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 "regex_impl.h"
  18
  19 #include <cassert>
  20 #include <string>
  21
  22 using namespace llvm;
  23
  24 Regex::Regex() : preg(nullptr), error(REG_BADPAT) {}
  25
  26 Regex::Regex(StringRef regex, RegexFlags Flags) {
  27   unsigned flags = 0;
  28   preg = new llvm_regex();
  29   preg->re_endp = regex.end();
  30   if (Flags & IgnoreCase)
  31     flags |= REG_ICASE;
  32   if (Flags & Newline)
  33     flags |= REG_NEWLINE;
  34   if (!(Flags & BasicRegex))
  35     flags |= REG_EXTENDED;
  36   error = llvm_regcomp(preg, regex.data(), flags|REG_PEND);
  37 }
  38
  39 Regex::Regex(StringRef regex, unsigned Flags)
  40     : Regex(regex, static_cast<RegexFlags>(Flags)) {}
  41
  42 Regex::Regex(Regex &&regex) {
  43   preg = regex.preg;
  44   error = regex.error;
  45   regex.preg = nullptr;
  46   regex.error = REG_BADPAT;
  47 }
  48
  49 Regex::~Regex() {
  50   if (preg) {
  51     llvm_regfree(preg);
  52     delete preg;
  53   }
  54 }
  55
  56 namespace {
  57
  58 /// Utility to convert a regex error code into a human-readable string.
  59 void RegexErrorToString(int error, struct llvm_regex *preg,
  60                         std::string &Error) {
  61   size_t len = llvm_regerror(error, preg, nullptr, 0);
  62
  63   Error.resize(len - 1);
  64   llvm_regerror(error, preg, &Error[0], len);
  65 }
  66
  67 } // namespace
  68
  69 bool Regex::isValid(std::string &Error) const {
  70   if (!error)
  71     return true;
  72
  73   RegexErrorToString(error, preg, Error);
  74   return false;
  75 }
  76
  77 /// getNumMatches - In a valid regex, return the number of parenthesized
  78 /// matches it contains.
  79 unsigned Regex::getNumMatches() const {
  80   return preg->re_nsub;
  81 }
  82
  83 bool Regex::match(StringRef String, SmallVectorImpl<StringRef> *Matches,
  84                   std::string *Error) const {
  85   // Reset error, if given.
  86   if (Error && !Error->empty())
  87     *Error = "";
  88
  89   // Check if the regex itself didn't successfully compile.
  90   if (Error ? !isValid(*Error) : !isValid())
  91     return false;
  92
  93   unsigned nmatch = Matches ? preg->re_nsub+1 : 0;
  94
  95   // pmatch needs to have at least one element.
  96   SmallVector<llvm_regmatch_t, 8> pm;
  97   pm.resize(nmatch > 0 ? nmatch : 1);
  98   pm[0].rm_so = 0;
  99   pm[0].rm_eo = String.size();
 100
 101   int rc = llvm_regexec(preg, String.data(), nmatch, pm.data(), REG_STARTEND);
 102
 103   // Failure to match is not an error, it's just a normal return value.
 104   // Any other error code is considered abnormal, and is logged in the Error.
 105   if (rc == REG_NOMATCH)
 106     return false;
 107   if (rc != 0) {
 108     if (Error)
 109       RegexErrorToString(error, preg, *Error);
 110     return false;
 111   }
 112
 113   // There was a match.
 114
 115   if (Matches) { // match position requested
 116     Matches->clear();
 117
 118     for (unsigned i = 0; i != nmatch; ++i) {
 119       if (pm[i].rm_so == -1) {
 120         // this group didn't match
 121         Matches->push_back(StringRef());
 122         continue;
 123       }
 124       assert(pm[i].rm_eo >= pm[i].rm_so);
 125       Matches->push_back(StringRef(String.data()+pm[i].rm_so,
 126                                    pm[i].rm_eo-pm[i].rm_so));
 127     }
 128   }
 129
 130   return true;
 131 }
 132
 133 std::string Regex::sub(StringRef Repl, StringRef String,
 134                        std::string *Error) const {
 135   SmallVector<StringRef, 8> Matches;
 136
 137   // Return the input if there was no match.
