1 //===-- CPPLanguageRuntime.cpp -------------------------------------------------*- C++ -*-===// 2 // 3 // The LLVM Compiler Infrastructure 4 // 5 // This file is distributed under the University of Illinois Open Source 6 // License. See LICENSE.TXT for details. 7 // 8 //===----------------------------------------------------------------------===// 9 10 #include "lldb/Target/CPPLanguageRuntime.h" 11 12 #include <string.h> 13 14 #include "llvm/ADT/StringRef.h" 15 16 #include "lldb/Core/PluginManager.h" 17 #include "lldb/Core/UniqueCStringMap.h" 18 #include "lldb/Target/ExecutionContext.h" 19 20 using namespace lldb; 21 using namespace lldb_private; 22 23 class CPPRuntimeEquivalents 24 { 25 public: 26 CPPRuntimeEquivalents () 27 { 28 29 m_impl.Append(ConstString("std::basic_string<char, std::char_traits<char>, std::allocator<char> >").AsCString(), ConstString("basic_string<char>")); 30 31 // these two (with a prefixed std::) occur when c++stdlib string class occurs as a template argument in some STL container 32 m_impl.Append(ConstString("std::basic_string<char, std::char_traits<char>, std::allocator<char> >").AsCString(), ConstString("std::basic_string<char>")); 33 34 m_impl.Sort(); 35 } 36 37 void 38 Add (ConstString& type_name, 39 ConstString& type_equivalent) 40 { 41 m_impl.Insert(type_name.AsCString(), type_equivalent); 42 } 43 44 uint32_t 45 FindExactMatches (ConstString& type_name, 46 std::vector<ConstString>& equivalents) 47 { 48 49 uint32_t count = 0; 50 51 for (ImplData match = m_impl.FindFirstValueForName(type_name.AsCString()); 52 match != NULL; 53 match = m_impl.FindNextValueForName(match)) 54 { 55 equivalents.push_back(match->value); 56 count++; 57 } 58 59 return count; 60 } 61 62 // partial matches can occur when a name with equivalents is a template argument. 63 // e.g. we may have "class Foo" be a match for "struct Bar". if we have a typename 64 // such as "class Templatized<class Foo, Anything>" we want this to be replaced with 65 // "class Templatized<struct Bar, Anything>". Since partial matching is time consuming 66 // once we get a partial match, we add it to the exact matches list for faster retrieval 67 uint32_t 68 FindPartialMatches (ConstString& type_name, 69 std::vector<ConstString>& equivalents) 70 { 71 72 uint32_t count = 0; 73 74 const char* type_name_cstr = type_name.AsCString(); 75 76 size_t items_count = m_impl.GetSize(); 77 78 for (size_t item = 0; item < items_count; item++) 79 { 80 const char* key_cstr = m_impl.GetCStringAtIndex(item); 81 if ( strstr(type_name_cstr,key_cstr) ) 82 { 83 count += AppendReplacements(type_name_cstr, 84 key_cstr, 85 equivalents); 86 } 87 } 88 89 return count; 90 91 } 92 93 private: 94 95 std::string& replace (std::string& target, 96 std::string& pattern, 97 std::string& with) 98 { 99 size_t pos; 100 size_t pattern_len = pattern.size(); 101 102 while ( (pos = target.find(pattern)) != std::string::npos ) 103 target.replace(pos, pattern_len, with); 104 105 return target; 106 } 107 108 uint32_t 109 AppendReplacements (const char* original, 110 const char *matching_key, 111 std::vector<ConstString>& equivalents) 112 { 113 114 std::string matching_key_str(matching_key); 115 ConstString original_const(original); 116 117 uint32_t count = 0; 118 119 for (ImplData match = m_impl.FindFirstValueForName(matching_key); 120 match != NULL; 121 match = m_impl.FindNextValueForName(match)) 122 { 123 std::string target(original); 124 std::string equiv_class(match->value.AsCString()); 125 126 replace (target, matching_key_str, equiv_class); 127 128 ConstString target_const(target.c_str()); 129 130 // you will most probably want to leave this off since it might make this map grow indefinitely 131 #ifdef ENABLE_CPP_EQUIVALENTS_MAP_TO_GROW 132 Add(original_const, target_const); 133 #endif 134 equivalents.push_back(target_const); 135 136 count++; 137 } 138 139 return count; 140 } 141 142 typedef UniqueCStringMap<ConstString> Impl; 143 typedef const Impl::Entry* ImplData; 