source/ps/CStr.h

   1 /* Copyright (C) 2022 Wildfire Games.
   2  * This file is part of 0 A.D.
   3  *
   4  * 0 A.D. is free software: you can redistribute it and/or modify
   5  * it under the terms of the GNU General Public License as published by
   6  * the Free Software Foundation, either version 2 of the License, or
   7  * (at your option) any later version.
   8  *
   9  * 0 A.D. is distributed in the hope that it will be useful,
  10  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  11  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  12  * GNU General Public License for more details.
  13  *
  14  * You should have received a copy of the GNU General Public License
  15  * along with 0 A.D.  If not, see <http://www.gnu.org/licenses/>.
  16  */
  17
  18 /**
  19  * Description : Contains CStr class which is a versatile class for making string use easy.
  20  *                         : The class implements a series of string manipulation/formatting functions.
  21  **/
  22
  23 #ifndef INCLUDED_CSTR
  24 #define INCLUDED_CSTR
  25
  26 /**
  27  * Whitespace trim identifier for Trim and Pad functions
  28  **/
  29 enum PS_TRIM_MODE
  30 {
  31         PS_TRIM_LEFT,   /// Trim all white space from the beginning of the string
  32         PS_TRIM_RIGHT,  /// Trim all white space from the end of the string
  33         PS_TRIM_BOTH    /// Trim all white space from the beginning and end of the string
  34 };
  35
  36 #ifndef IN_UNIDOUBLER
  37  #define UNIDOUBLER_HEADER "CStr.h"
  38  #include "UniDoubler.h"
  39 #endif
  40
  41 #endif
  42
  43 // Include this section when in unidoubler mode, and when this unicode/ascii
  44 // version has not already been included.
  45 #if defined(IN_UNIDOUBLER) && ( (defined(_UNICODE) && !defined(CSTR_H_U)) || (!defined(_UNICODE) && !defined(CSTR_H_A)) )
  46
  47 #ifdef _UNICODE
  48 #define CSTR_H_U
  49 #else
  50 #define CSTR_H_A
  51 #endif
  52
  53 #include "ps/CStrForward.h"
  54
  55 #include <string>
  56
  57 /**
  58  * The base class of all strings
  59  **/
  60 class CStr : public std::tstring
  61 {
  62 public:
  63         using StrBase = std::tstring;
  64         using Char = typename std::tstring::value_type;
  65
  66         CStr(const StrBase& str) : StrBase(str) {}
  67
  68         using StrBase::StrBase;
  69
  70         /**
  71          * Repeat: Named constructor, to avoid overload overload.
  72          *
  73          * @param const CStr & str reference to another CStr object to be repeated for initialization
  74          * @param size_t Reps number of times to repeat the initialization
  75          * @return CStr new CStr object
  76          **/
  77         static CStr Repeat(const CStr& str, size_t reps);
  78
  79         /**
  80          * Construction from u16strings.
  81          *
  82          * @param u16string String u16string to be used for initialization.
  83          **/
  84         explicit CStr(const std::u16string& str) : StrBase(str.begin(), str.end()) {}
  85
  86         // Conversion to/from UTF-8, encoded in a CStr8.
  87         // Invalid bytes/characters (e.g. broken UTF-8, and Unicode characters
  88         // above U+FFFF) are silently replaced with U+FFFD.
  89         #ifdef _UNICODE
  90                 CStr8 ToUTF8() const;
  91         #else
  92                 CStrW FromUTF8() const;
  93         #endif
  94
  95         // Conversions:
  96
  97         static CStr FromInt(int n);
  98         static CStr FromUInt(unsigned int n);
  99         static CStr FromInt64(i64 n);
 100         static CStr FromDouble(double n);
 101
 102         /**
 103          * Return CStr as Integer.
 104          * Conversion is from the beginning of CStr.
 105          *
 106          * @return int CStr represented as an integer.
 107          **/
 108         int ToInt() const;
 109         /**
 110          * Return CStr as Unsigned Integer.
 111          * Conversion is from the beginning of CStr.
 112          *
 113          * @return unsigned int CStr represented as an unsigned integer.
