epan/strutil.h

   1 /* strutil.h
   2  * String utility definitions
   3  *
   4  * $Id$
   5  *
   6  * Wireshark - Network traffic analyzer
   7  * By Gerald Combs <gerald@wireshark.org>
   8  * Copyright 1998 Gerald Combs
   9  *
  10  * This program is free software; you can redistribute it and/or
  11  * modify it under the terms of the GNU General Public License
  12  * as published by the Free Software Foundation; either version 2
  13  * of the License, or (at your option) any later version.
  14  *
  15  * This program is distributed in the hope that it will be useful,
  16  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  17  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  18  * GNU General Public License for more details.
  19  *
  20  * You should have received a copy of the GNU General Public License
  21  * along with this program; if not, write to the Free Software
  22  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
  23  */
  24
  25 #ifndef __STRUTIL_H__
  26 #define __STRUTIL_H__
  27
  28 #include "ws_symbol_export.h"
  29
  30 #ifdef __cplusplus
  31 extern "C" {
  32 #endif /* __cplusplus */
  33
  34 /* ... thus, config.h needs to be #included */
  35
  36 /** @file
  37  * String handling and conversion utilities.
  38  */
  39
  40 /** Given a pointer into a data buffer, and to the end of the buffer,
  41  *  find the end of the (putative) line at that position in the data
  42  *  buffer.
  43  *
  44  * @param data A pointer to the beginning of the data
  45  * @param dataend A pointer to the end of the data
  46  * @param eol A pointer that will receive the EOL location
  47  * @return A pointer to the EOL character(s) in "*eol".
  48  */
  49 const guchar *find_line_end(const guchar *data, const guchar *dataend,
  50     const guchar **eol);
  51
  52 /** Get the length of the next token in a line, and the beginning of the
  53  *  next token after that (if any).
  54  * @param linep A pointer to the beginning of the line
  55  * @param lineend A pointer to the end of the line
  56  * @param next_token Receives the location of the next token
  57  * @return 0 if there is no next token.
  58  */
  59 WS_DLL_PUBLIC
  60 int        get_token_len(const guchar *linep, const guchar *lineend,
  61     const guchar **next_token);
  62
  63 /** Given a string, generate a string from it that shows non-printable
  64  *  characters as C-style escapes, and return a pointer to it.
  65  *
  66  * @param line A pointer to the input string
  67  * @param len The length of the input string
  68  * @return A pointer to the formatted string
  69  *
  70  * @see tvb_format_text()
  71  */
  72 WS_DLL_PUBLIC
  73 gchar*     format_text(const guchar *line, size_t len);
  74
  75 /**
  76  * Given a string, generate a string from it that shows non-printable
  77  * characters as C-style escapes except a whitespace character
  78  * (space, tab, carriage return, new line, vertical tab, or formfeed)
  79  * which will be replaced by a space, and return a pointer to it.
  80  *
  81  * @param line A pointer to the input string
  82  * @param len The length of the input string
  83  * @return A pointer to the formatted string
  84  *
  85  */
  86 WS_DLL_PUBLIC
  87 gchar*     format_text_wsp(const guchar *line, size_t len);
  88
  89 /**
  90  * Given a string, generate a string from it that shows non-printable
  91  * characters as the chr parameter passed, except a whitespace character
  92  * (space, tab, carriage return, new line, vertical tab, or formfeed)
  93  * which will be replaced by a space, and return a pointer to it.
  94  *
  95  * @param line A pointer to the input string
  96  * @param len The length of the input string
  97  * @param chr The character to use to replace non-printable characters
  98  * @return A pointer to the formatted string
  99  *
 100  */
 101 WS_DLL_PUBLIC
 102 gchar*     format_text_chr(const guchar *string, const size_t len, const guchar chr);
 103
 104
 105 /** Turn a string of hex digits with optional separators (defined by
 106  *  is_byte_sep() into a byte array.
 107  *
 108  * @param hex_str The string of hex digits.
 109  * @param bytes The GByteArray that will receive the bytes.  This
 110  *        must be initialized by the caller.