 138   if (!match(String, &Matches, Error))
 139     return std::string(String);
 140
 141   // Otherwise splice in the replacement string, starting with the prefix before
 142   // the match.
 143   std::string Res(String.begin(), Matches[0].begin());
 144
 145   // Then the replacement string, honoring possible substitutions.
 146   while (!Repl.empty()) {
 147     // Skip to the next escape.
 148     std::pair<StringRef, StringRef> Split = Repl.split('\\');
 149
 150     // Add the skipped substring.
 151     Res += Split.first;
 152
 153     // Check for terminimation and trailing backslash.
 154     if (Split.second.empty()) {
 155       if (Repl.size() != Split.first.size() &&
 156           Error && Error->empty())
 157         *Error = "replacement string contained trailing backslash";
 158       break;
 159     }
 160
 161     // Otherwise update the replacement string and interpret escapes.
 162     Repl = Split.second;
 163
 164     // FIXME: We should have a StringExtras function for mapping C99 escapes.
 165     switch (Repl[0]) {
 166
 167       // Backreference with the "\g<ref>" syntax
 168     case 'g':
 169       if (Repl.size() >= 4 && Repl[1] == '<') {
 170         size_t End = Repl.find('>');
 171         StringRef Ref = Repl.slice(2, End);
 172         unsigned RefValue;
 173         if (End != StringRef::npos && !Ref.getAsInteger(10, RefValue)) {
 174           Repl = Repl.substr(End + 1);
 175           if (RefValue < Matches.size())
 176             Res += Matches[RefValue];
 177           else if (Error && Error->empty())
 178             *Error =
 179                 ("invalid backreference string 'g<" + Twine(Ref) + ">'").str();
 180           break;
 181         }
 182       }
 183       [[fallthrough]];
 184
 185       // Treat all unrecognized characters as self-quoting.
 186     default:
 187       Res += Repl[0];
 188       Repl = Repl.substr(1);
 189       break;
 190
 191       // Single character escapes.
 192     case 't':
 193       Res += '\t';
 194       Repl = Repl.substr(1);
 195       break;
 196     case 'n':
 197       Res += '\n';
 198       Repl = Repl.substr(1);
 199       break;
 200
 201       // Decimal escapes are backreferences.
 202     case '0': case '1': case '2': case '3': case '4':
 203     case '5': case '6': case '7': case '8': case '9': {
 204       // Extract the backreference number.
 205       StringRef Ref = Repl.slice(0, Repl.find_first_not_of("0123456789"));
 206       Repl = Repl.substr(Ref.size());
 207
 208       unsigned RefValue;
 209       if (!Ref.getAsInteger(10, RefValue) &&
 210           RefValue < Matches.size())
 211         Res += Matches[RefValue];
 212       else if (Error && Error->empty())
 213         *Error = ("invalid backreference string '" + Twine(Ref) + "'").str();
 214       break;
 215     }
 216     }
 217   }
 218
 219   // And finally the suffix.
 220   Res += StringRef(Matches[0].end(), String.end() - Matches[0].end());
 221
 222   return Res;
 223 }
 224
 225 // These are the special characters matched in functions like "p_ere_exp".
 226 static const char RegexMetachars[] = "()^$|*+?.[]\\{}";
 227
 228 bool Regex::isLiteralERE(StringRef Str) {
 229   // Check for regex metacharacters.  This list was derived from our regex
 230   // implementation in regcomp.c and double checked against the POSIX extended
 231   // regular expression specification.
 232   return Str.find_first_of(RegexMetachars) == StringRef::npos;
 233 }
 234
 235 std::string Regex::escape(StringRef String) {
 236   std::string RegexStr;
 237   for (char C : String) {
 238     if (strchr(RegexMetachars, C))
 239       RegexStr += '\\';
 240     RegexStr += C;
 241   }
 242
 243   return RegexStr;
 244 }