Rename CStringMap -> StringMap, since it now supports nul characters in the

strings.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@34064 91177308-0d34-0410-b5e6-96231b3b80d8
This commit is contained in:
Chris Lattner 2007-02-08 19:20:57 +00:00
parent 796f9faffd
commit bb28a81ba3
4 changed files with 25 additions and 345 deletions

View File

@ -1,182 +0,0 @@
//===--- CStringMap.h - CString Hash table map interface --------*- C++ -*-===//
//
// The LLVM Compiler Infrastructure
//
// This file was developed by Chris Lattner and is distributed under
// the University of Illinois Open Source License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
// This file defines the CStringMap class.
//
//===----------------------------------------------------------------------===//
#ifndef LLVM_ADT_CSTRINGMAP_H
#define LLVM_ADT_CSTRINGMAP_H
#include "llvm/Support/Allocator.h"
#include <cstring>
namespace llvm {
/// StringMapEntryBase - Shared base class of StringMapEntry instances.
class StringMapEntryBase {
unsigned StrLen;
public:
StringMapEntryBase(unsigned Len) : StrLen(Len) {}
unsigned getKeyLength() const { return StrLen; }
};
/// CStringMapVisitor - Subclasses of this class may be implemented to walk all
/// of the items in a CStringMap.
class CStringMapVisitor {
public:
virtual ~CStringMapVisitor();
virtual void Visit(const char *Key, StringMapEntryBase *Value) const = 0;
};
/// CStringMapImpl - This is the base class of CStringMap that is shared among
/// all of its instantiations.
class CStringMapImpl {
protected:
/// ItemBucket - The hash table consists of an array of these. If Item is
/// non-null, this is an extant entry, otherwise, it is a hole.
struct ItemBucket {
/// FullHashValue - This remembers the full hash value of the key for
/// easy scanning.
unsigned FullHashValue;
/// Item - This is a pointer to the actual item object.
StringMapEntryBase *Item;
};
ItemBucket *TheTable;
unsigned NumBuckets;
unsigned NumItems;
unsigned ItemSize;
protected:
CStringMapImpl(unsigned InitSize, unsigned ItemSize);
void RehashTable();
/// LookupBucketFor - Look up the bucket that the specified string should end
/// up in. If it already exists as a key in the map, the Item pointer for the
/// specified bucket will be non-null. Otherwise, it will be null. In either
/// case, the FullHashValue field of the bucket will be set to the hash value
/// of the string.
unsigned LookupBucketFor(const char *KeyStart, const char *KeyEnd);
public:
unsigned getNumBuckets() const { return NumBuckets; }
unsigned getNumItems() const { return NumItems; }
void VisitEntries(const CStringMapVisitor &Visitor) const;
};
/// StringMapEntry - This is used to represent one value that is inserted into
/// a StringMap. It contains the Value itself and the key: the string length
/// and data.
template<typename ValueTy>
class StringMapEntry : public StringMapEntryBase {
ValueTy Val;
public:
StringMapEntry(unsigned StrLen)
: StringMapEntryBase(StrLen), Val() {}
StringMapEntry(unsigned StrLen, const ValueTy &V)
: StringMapEntryBase(StrLen), Val(V) {}
const ValueTy &getValue() const { return Val; }
ValueTy &getValue() { return Val; }
void setValue(const ValueTy &V) { Val = V; }
/// getKeyData - Return the start of the string data that is the key for this
/// value. The string data is always stored immediately after the
/// StringMapEntry object.
const char *getKeyData() const {return reinterpret_cast<const char*>(this+1);}
};
/// CStringMap - This is an unconventional map that is specialized for handling
/// keys that are "strings", which are basically ranges of bytes. This does some
/// funky memory allocation and hashing things to make it extremely efficient,
/// storing the string data *after* the value in the map.
template<typename ValueTy, typename AllocatorTy = MallocAllocator>
class CStringMap : public CStringMapImpl {
AllocatorTy Allocator;
typedef StringMapEntry<ValueTy> MapEntryTy;
public:
CStringMap(unsigned InitialSize = 0)
: CStringMapImpl(InitialSize, sizeof(MapEntryTy)) {}
AllocatorTy &getAllocator() { return Allocator; }
const AllocatorTy &getAllocator() const { return Allocator; }
/// FindValue - Look up the specified key in the map. If it exists, return a
/// pointer to the element, otherwise return null.
MapEntryTy *FindValue(const char *KeyStart, const char *KeyEnd) {
unsigned BucketNo = LookupBucketFor(KeyStart, KeyEnd);
return static_cast<MapEntryTy*>(TheTable[BucketNo].Item);
}
/// GetOrCreateValue - Look up the specified key in the table. If a value
/// exists, return it. Otherwise, default construct a value, insert it, and
/// return.
StringMapEntry<ValueTy> &GetOrCreateValue(const char *KeyStart,
const char *KeyEnd) {
unsigned BucketNo = LookupBucketFor(KeyStart, KeyEnd);
ItemBucket &Bucket = TheTable[BucketNo];
if (Bucket.Item)
return *static_cast<MapEntryTy*>(Bucket.Item);
unsigned KeyLength = KeyEnd-KeyStart;
// Okay, the item doesn't already exist, and 'Bucket' is the bucket to fill
// in. Allocate a new item with space for the string at the end and a null
// terminator.
unsigned AllocSize = sizeof(MapEntryTy)+KeyLength+1;
#ifdef __GNUC__
unsigned Alignment = __alignof__(MapEntryTy);
#else
// FIXME: ugly.
unsigned Alignment = 8;
#endif
MapEntryTy *NewItem =
static_cast<MapEntryTy*>(Allocator.Allocate(AllocSize, Alignment));
// Default construct the value.
new (NewItem) MapEntryTy(KeyLength);
++NumItems;
// Copy the string information.
char *StrBuffer = const_cast<char*>(NewItem->getKeyData());
memcpy(StrBuffer, KeyStart, KeyLength);
StrBuffer[KeyLength] = 0; // Null terminate for convenience of clients.
// Fill in the bucket for the hash table. The FullHashValue was already
// filled in by LookupBucketFor.
Bucket.Item = NewItem;
// If the hash table is now more than 3/4 full, rehash into a larger table.
if (NumItems > NumBuckets*3/4)
RehashTable();
return *NewItem;
}
~CStringMap() {
for (ItemBucket *I = TheTable, *E = TheTable+NumBuckets; I != E; ++I) {
if (MapEntryTy *Id = static_cast<MapEntryTy*>(I->Item)) {
// Free memory referenced by the item.
Id->~MapEntryTy();
Allocator.Deallocate(Id);
}
}
delete [] TheTable;
}
};
}
#endif