 114          **/
 115         unsigned int ToUInt() const;
 116         /**
 117          * Return CStr as Long.
 118          * Conversion is from the beginning of CStr.
 119          *
 120          * @return long CStr represented as a long.
 121          **/
 122         long ToLong() const;
 123         /**
 124          * Return CStr as Unsigned Long.
 125          * Conversion is from the beginning of CStr.
 126          *
 127          * @return unsigned long CStr represented as an unsigned long.
 128          **/
 129         unsigned long ToULong() const;
 130         /**
 131          * Return CStr as Float.
 132          * Conversion is from the beginning of CStr.
 133          *
 134          * @return float CStr represented as a float.
 135          **/
 136         float ToFloat() const;
 137         /**
 138          * Return CStr as Double.
 139          * Conversion is from the beginning of CStr.
 140          *
 141          * @return double CStr represented as a double.
 142          **/
 143         double ToDouble() const;
 144
 145         /**
 146          * Search the CStr for another string.
 147          * The search is case-sensitive.
 148          *
 149          * @param const CStr & str reference to the search string
 150          * @return long offset into the CStr of the first occurrence of the search string
 151          *                              -1 if the search string is not found
 152          **/
 153         long Find(const CStr& str) const;
 154         /**
 155          * Search the CStr for another string.
 156          * The search is case-sensitive.
 157          *
 158          * @param const {t|w}char_t & chr reference to the search string
 159          * @return long offset into the CStr of the first occurrence of the search string
 160          *                              -1 if the search string is not found
 161          **/
 162         long Find(const Char chr) const;
 163         /**
 164          * Search the CStr for another string with starting offset.
 165          * The search is case-sensitive.
 166          *
 167          * @param const int & start character offset into CStr to begin search
 168          * @param const {t|w}char_t & chr reference to the search string
 169          * @return long offset into the CStr of the first occurrence of the search string
 170          *                              -1 if the search string is not found
 171          **/
 172         long Find(const int start, const Char chr) const;
 173
 174         /**
 175          * Search the CStr for another string.
 176          * The search is case-insensitive.
 177          *
 178          * @param const CStr & str reference to the search string
 179          * @return long offset into the CStr of the first occurrence of the search string
 180          *                              -1 if the search string is not found
 181          **/
 182         long FindInsensitive(const CStr& str) const;
 183         /**
 184          * Search the CStr for another string.
 185          * The search is case-insensitive.
 186          *
 187          * @param const {t|w}char_t & chr reference to the search string
 188          * @return long offset into the CStr of the first occurrence of the search string
 189          *                              -1 if the search string is not found
 190          **/
 191         long FindInsensitive(const Char chr) const;
 192         /**
 193          * Search the CStr for another string with starting offset.
 194          * The search is case-insensitive.
 195          *
 196          * @param const int & start character offset into CStr to begin search
 197          * @param const {t|w}char_t & chr reference to the search string
 198          * @return long offset into the CStr of the first occurrence of the search string
 199          *                              -1 if the search string is not found
 200          **/
 201         long FindInsensitive(const int start, const Char chr) const;
 202
 203         /**
 204          * Search the CStr for another string.
 205          * The search is case-sensitive.
 206          *
 207          * @param const CStr & str reference to the search string
 208          * @return long offset into the CStr of the last occurrence of the search string
 209          *                              -1 if the search string is not found
 210          **/
 211         long ReverseFind(const CStr& str) const;
 212
 213         /**
 214          * Make a copy of the CStr in lower-case.
 215          *
 216          * @return CStr converted copy of CStr.
 217          **/
 218         CStr LowerCase() const;
 219         /**
 220          * Make a copy of the CStr in upper-case.
 221          *
 222          * @return CStr converted copy of CStr.
 223          **/
 224         CStr UpperCase() const;
 225
 226         /**
 227          * Retrieve first n characters of the CStr.
 228          *
 229          * @param size_t len the number of characters to retrieve.
 230          * @return CStr retrieved substring.
 231          **/
 232         CStr Left(size_t len) const;
 233
 234         /**
 235          * Retrieve last n characters of the CStr.
 236          *
 237          * @param size_t len the number of characters to retrieve.
