llvm/lib/Support/Regex.cpp

   1 //===-- Regex.cpp - Regular Expression matcher implementation -------------===//
   2 //
   3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
   4 // See https://llvm.org/LICENSE.txt for license information.
   5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
   6 //
   7 //===----------------------------------------------------------------------===//
   8 //
   9 // This file implements a POSIX regular expression matcher.
  10 //
  11 //===----------------------------------------------------------------------===//
  12
  13 #include "llvm/Support/Regex.h"
  14 #include "llvm/ADT/SmallVector.h"
  15 #include "llvm/ADT/StringRef.h"
  16 #include "llvm/ADT/Twine.h"
  17 #include <cassert>
  18 #include <string>
  19
  20 // Important this comes last because it defines "_REGEX_H_". At least on
  21 // Darwin, if included before any header that (transitively) includes
  22 // xlocale.h, this will cause trouble, because of missing regex-related types.
  23 #include "regex_impl.h"
  24
  25 using namespace llvm;
  26
  27 Regex::Regex() : preg(nullptr), error(REG_BADPAT) {}
  28
  29 Regex::Regex(StringRef regex, RegexFlags Flags) {
  30   unsigned flags = 0;
  31   preg = new llvm_regex();
  32   preg->re_endp = regex.end();
  33   if (Flags & IgnoreCase)
  34     flags |= REG_ICASE;
  35   if (Flags & Newline)
  36     flags |= REG_NEWLINE;
  37   if (!(Flags & BasicRegex))
  38     flags |= REG_EXTENDED;
  39   error = llvm_regcomp(preg, regex.data(), flags|REG_PEND);
  40 }
  41
  42 Regex::Regex(StringRef regex, unsigned Flags)
  43     : Regex(regex, static_cast<RegexFlags>(Flags)) {}
  44
  45 Regex::Regex(Regex &&regex) {
  46   preg = regex.preg;
  47   error = regex.error;
  48   regex.preg = nullptr;
  49   regex.error = REG_BADPAT;
  50 }
  51
  52 Regex::~Regex() {
  53   if (preg) {
  54     llvm_regfree(preg);
  55     delete preg;
  56   }
  57 }
  58
  59 namespace {
  60
  61 /// Utility to convert a regex error code into a human-readable string.
  62 void RegexErrorToString(int error, struct llvm_regex *preg,
  63                         std::string &Error) {
  64   size_t len = llvm_regerror(error, preg, nullptr, 0);
  65
  66   Error.resize(len - 1);
  67   llvm_regerror(error, preg, &Error[0], len);
  68 }
  69
  70 } // namespace
  71
  72 bool Regex::isValid(std::string &Error) const {
  73   if (!error)
  74     return true;
  75
  76   RegexErrorToString(error, preg, Error);
  77   return false;
  78 }
  79
  80 /// getNumMatches - In a valid regex, return the number of parenthesized
  81 /// matches it contains.
  82 unsigned Regex::getNumMatches() const {
  83   return preg->re_nsub;
  84 }
  85
  86 bool Regex::match(StringRef String, SmallVectorImpl<StringRef> *Matches,
  87                   std::string *Error) const {
  88   // Reset error, if given.
  89   if (Error && !Error->empty())
  90     *Error = "";
  91
  92   // Check if the regex itself didn't successfully compile.
  93   if (Error ? !isValid(*Error) : !isValid())
  94     return false;
  95
  96   unsigned nmatch = Matches ? preg->re_nsub+1 : 0;
  97
  98   // pmatch needs to have at least one element.
  99   SmallVector<llvm_regmatch_t, 8> pm;
 100   pm.resize(nmatch > 0 ? nmatch : 1);
 101   pm[0].rm_so = 0;
 102   pm[0].rm_eo = String.size();
 103
 104   int rc = llvm_regexec(preg, String.data(), nmatch, pm.data(), REG_STARTEND);
 105
 106   // Failure to match is not an error, it's just a normal return value.
 107   // Any other error code is considered abnormal, and is logged in the Error.