View File

@ -1,4 +1,4 @@
//===--- CStringMap.h - CString Hash table map interface --------*- C++ -*-===//
//===--- StringMap.h - String Hash table map interface ----------*- C++ -*-===//
//
// The LLVM Compiler Infrastructure
//
@ -7,12 +7,12 @@
//
//===----------------------------------------------------------------------===//
//
// This file defines the CStringMap class.
// This file defines the StringMap class.
//
//===----------------------------------------------------------------------===//
#ifndef LLVM_ADT_CSTRINGMAP_H
#define LLVM_ADT_CSTRINGMAP_H
#ifndef LLVM_ADT_STRINGMAP_H
#define LLVM_ADT_STRINGMAP_H
#include "llvm/Support/Allocator.h"
#include <cstring>
@ -28,17 +28,17 @@ public:
unsigned getKeyLength() const { return StrLen; }
};
/// CStringMapVisitor - Subclasses of this class may be implemented to walk all
/// of the items in a CStringMap.
class CStringMapVisitor {
/// StringMapVisitor - Subclasses of this class may be implemented to walk all
/// of the items in a StringMap.
class StringMapVisitor {
public:
virtual ~CStringMapVisitor();
virtual ~StringMapVisitor();
virtual void Visit(const char *Key, StringMapEntryBase *Value) const = 0;
};
/// CStringMapImpl - This is the base class of CStringMap that is shared among
/// StringMapImpl - This is the base class of StringMap that is shared among
/// all of its instantiations.
class CStringMapImpl {
class StringMapImpl {
protected:
/// ItemBucket - The hash table consists of an array of these. If Item is
/// non-null, this is an extant entry, otherwise, it is a hole.
@ -56,7 +56,7 @@ protected:
unsigned NumItems;
unsigned ItemSize;
protected:
CStringMapImpl(unsigned InitSize, unsigned ItemSize);
StringMapImpl(unsigned InitSize, unsigned ItemSize);
void RehashTable();
/// LookupBucketFor - Look up the bucket that the specified string should end
@ -70,7 +70,7 @@ public:
unsigned getNumBuckets() const { return NumBuckets; }
unsigned getNumItems() const { return NumItems; }
void VisitEntries(const CStringMapVisitor &Visitor) const;
void VisitEntries(const StringMapVisitor &Visitor) const;
};
/// StringMapEntry - This is used to represent one value that is inserted into
@ -97,17 +97,17 @@ public:
};
/// CStringMap - This is an unconventional map that is specialized for handling
/// StringMap - This is an unconventional map that is specialized for handling
/// keys that are "strings", which are basically ranges of bytes. This does some
/// funky memory allocation and hashing things to make it extremely efficient,
/// storing the string data *after* the value in the map.
template<typename ValueTy, typename AllocatorTy = MallocAllocator>
class CStringMap : public CStringMapImpl {
class StringMap : public StringMapImpl {
AllocatorTy Allocator;
typedef StringMapEntry<ValueTy> MapEntryTy;
public:
CStringMap(unsigned InitialSize = 0)
: CStringMapImpl(InitialSize, sizeof(MapEntryTy)) {}
StringMap(unsigned InitialSize = 0)
: StringMapImpl(InitialSize, sizeof(MapEntryTy)) {}
AllocatorTy &getAllocator() { return Allocator; }
const AllocatorTy &getAllocator() const { return Allocator; }
@ -164,7 +164,7 @@ public:
return *NewItem;
}
~CStringMap() {
~StringMap() {
for (ItemBucket *I = TheTable, *E = TheTable+NumBuckets; I != E; ++I) {
if (MapEntryTy *Id = static_cast<MapEntryTy*>(I->Item)) {
// Free memory referenced by the item.

