utils/TableGen/SequenceToOffsetTable.h

   1 //===-- SequenceToOffsetTable.h - Compress similar sequences ----*- 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 // SequenceToOffsetTable can be used to emit a number of null-terminated
  11 // sequences as one big array.  Use the same memory when a sequence is a suffix
  12 // of another.
  13 //
  14 //===----------------------------------------------------------------------===//
  15
  16 #ifndef LLVM_UTILS_TABLEGEN_SEQUENCETOOFFSETTABLE_H
  17 #define LLVM_UTILS_TABLEGEN_SEQUENCETOOFFSETTABLE_H
  18
  19 #include "llvm/Support/raw_ostream.h"
  20 #include <algorithm>
  21 #include <cassert>
  22 #include <cctype>
  23 #include <functional>
  24 #include <map>
  25
  26 namespace llvm {
  27
  28 /// SequenceToOffsetTable - Collect a number of terminated sequences of T.
  29 /// Compute the layout of a table that contains all the sequences, possibly by
  30 /// reusing entries.
  31 ///
  32 /// @tparam SeqT The sequence container. (vector or string).
  33 /// @tparam Less A stable comparator for SeqT elements.
  34 template<typename SeqT, typename Less = std::less<typename SeqT::value_type> >
  35 class SequenceToOffsetTable {
  36   typedef typename SeqT::value_type ElemT;
  37
  38   // Define a comparator for SeqT that sorts a suffix immediately before a
  39   // sequence with that suffix.
  40   struct SeqLess {
  41     Less L;
  42     bool operator()(const SeqT &A, const SeqT &B) const {
  43       return std::lexicographical_compare(A.rbegin(), A.rend(),
  44                                           B.rbegin(), B.rend(), L);
  45     }
  46   };
  47
  48   // Keep sequences ordered according to SeqLess so suffixes are easy to find.
  49   // Map each sequence to its offset in the table.
  50   typedef std::map<SeqT, unsigned, SeqLess> SeqMap;
  51
  52   // Sequences added so far, with suffixes removed.
  53   SeqMap Seqs;
  54
  55   // Entries in the final table, or 0 before layout was called.
  56   unsigned Entries;
  57
  58   // isSuffix - Returns true if A is a suffix of B.
  59   static bool isSuffix(const SeqT &A, const SeqT &B) {
  60     return A.size() <= B.size() && std::equal(A.rbegin(), A.rend(), B.rbegin());
  61   }
  62
  63 public:
  64   SequenceToOffsetTable() : Entries(0) {}
  65
  66   /// add - Add a sequence to the table.
  67   /// This must be called before layout().
  68   void add(const SeqT &Seq) {
  69     assert(Entries == 0 && "Cannot call add() after layout()");
  70     typename SeqMap::iterator I = Seqs.lower_bound(Seq);
  71
  72     // If SeqMap contains a sequence that has Seq as a suffix, I will be
  73     // pointing to it.
  74     if (I != Seqs.end() && isSuffix(Seq, I->first))
  75       return;
  76
  77     I = Seqs.insert(I, std::make_pair(Seq, 0u));
  78
  79     // The entry before I may be a suffix of Seq that can now be erased.
  80     if (I != Seqs.begin() && isSuffix((--I)->first, Seq))
  81       Seqs.erase(I);
  82   }
  83
  84   bool empty() const { return Seqs.empty(); }
  85
  86   unsigned size() const {
  87     assert(Entries && "Call layout() before size()");
  88     return Entries;
  89   }
  90
  91   /// layout - Computes the final table layout.
  92   void layout() {
  93     assert(Entries == 0 && "Can only call layout() once");
  94     // Lay out the table in Seqs iteration order.
  95     for (typename SeqMap::iterator I = Seqs.begin(), E = Seqs.end(); I != E;
  96          ++I) {
  97       I->second = Entries;
  98       // Include space for a terminator.
  99       Entries += I->first.size() + 1;
 100     }
 101   }
 102
 103   /// get - Returns the offset of Seq in the final table.
 104   unsigned get(const SeqT &Seq) const {
 105     assert(Entries && "Call layout() before get()");
 106     typename SeqMap::const_iterator I = Seqs.lower_bound(Seq);
 107     assert(I != Seqs.end() && isSuffix(Seq, I->first) &&
 108            "get() called with sequence that wasn't added first");
 109     return I->second + (I->first.size() - Seq.size());
 110   }
 111
 112   /// emit - Print out the table as the body of an array initializer.
 113   /// Use the Print function to print elements.
 114   void emit(raw_ostream &OS,
 115             void (*Print)(raw_ostream&, ElemT),
 116             const char *Term = "0") const {
 117     assert(Entries && "Call layout() before emit()");
 118     for (typename SeqMap::const_iterator I = Seqs.begin(), E = Seqs.end();
 119          I != E; ++I) {
 120       OS << "  /* " << I->second << " */ ";
 121       for (typename SeqT::const_iterator SI = I->first.begin(),
 122              SE = I->first.end(); SI != SE; ++SI) {
 123         Print(OS, *SI);
 124         OS << ", ";
 125       }
 126       OS << Term << ",\n";
 127     }
 128   }
 129 };
 130
 131 // Helper function for SequenceToOffsetTable<string>.
 132 static inline void printChar(raw_ostream &OS, char C) {
 133   unsigned char UC(C);
 134   if (isalnum(UC) || ispunct(UC)) {
 135     OS << '\'';
 136     if (C == '\\' || C == '\'')
 137       OS << '\\';
 138     OS << C << '\'';
 139   } else {
 140     OS << unsigned(UC);
 141   }
 142 }
 143
 144 } // end namespace llvm
 145
 146 #endif