144 Impl m_impl; 145 }; 146 147 static CPPRuntimeEquivalents& 148 GetEquivalentsMap () 149 { 150 static CPPRuntimeEquivalents g_equivalents_map; 151 return g_equivalents_map; 152 } 153 154 //---------------------------------------------------------------------- 155 // Destructor 156 //---------------------------------------------------------------------- 157 CPPLanguageRuntime::~CPPLanguageRuntime() 158 { 159 } 160 161 CPPLanguageRuntime::CPPLanguageRuntime (Process *process) : 162 LanguageRuntime (process) 163 { 164 165 } 166 167 bool 168 CPPLanguageRuntime::GetObjectDescription (Stream &str, ValueObject &object) 169 { 170 // C++ has no generic way to do this. 171 return false; 172 } 173 174 bool 175 CPPLanguageRuntime::GetObjectDescription (Stream &str, Value &value, ExecutionContextScope *exe_scope) 176 { 177 // C++ has no generic way to do this. 178 return false; 179 } 180 181 bool 182 CPPLanguageRuntime::IsCPPMangledName (const char *name) 183 { 184 // FIXME, we should really run through all the known C++ Language plugins and ask each one if 185 // this is a C++ mangled name, but we can put that off till there is actually more than one 186 // we care about. 187 188 if (name && name[0] == '_' && name[1] == 'Z') 189 return true; 190 else 191 return false; 192 } 193 194 bool 195 CPPLanguageRuntime::ExtractContextAndIdentifier (const char *name, llvm::StringRef &context, llvm::StringRef &identifier) 196 { 197 static RegularExpression g_basename_regex("^(([A-Za-z_][A-Za-z_0-9]*::)*)([A-Za-z_][A-Za-z_0-9]*)$"); 198 RegularExpression::Match match(4); 199 if (g_basename_regex.Execute (name, &match)) 200 { 201 match.GetMatchAtIndex(name, 1, context); 202 match.GetMatchAtIndex(name, 3, identifier); 203 return true; 204 } 205 return false; 206 } 207 208 uint32_t 209 CPPLanguageRuntime::FindEquivalentNames(ConstString type_name, std::vector<ConstString>& equivalents) 210 { 211 uint32_t count = GetEquivalentsMap().FindExactMatches(type_name, equivalents); 212 213 bool might_have_partials= 214 ( count == 0 ) // if we have a full name match just use it 215 && (strchr(type_name.AsCString(), '<') != NULL // we should only have partial matches when templates are involved, check that we have 216 && strchr(type_name.AsCString(), '>') != NULL); // angle brackets in the type_name before trying to scan for partial matches 217 218 if ( might_have_partials ) 219 count = GetEquivalentsMap().FindPartialMatches(type_name, equivalents); 220 221 return count; 222 } 223 224 void 225 CPPLanguageRuntime::MethodName::Clear() 226 { 227 m_full.Clear(); 228 m_basename = llvm::StringRef(); 229 m_context = llvm::StringRef(); 230 m_arguments = llvm::StringRef(); 231 m_qualifiers = llvm::StringRef(); 232 m_type = eTypeInvalid; 233 m_parsed = false; 234 m_parse_error = false; 235 } 236 237 bool 238 ReverseFindMatchingChars (const llvm::StringRef &s, 239 const llvm::StringRef &left_right_chars, 240 size_t &left_pos, 241 size_t &right_pos, 242 size_t pos = llvm::StringRef::npos) 243 { 244 assert (left_right_chars.size() == 2); 245 left_pos = llvm::StringRef::npos; 246 const char left_char = left_right_chars[0]; 247 const char right_char = left_right_chars[1]; 248 pos = s.find_last_of(left_right_chars, pos); 249 if (pos == llvm::StringRef::npos || s[pos] == left_char) 250 return false; 251 right_pos = pos; 252 uint32_t depth = 1; 253 while (pos > 0 && depth > 0) 254 { 255 pos = s.find_last_of(left_right_chars, pos); 256 if (pos == llvm::StringRef::npos) 257 return false; 258 if (s[pos] == left_char) 259 { 260 if (--depth == 0) 261 { 262 left_pos = pos; 263 return left_pos < right_pos; 264 } 265 } 266 else if (s[pos] == right_char) 267 { 268 ++depth; 269 } 270 } 271 return false; 272 } 273 274 275 void 276 CPPLanguageRuntime::MethodName::Parse() 277 { 278 if (!m_parsed && m_full) 279 { 280 // ConstString mangled; 281 // m_full.GetMangledCounterpart(mangled); 282 // printf ("\n parsing = '%s'\n", m_full.GetCString()); 283 // if (mangled) 