View File

@ -1,138 +0,0 @@
//===--- CStringMap.cpp - CString Hash table map implementation -----------===//
//
// The LLVM Compiler Infrastructure
//
// This file was developed by Chris Lattner and is distributed under
// the University of Illinois Open Source License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
// This file implements the CStringMap class.
//
//===----------------------------------------------------------------------===//
#include "llvm/ADT/CStringMap.h"
#include <cassert>
using namespace llvm;
CStringMapVisitor::~CStringMapVisitor() {
}
CStringMapImpl::CStringMapImpl(unsigned InitSize, unsigned itemSize) {
assert((InitSize & (InitSize-1)) == 0 &&
"Init Size must be a power of 2 or zero!");
NumBuckets = InitSize ? InitSize : 512;
ItemSize = itemSize;
NumItems = 0;
TheTable = new ItemBucket[NumBuckets]();
memset(TheTable, 0, NumBuckets*sizeof(ItemBucket));
}
/// HashString - Compute a hash code for the specified string.
///
static unsigned HashString(const char *Start, const char *End) {
// Bernstein hash function.
unsigned int Result = 0;
// TODO: investigate whether a modified bernstein hash function performs
// better: http://eternallyconfuzzled.com/tuts/algorithms/jsw_tut_hashing.aspx
// X*33+c -> X*33^c
while (Start != End)
Result = Result * 33 + *Start++;
Result = Result + (Result >> 5);
return Result;
}
/// LookupBucketFor - Look up the bucket that the specified string should end
/// up in. If it already exists as a key in the map, the Item pointer for the
/// specified bucket will be non-null. Otherwise, it will be null. In either
/// case, the FullHashValue field of the bucket will be set to the hash value
/// of the string.
unsigned CStringMapImpl::LookupBucketFor(const char *NameStart,
const char *NameEnd) {
unsigned HTSize = NumBuckets;
unsigned FullHashValue = HashString(NameStart, NameEnd);
unsigned BucketNo = FullHashValue & (HTSize-1);
unsigned ProbeAmt = 1;
while (1) {
ItemBucket &Bucket = TheTable[BucketNo];
StringMapEntryBase *BucketItem = Bucket.Item;
// If we found an empty bucket, this key isn't in the table yet, return it.
if (BucketItem == 0) {
Bucket.FullHashValue = FullHashValue;
return BucketNo;
}
// If the full hash value matches, check deeply for a match. The common
// case here is that we are only looking at the buckets (for item info
// being non-null and for the full hash value) not at the items. This
// is important for cache locality.
if (Bucket.FullHashValue == FullHashValue) {
// Do the comparison like this because NameStart isn't necessarily
// null-terminated!
char *ItemStr = (char*)BucketItem+ItemSize;
unsigned ItemStrLen = BucketItem->getKeyLength();
if (unsigned(NameEnd-NameStart) == ItemStrLen &&
memcmp(ItemStr, NameStart, ItemStrLen) == 0) {
// We found a match!
return BucketNo;
}
}
// Okay, we didn't find the item. Probe to the next bucket.
BucketNo = (BucketNo+ProbeAmt) & (HTSize-1);
// Use quadratic probing, it has fewer clumping artifacts than linear
// probing and has good cache behavior in the common case.
++ProbeAmt;
}
}
/// RehashTable - Grow the table, redistributing values into the buckets with
/// the appropriate mod-of-hashtable-size.
void CStringMapImpl::RehashTable() {
unsigned NewSize = NumBuckets*2;
ItemBucket *NewTableArray = new ItemBucket[NewSize]();
memset(NewTableArray, 0, NewSize*sizeof(ItemBucket));
// Rehash all the items into their new buckets. Luckily :) we already have
// the hash values available, so we don't have to rehash any strings.
for (ItemBucket *IB = TheTable, *E = TheTable+NumBuckets; IB != E; ++IB) {
if (IB->Item) {
// Fast case, bucket available.
unsigned FullHash = IB->FullHashValue;
unsigned NewBucket = FullHash & (NewSize-1);
if (NewTableArray[NewBucket].Item == 0) {
NewTableArray[FullHash & (NewSize-1)].Item = IB->Item;
NewTableArray[FullHash & (NewSize-1)].FullHashValue = FullHash;
continue;
}
unsigned ProbeSize = 1;
do {
NewBucket = (NewBucket + ProbeSize++) & (NewSize-1);
} while (NewTableArray[NewBucket].Item);
// Finally found a slot. Fill it in.
NewTableArray[NewBucket].Item = IB->Item;
NewTableArray[NewBucket].FullHashValue = FullHash;
}
}
delete[] TheTable;
TheTable = NewTableArray;
NumBuckets = NewSize;
}
/// VisitEntries - This method walks through all of the items,
/// invoking Visitor.Visit for each of them.
void CStringMapImpl::VisitEntries(const CStringMapVisitor &Visitor) const {
for (ItemBucket *IB = TheTable, *E = TheTable+NumBuckets; IB != E; ++IB) {
if (StringMapEntryBase *Id = IB->Item)
Visitor.Visit((char*)Id + ItemSize, Id);
}
}