 108   if (rc == REG_NOMATCH)
 109     return false;
 110   if (rc != 0) {
 111     if (Error)
 112       RegexErrorToString(error, preg, *Error);
 113     return false;
 114   }
 115
 116   // There was a match.
 117
 118   if (Matches) { // match position requested
 119     Matches->clear();
 120
 121     for (unsigned i = 0; i != nmatch; ++i) {
 122       if (pm[i].rm_so == -1) {
 123         // this group didn't match
 124         Matches->push_back(StringRef());
 125         continue;
 126       }
 127       assert(pm[i].rm_eo >= pm[i].rm_so);
 128       Matches->push_back(StringRef(String.data()+pm[i].rm_so,
 129                                    pm[i].rm_eo-pm[i].rm_so));
 130     }
 131   }
 132
 133   return true;
 134 }
 135
 136 std::string Regex::sub(StringRef Repl, StringRef String,
 137                        std::string *Error) const {
 138   SmallVector<StringRef, 8> Matches;
 139
 140   // Return the input if there was no match.
 141   if (!match(String, &Matches, Error))
 142     return std::string(String);
 143
 144   // Otherwise splice in the replacement string, starting with the prefix before
 145   // the match.
 146   std::string Res(String.begin(), Matches[0].begin());
 147
 148   // Then the replacement string, honoring possible substitutions.
 149   while (!Repl.empty()) {
 150     // Skip to the next escape.
 151     std::pair<StringRef, StringRef> Split = Repl.split('\\');
 152
 153     // Add the skipped substring.
 154     Res += Split.first;
 155
 156     // Check for terminimation and trailing backslash.
 157     if (Split.second.empty()) {
 158       if (Repl.size() != Split.first.size() &&
 159           Error && Error->empty())
 160         *Error = "replacement string contained trailing backslash";
 161       break;
 162     }
 163
 164     // Otherwise update the replacement string and interpret escapes.
 165     Repl = Split.second;
 166
 167     // FIXME: We should have a StringExtras function for mapping C99 escapes.
 168     switch (Repl[0]) {
 169       // Treat all unrecognized characters as self-quoting.
 170     default:
 171       Res += Repl[0];
 172       Repl = Repl.substr(1);
 173       break;
 174
 175       // Single character escapes.
 176     case 't':
 177       Res += '\t';
 178       Repl = Repl.substr(1);
 179       break;
 180     case 'n':
 181       Res += '\n';
 182       Repl = Repl.substr(1);
 183       break;
 184
 185       // Decimal escapes are backreferences.
 186     case '0': case '1': case '2': case '3': case '4':
 187     case '5': case '6': case '7': case '8': case '9': {
 188       // Extract the backreference number.
 189       StringRef Ref = Repl.slice(0, Repl.find_first_not_of("0123456789"));
 190       Repl = Repl.substr(Ref.size());
 191
 192       unsigned RefValue;
 193       if (!Ref.getAsInteger(10, RefValue) &&
 194           RefValue < Matches.size())
 195         Res += Matches[RefValue];
 196       else if (Error && Error->empty())
 197         *Error = ("invalid backreference string '" + Twine(Ref) + "'").str();
 198       break;
 199     }
 200     }
 201   }
 202
 203   // And finally the suffix.
 204   Res += StringRef(Matches[0].end(), String.end() - Matches[0].end());
 205
 206   return Res;
 207 }
 208
 209 // These are the special characters matched in functions like "p_ere_exp".
 210 static const char RegexMetachars[] = "()^$|*+?.[]\\{}";
 211
 212 bool Regex::isLiteralERE(StringRef Str) {
 213   // Check for regex metacharacters.  This list was derived from our regex
 214   // implementation in regcomp.c and double checked against the POSIX extended
 215   // regular expression specification.
 216   return Str.find_first_of(RegexMetachars) == StringRef::npos;
 217 }
 218
 219 std::string Regex::escape(StringRef String) {
 220   std::string RegexStr;
 221   for (unsigned i = 0, e = String.size(); i != e; ++i) {
 222     if (strchr(RegexMetachars, String[i]))
 223       RegexStr += '\\';
 224     RegexStr += String[i];
 225   }
 226
 227   return RegexStr;
 228 }