284 // printf (" mangled = '%s'\n", mangled.GetCString()); 285 m_parse_error = false; 286 m_parsed = true; 287 llvm::StringRef full (m_full.GetCString()); 288 289 size_t arg_start, arg_end; 290 llvm::StringRef parens("()", 2); 291 if (ReverseFindMatchingChars (full, parens, arg_start, arg_end)) 292 { 293 m_arguments = full.substr(arg_start, arg_end - arg_start + 1); 294 if (arg_end + 1 < full.size()) 295 m_qualifiers = full.substr(arg_end + 1); 296 if (arg_start > 0) 297 { 298 size_t basename_end = arg_start; 299 size_t context_start = 0; 300 size_t context_end = llvm::StringRef::npos; 301 if (basename_end > 0 && full[basename_end-1] == '>') 302 { 303 // TODO: handle template junk... 304 // Templated function 305 size_t template_start, template_end; 306 llvm::StringRef lt_gt("<>", 2); 307 if (ReverseFindMatchingChars (full, lt_gt, template_start, template_end, basename_end)) 308 { 309 context_end = full.rfind(':', template_start); 310 if (context_end == llvm::StringRef::npos) 311 { 312 // Check for templated functions that include return type like: 313 // 'void foo<Int>()' 314 context_end = full.rfind(' ', template_start); 315 if (context_end != llvm::StringRef::npos) 316 { 317 context_start = context_end; 318 } 319 } 320 } 321 else 322 { 323 context_end = full.rfind(':', basename_end); 324 } 325 } 326 else if (context_end == llvm::StringRef::npos) 327 { 328 context_end = full.rfind(':', basename_end); 329 } 330 331 if (context_end == llvm::StringRef::npos) 332 m_basename = full.substr(0, basename_end); 333 else 334 { 335 if (context_start < context_end) 336 m_context = full.substr(context_start, context_end - 1); 337 const size_t basename_begin = context_end + 1; 338 m_basename = full.substr(basename_begin, basename_end - basename_begin); 339 } 340 m_type = eTypeUnknownMethod; 341 } 342 else 343 { 344 m_parse_error = true; 345 return; 346 } 347 348 // if (!m_context.empty()) 349 // printf (" context = '%s'\n", m_context.str().c_str()); 350 // if (m_basename) 351 // printf (" basename = '%s'\n", m_basename.GetCString()); 352 // if (!m_arguments.empty()) 353 // printf (" arguments = '%s'\n", m_arguments.str().c_str()); 354 // if (!m_qualifiers.empty()) 355 // printf ("qualifiers = '%s'\n", m_qualifiers.str().c_str()); 356 357 // Make sure we have a valid C++ basename with optional template args 358 static RegularExpression g_identifier_regex("^~?([A-Za-z_][A-Za-z_0-9]*)(<.*>)?$"); 359 std::string basename_str(m_basename.str()); 360 bool basename_is_valid = g_identifier_regex.Execute (basename_str.c_str(), NULL); 361 if (!basename_is_valid) 362 { 363 // Check for C++ operators 364 if (m_basename.startswith("operator")) 365 { 366 static RegularExpression g_operator_regex("^(operator)( ?)([A-Za-z_][A-Za-z_0-9]*|\\(\\)|\\[\\]|[\\^<>=!\\/*+-]+)(<.*>)?(\\[\\])?$"); 367 basename_is_valid = g_operator_regex.Execute(basename_str.c_str(), NULL); 368 } 369 } 370 if (!basename_is_valid) 371 { 372 // The C++ basename doesn't match our regular expressions so this can't 373 // be a valid C++ method, clear everything out and indicate an error 374 m_context = llvm::StringRef(); 375 m_basename = llvm::StringRef(); 376 m_arguments = llvm::StringRef(); 377 m_qualifiers = llvm::StringRef(); 378 m_parse_error = true; 379 } 380 } 381 else 382 { 383 m_parse_error = true; 384 // printf ("error: didn't find matching parens for arguments\n"); 385 } 386 } 387 } 388 389 llvm::StringRef 390 CPPLanguageRuntime::MethodName::GetBasename () 391 { 392 if (!m_parsed) 393 Parse(); 394 return m_basename; 395 } 396 397 llvm::StringRef 398 CPPLanguageRuntime::MethodName::GetContext () 399 { 400 if (!m_parsed) 401 Parse(); 402 return m_context; 403 } 404 405 llvm::StringRef 406 CPPLanguageRuntime::MethodName::GetArguments () 407 { 408 if (!m_parsed) 409 Parse(); 410 return m_arguments; 411 } 412 413 llvm::StringRef 414 CPPLanguageRuntime::MethodName::GetQualifiers () 415 { 416 if (!m_parsed) 417 Parse(); 418 return m_qualifiers; 419 } 420 421