
|
| 136 |
Returns: TRUE or FALSE |
Returns: TRUE or FALSE |
| 137 |
*/ |
*/ |
| 138 |
|
|
| 139 |
BOOL |
BOOL |
| 140 |
_pcre_was_newline(USPTR ptr, int type, USPTR startptr, int *lenptr, BOOL utf8) |
PRIV(was_newline)(PCRE_PUCHAR ptr, int type, PCRE_PUCHAR startptr, int *lenptr, |
| 141 |
|
BOOL utf) |
| 142 |
{ |
{ |
| 143 |
int c; |
int c; |
| 144 |
|
(void)utf; |
| 145 |
ptr--; |
ptr--; |
| 146 |
#ifdef SUPPORT_UTF8 |
#ifdef SUPPORT_UTF |
| 147 |
if (utf8) |
if (utf) |
| 148 |
{ |
{ |
| 149 |
BACKCHAR(ptr); |
BACKCHAR(ptr); |
| 150 |
GETCHAR(c, ptr); |
GETCHAR(c, ptr); |
| 151 |
} |
} |
| 152 |
else c = *ptr; |
else |
| 153 |
#else /* no UTF-8 support */ |
#endif /* SUPPORT_UTF */ |
| 154 |
c = *ptr; |
c = *ptr; |
|
#endif /* SUPPORT_UTF8 */ |
|
| 155 |
|
|
| 156 |
if (type == NLTYPE_ANYCRLF) switch(c) |
if (type == NLTYPE_ANYCRLF) switch(c) |
| 157 |
{ |
{ |
| 168 |
case 0x000b: /* VT */ |
case 0x000b: /* VT */ |
| 169 |
case 0x000c: /* FF */ |
case 0x000c: /* FF */ |
| 170 |
case 0x000d: *lenptr = 1; return TRUE; /* CR */ |
case 0x000d: *lenptr = 1; return TRUE; /* CR */ |
| 171 |
case 0x0085: *lenptr = utf8? 2 : 1; return TRUE; /* NEL */ |
#ifdef COMPILE_PCRE8 |
| 172 |
|
case 0x0085: *lenptr = utf? 2 : 1; return TRUE; /* NEL */ |
| 173 |
case 0x2028: /* LS */ |
case 0x2028: /* LS */ |
| 174 |
case 0x2029: *lenptr = 3; return TRUE; /* PS */ |
case 0x2029: *lenptr = 3; return TRUE; /* PS */ |
| 175 |
|
#else |
| 176 |
|
case 0x0085: /* NEL */ |
| 177 |
|
case 0x2028: /* LS */ |
| 178 |
|
case 0x2029: *lenptr = 1; return TRUE; /* PS */ |
| 179 |
|
#endif /* COMPILE_PCRE8 */ |
| 180 |
default: return FALSE; |
default: return FALSE; |
| 181 |
} |
} |
| 182 |
} |
} |