 238          * @return CStr retrieved substring.
 239          **/
 240         CStr Right(size_t len) const;
 241
 242         /**
 243          * Retrieve substring of the CStr after last occurrence of a string.
 244          * Return substring of the CStr after the last occurrence of the search string.
 245          *
 246          * @param const CStr & str reference to search string
 247          * @param size_t startPos character position to start searching from
 248          * @return CStr substring remaining after match
 249          *                                       the CStr if no match is found
 250          **/
 251         CStr AfterLast(const CStr& str, size_t startPos = npos) const;
 252
 253         /**
 254          * Retrieve substring of the CStr preceding last occurrence of a string.
 255          * Return substring of the CStr preceding the last occurrence of the search string.
 256          *
 257          * @param const CStr & str reference to search string
 258          * @param size_t startPos character position to start searching from
 259          * @return CStr substring preceding before match
 260          *                                       the CStr if no match is found
 261          **/
 262         CStr BeforeLast(const CStr& str, size_t startPos = npos) const;
 263
 264         /**
 265          * Retrieve substring of the CStr after first occurrence of a string.
 266          * Return substring of the CStr after the first occurrence of the search string.
 267          *
 268          * @param const CStr & str reference to search string
 269          * @param size_t startPos character position to start searching from
 270          * @return CStr substring remaining after match
 271          *                                       the CStr if no match is found
 272          **/
 273         CStr AfterFirst(const CStr& str, size_t startPos = 0) const;
 274
 275         /**
 276          * Retrieve substring of the CStr preceding first occurrence of a string.
 277          * Return substring of the CStr preceding the first occurrence of the search string.
 278          *
 279          * @param const CStr & str reference to search string
 280          * @param size_t startPos character position to start searching from
 281          * @return CStr substring preceding before match
 282          *                                       the CStr if no match is found
 283          **/
 284         CStr BeforeFirst(const CStr& str, size_t startPos = 0) const;
 285
 286         /**
 287          * Remove all occurrences of a string from the CStr.
 288          *
 289          * @param const CStr & str reference to search string to remove.
 290          **/
 291         void Remove(const CStr& str);
 292
 293         /**
 294          * Replace all occurrences of one string by another string in the CStr.
 295          *
 296          * @param const CStr & strToReplace reference to search string.
 297          * @param const CStr & replaceWith reference to replace string.
 298          **/
 299         void Replace(const CStr& toReplace, const CStr& replaceWith);
 300
 301         /**
 302          * Convert strings to printable ASCII characters with JSON-style escapes.
 303          */
 304         std::string EscapeToPrintableASCII() const;
 305
 306         /**
 307          * Return a trimmed copy of the CStr.
 308          *
 309          * @param PS_TRIM_MODE Mode value from trim mode enumeration.
 310          * @return CStr copy of trimmed CStr.
 311          **/
 312         CStr Trim(PS_TRIM_MODE mode) const;
 313
 314         /**
 315          * Return a space padded copy of the CStr.
 316          *
 317          * @param PS_TRIM_MODE Mode value from trim mode enumeration.
 318          * @param size_t Length number of pad spaces to add
 319          * @return CStr copy of padded CStr.
 320          **/
 321         CStr Pad(PS_TRIM_MODE mode, size_t len) const;
 322
 323         // Conversion to u16string
 324         std::u16string utf16() const { return std::u16string(begin(), end()); }
 325
 326         // Calculates a hash of the string's contents
 327         size_t GetHashCode() const;
 328
 329         // Serialization functions
 330         // (These are not virtual or inherited from ISerializable, to avoid
 331         // adding a vtable and making the strings larger than std::string)
 332         size_t GetSerializedLength() const;
 333         u8* Serialize(u8* buffer) const;
 334         const u8* Deserialize(const u8* buffer, const u8* bufferend);
 335 };
 336
 337 namespace std
 338 {
 339 template <>
 340 struct hash<CStr>
 341 {
 342         std::size_t operator()(const CStr& str) const
 343         {
 344                 return str.GetHashCode();
 345         }
 346 };
 347 }
 348
 349 #endif // INCLUDED_CSTR