View File

@ -1,4 +1,4 @@
//===--- CStringMap.cpp - CString Hash table map implementation -----------===//
//===--- StringMap.cpp - String Hash table map implementation -------------===//
//
// The LLVM Compiler Infrastructure
//
@ -7,18 +7,18 @@
//
//===----------------------------------------------------------------------===//
//
// This file implements the CStringMap class.
// This file implements the StringMap class.
//
//===----------------------------------------------------------------------===//
#include "llvm/ADT/CStringMap.h"
#include "llvm/ADT/StringMap.h"
#include <cassert>
using namespace llvm;
CStringMapVisitor::~CStringMapVisitor() {
StringMapVisitor::~StringMapVisitor() {
}
CStringMapImpl::CStringMapImpl(unsigned InitSize, unsigned itemSize) {
StringMapImpl::StringMapImpl(unsigned InitSize, unsigned itemSize) {
assert((InitSize & (InitSize-1)) == 0 &&
"Init Size must be a power of 2 or zero!");
NumBuckets = InitSize ? InitSize : 512;
@ -49,7 +49,7 @@ static unsigned HashString(const char *Start, const char *End) {
/// specified bucket will be non-null. Otherwise, it will be null. In either
/// case, the FullHashValue field of the bucket will be set to the hash value
/// of the string.
unsigned CStringMapImpl::LookupBucketFor(const char *NameStart,
unsigned StringMapImpl::LookupBucketFor(const char *NameStart,
const char *NameEnd) {
unsigned HTSize = NumBuckets;
unsigned FullHashValue = HashString(NameStart, NameEnd);
@ -92,7 +92,7 @@ unsigned CStringMapImpl::LookupBucketFor(const char *NameStart,
/// RehashTable - Grow the table, redistributing values into the buckets with
/// the appropriate mod-of-hashtable-size.
void CStringMapImpl::RehashTable() {
void StringMapImpl::RehashTable() {
unsigned NewSize = NumBuckets*2;
ItemBucket *NewTableArray = new ItemBucket[NewSize]();
memset(NewTableArray, 0, NewSize*sizeof(ItemBucket));
@ -130,7 +130,7 @@ void CStringMapImpl::RehashTable() {
/// VisitEntries - This method walks through all of the items,
/// invoking Visitor.Visit for each of them.
void CStringMapImpl::VisitEntries(const CStringMapVisitor &Visitor) const {
void StringMapImpl::VisitEntries(const StringMapVisitor &Visitor) const {
for (ItemBucket *IB = TheTable, *E = TheTable+NumBuckets; IB != E; ++IB) {
if (StringMapEntryBase *Id = IB->Item)
Visitor.Visit((char*)Id + ItemSize, Id);