 111  * @param force_separators If set to TRUE, separators MUST exist between
 112  *        bytes.
 113  * @return True if the string was converted successfully
 114  */
 115 WS_DLL_PUBLIC
 116 gboolean   hex_str_to_bytes(const char *hex_str, GByteArray *bytes,
 117     gboolean force_separators);
 118
 119 /** Turn an RFC 3986 percent-encoded string into a byte array.
 120  *
 121  * @param uri_str The string of hex digits.
 122  * @param bytes The GByteArray that will receive the bytes.  This
 123  *        must be initialized by the caller.
 124  * @return True if the string was converted successfully
 125  * @see format_uri()
 126  */
 127 WS_DLL_PUBLIC
 128 gboolean   uri_str_to_bytes(const char *uri_str, GByteArray *bytes);
 129
 130 /** Turn a byte array into an RFC 3986 percent-encoded string.
 131  *
 132  * @param bytes The GByteArray that will receive the bytes.  This
 133  *        must be initialized by the caller.
 134  * @param reserved_chars Normally the "gen-delims" and "sub-delims"
 135  *        from RFC 3986 (":/?#[]@" and "!$&'()*+,;=" respectively)
 136  *        plus space (hex value 20) are treated as reserved characters.
 137  *        If this variable is non-NULL, its contents will be used
 138  *        instead.
 139  * @note Any non-printing character determined by isprint(), along
 140  *       with the % character itself are always reserved.
 141  * @see uri_str_to_bytes(),  format_text(), isprint()
 142  */
 143 WS_DLL_PUBLIC
 144 const gchar* format_uri(const GByteArray *bytes, const gchar *reserved_chars);
 145
 146 /** Turn a OID string representation (dot notation) into a byte array.
 147  *
 148  * @param oid_str The OID string (dot notaion).
 149  * @param bytes The GByteArray that will receive the bytes.  This
 150  *        must be initialized by the caller.
 151  * @param is_absolute True if this is an absolute OID; false for relative OID.
 152  * @return True if the string was converted successfully
 153  */
 154 WS_DLL_PUBLIC
 155 gboolean   rel_oid_str_to_bytes(const char *oid_str, GByteArray *bytes, gboolean is_absolute);
 156
 157 /** Turn a OID string representation (dot notation) into a byte array.
 158  *
 159  * @param oid_str The OID string (dot notaion).
 160  * @param bytes The GByteArray that will receive the bytes.  This
 161  *        must be initialized by the caller.
 162  * @return True if the string was converted successfully
 163  */
 164 WS_DLL_PUBLIC
 165 gboolean   oid_str_to_bytes(const char *oid_str, GByteArray *bytes);
 166
 167 /**
 168  * Create a copy of a GByteArray
 169  *
 170  * @param ba The byte array to be copied.
 171  * @return If ba exists, a freshly allocated copy.  NULL otherwise.
 172  *
 173  * @todo - Should this be in strutil.c?
 174  */
 175 WS_DLL_PUBLIC
 176 GByteArray *byte_array_dup(GByteArray *ba);
 177
 178 /**
 179  * Compare the contents of two GByteArrays
 180  *
 181  * @param ba1 A byte array
 182  * @param ba2 A byte array
 183  * @return If both arrays are non-NULL and their lengths are equal and
 184  *         their contents are equal, returns TRUE.  Otherwise, returns
 185  *         FALSE.
 186  *
 187  * @todo - Should this be in strutil.c?
 188  */
 189 WS_DLL_PUBLIC
 190 gboolean byte_array_equal(GByteArray *ba1, GByteArray *ba2);
 191
 192
 193 /** Return a XML escaped representation of the unescaped string.
 194  *  The returned string must be freed when no longer in use.
 195  *
 196  * @param unescaped The unescaped string
 197  * @return An XML-escaped representation of the input string
 198  */
 199 WS_DLL_PUBLIC
 200 gchar*     xml_escape(const gchar *unescaped);
 201
 202 /**
 203  * Return the first occurrence of needle in haystack.
