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