/[pcre]/code/trunk/pcre_newline.c
ViewVC logotype

Diff of /code/trunk/pcre_newline.c

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 93 by nigel, Sat Feb 24 21:41:42 2007 UTC revision 305 by ph10, Sun Jan 20 20:07:32 2008 UTC
# Line 6  Line 6 
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) 1997-2006 University of Cambridge             Copyright (c) 1997-2008 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
# Line 42  POSSIBILITY OF SUCH DAMAGE. Line 42  POSSIBILITY OF SUCH DAMAGE.
42  one kind of newline is to be recognized. When a newline is found, its length is  one kind of newline is to be recognized. When a newline is found, its length is
43  returned. In principle, we could implement several newline "types", each  returned. In principle, we could implement several newline "types", each
44  referring to a different set of newline characters. At present, PCRE supports  referring to a different set of newline characters. At present, PCRE supports
45  only NLTYPE_FIXED, which gets handled without these functions, and NLTYPE_ALL,  only NLTYPE_FIXED, which gets handled without these functions, NLTYPE_ANYCRLF,
46  so for now the type isn't passed into the functions. It can easily be added  and NLTYPE_ANY. The full list of Unicode newline characters is taken from
 later if required. The full list of Unicode newline characters is taken from  
47  http://unicode.org/unicode/reports/tr18/. */  http://unicode.org/unicode/reports/tr18/. */
48    
49    
50    #ifdef HAVE_CONFIG_H
51    #include "config.h"
52    #endif
53    
54  #include "pcre_internal.h"  #include "pcre_internal.h"
55    
56    
# Line 61  string that is being processed. Line 64  string that is being processed.
64    
65  Arguments:  Arguments:
66    ptr          pointer to possible newline    ptr          pointer to possible newline
67      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    utf8         TRUE if in utf8 mode
# Line 69  Returns:       TRUE or FALSE Line 73  Returns:       TRUE or FALSE
73  */  */
74    
75  BOOL  BOOL
76  _pcre_is_newline(const uschar *ptr, const uschar *endptr, int *lenptr,  _pcre_is_newline(const uschar *ptr, int type, const uschar *endptr,
77    BOOL utf8)    int *lenptr, BOOL utf8)
78  {  {
79  int c;  int c;
80  if (utf8) { GETCHAR(c, ptr); } else c = *ptr;  if (utf8) { GETCHAR(c, ptr); } else c = *ptr;
81  switch(c)  
82    if (type == NLTYPE_ANYCRLF) switch(c)
83      {
84      case 0x000a: *lenptr = 1; return TRUE;             /* LF */
85      case 0x000d: *lenptr = (ptr < endptr - 1 && ptr[1] == 0x0a)? 2 : 1;
86                   return TRUE;                          /* CR */
87      default: return FALSE;
88      }
89    
90    /* NLTYPE_ANY */
91    
92    else switch(c)
93    {    {
94    case 0x000a:                                       /* LF */    case 0x000a:                                       /* LF */
95    case 0x000b:                                       /* VT */    case 0x000b:                                       /* VT */
# Line 99  the string that is being processed. Line 114  the string that is being processed.
114    
115  Arguments:  Arguments:
116    ptr          pointer to possible newline    ptr          pointer to possible newline
117      type         the newline type
118    startptr     pointer to the start of the string    startptr     pointer to the start of the string
119    lenptr       where to return the length    lenptr       where to return the length
120    utf8         TRUE if in utf8 mode    utf8         TRUE if in utf8 mode
# Line 107  Returns:       TRUE or FALSE Line 123  Returns:       TRUE or FALSE
123  */  */
124    
125  BOOL  BOOL
126  _pcre_was_newline(const uschar *ptr, const uschar *startptr, int *lenptr,  _pcre_was_newline(const uschar *ptr, int type, const uschar *startptr,
127    BOOL utf8)    int *lenptr, BOOL utf8)
128  {  {
129  int c;  int c;
130  ptr--;  ptr--;
131    #ifdef SUPPORT_UTF8
132  if (utf8)  if (utf8)
133    {    {
134    BACKCHAR(ptr);    BACKCHAR(ptr);
135    GETCHAR(c, ptr);    GETCHAR(c, ptr);
136    }    }
137  else c = *ptr;  else c = *ptr;
138  switch(c)  #else   /* no UTF-8 support */
139    c = *ptr;
140    #endif  /* SUPPORT_UTF8 */
141    
142    if (type == NLTYPE_ANYCRLF) switch(c)
143      {
144      case 0x000a: *lenptr = (ptr > startptr && ptr[-1] == 0x0d)? 2 : 1;
145                   return TRUE;                         /* LF */
146      case 0x000d: *lenptr = 1; return TRUE;            /* CR */
147      default: return FALSE;
148      }
149    
150    else switch(c)
151    {    {
152    case 0x000a: *lenptr = (ptr > startptr && ptr[-1] == 0x0d)? 2 : 1;    case 0x000a: *lenptr = (ptr > startptr && ptr[-1] == 0x0d)? 2 : 1;
153                 return TRUE;                         /* LF */                 return TRUE;                         /* LF */
Legend:
Removed from v.93  
changed lines
  Added in v.305
  ViewVC Help
Powered by ViewVC 1.1.5