 204  * Algorithm copied from GNU's glibc 2.3.2 memcmp()
 205  *
 206  * @param haystack The data to search
 207  * @param haystack_len The length of the search data
 208  * @param needle The string to look for
 209  * @param needle_len The length of the search string
 210  * @return A pointer to the first occurrence of "needle" in
 211  *         "haystack".  If "needle" isn't found or is NULL, or if
 212  *         "needle_len" is 0, NULL is returned.
 213  */
 214 WS_DLL_PUBLIC
 215 const guint8 * epan_memmem(const guint8 *haystack, guint haystack_len,
 216                 const guint8 *needle, guint needle_len);
 217
 218 /** Scan a string to make sure it's valid hex.
 219  *
 220  * @param string The string to validate
 221  * @param nbytes The length of the return buffer
 222  * @return A pointer to a buffer containing the converted raw bytes.  This
 223  *         buffer must be g_free()d by the caller.
 224  */
 225 WS_DLL_PUBLIC
 226 guint8 * convert_string_to_hex(const char *string, size_t *nbytes);
 227
 228 /** Prep a string for case-sensitive vs case-insensitive searching.
 229  *
 230  * @param string The search string
 231  * @param case_insensitive TRUE if case-insensitive, FALSE if not
 232  * @return A direct copy of the string if it's a case-sensitive search and
 233  * an uppercased version if not.  In either case the string must be g_free()d
 234  * by the caller.
 235  */
 236 WS_DLL_PUBLIC
 237 char * convert_string_case(const char *string, gboolean case_insensitive);
 238
 239 /** Finds the first occurrence of string 'needle' in string 'haystack'.
 240  *  The matching is done in a case insensitive manner.
 241  *
 242  * @param haystack The string possibly containing the substring
 243  * @param needle The substring to be searched
 244  * @return A pointer into 'haystack' where 'needle' is first found.
 245  *   Otherwise it returns NULL.
 246  */
 247 WS_DLL_PUBLIC
 248 char * epan_strcasestr(const char *haystack, const char *needle);
 249
 250 /** Guarantee a non-null string.
 251  *
 252  * @param string The string to check
 253  * @return A pointer 'string' if it's non-null, otherwise "[NULL]".
 254  */
 255 WS_DLL_PUBLIC
 256 const char * string_or_null(const char *string);
 257
 258 WS_DLL_PUBLIC
 259 int escape_string_len(const char *string);
 260 WS_DLL_PUBLIC
 261 char * escape_string(char *dst, const char *string);
 262
 263
 264 WS_DLL_PUBLIC
 265 void IA5_7BIT_decode(unsigned char * dest, const unsigned char* src, int len);
 266
 267 /** Copy a string, escaping the 'chr' characters in it
 268  *
 269  * @param str The string to be copied
 270  * @param chr The character to be escaped
 271  * @return A copy of the string with every original 'chr' being
 272  * transformed into double 'chr'.
 273  */
 274 WS_DLL_PUBLIC
 275 gchar* ws_strdup_escape_char (const gchar *str, const gchar chr);
 276
 277 /** Copy a string, unescaping the 'chr' characters in it
 278  *
 279  * @param str The string to be copied
 280  * @param chr The character to be escaped
 281  * @return A copy of the string with every occurrence of double 'chr' in
 282  * the original string being copied as a single 'chr'.
 283  */
 284 WS_DLL_PUBLIC
 285 gchar* ws_strdup_unescape_char (const gchar *str, const gchar chr);
 286
 287 /** Replace values in a string
 288  *
 289  * @param str String containing 0 or more values to be replaced.
 290  * @param old_val Old value.
 291  * @param new_val New value. May be NULL, in which case occurences of
 292  *                           old_value will be removed.
 293  * @return A newly-allocated version of str with replacement values or
 294  * NULL on failure.
 295  */
 296 WS_DLL_PUBLIC
 297 gchar *string_replace(const gchar* str, const gchar *old_val, const gchar *new_val);
 298
 299 #ifdef __cplusplus
 300 }
 301 #endif /* __cplusplus */
 302
 303 #endif /* __STRUTIL_H__ */