Merge remote-tracking branch 'mvo/feature/hash-stats' into debian/experimental

Conflicts:
	apt-pkg/acquire-item.cc
	apt-pkg/acquire-item.h
	apt-pkg/deb/debmetaindex.h
	apt-pkg/pkgcache.cc
	test/integration/test-apt-ftparchive-src-cachedb

4 files changed, 56 insertions, 16 deletions

diff --git a/apt-pkg/pkgcache.cc b/apt-pkg/pkgcache.cc
index 2b6153634..4fbdc93d5 100644
--- a/apt-pkg/pkgcache.cc
+++ b/apt-pkg/pkgcache.cc
@@ -55,11 +55,7 @@ pkgCache::Header::Header()
    /* Whenever the structures change the major version should be bumped,
       whenever the generator changes the minor version should be bumped. */
    MajorVersion = 9;
-#if (APT_PKG_MAJOR >= 4 && APT_PKG_MINOR >= 13)
    MinorVersion = 2;
-#else
-   MinorVersion = 1;
-#endif
    Dirty = false;
    
    HeaderSz = sizeof(pkgCache::Header);
@@ -218,7 +214,7 @@ pkgCache::PkgIterator pkgCache::SingleArchFindPkg(const string &Name)
 {
    // Look at the hash bucket
    Package *Pkg = PkgP + HeaderP->PkgHashTable[Hash(Name)];
-   for (; Pkg != PkgP; Pkg = PkgP + Pkg->NextPackage)
+   for (; Pkg != PkgP; Pkg = PkgP + Pkg->Next)
    {
       if (unlikely(Pkg->Name == 0))
 	 continue;
@@ -374,7 +370,7 @@ pkgCache::PkgIterator pkgCache::GrpIterator::FindPkg(string Arch) const {
 	   (= different packages with same calculated hash),
 	   so we need to check the name also */
 	for (pkgCache::Package *Pkg = PackageList(); Pkg != Owner->PkgP;
-	     Pkg = Owner->PkgP + Pkg->NextPackage) {
+	     Pkg = Owner->PkgP + Pkg->Next) {
 		if (S->Name == Pkg->Name &&
 		    stringcasecmp(Arch, Owner->StrP + Pkg->Arch) == 0)
 			return PkgIterator(*Owner, Pkg);
@@ -423,7 +419,7 @@ pkgCache::PkgIterator pkgCache::GrpIterator::NextPkg(pkgCache::PkgIterator const
 	if (S->LastPackage == LastPkg.Index())
 		return PkgIterator(*Owner, 0);
 
-	return PkgIterator(*Owner, Owner->PkgP + LastPkg->NextPackage);
+	return PkgIterator(*Owner, Owner->PkgP + LastPkg->Next);
 }
 									/*}}}*/
 // GrpIterator::operator ++ - Postfix incr				/*{{{*/
@@ -450,7 +446,7 @@ void pkgCache::PkgIterator::operator ++(int)
 {
    // Follow the current links
    if (S != Owner->PkgP)
-      S = Owner->PkgP + S->NextPackage;
+      S = Owner->PkgP + S->Next;
 
    // Follow the hash table
    while (S == Owner->PkgP && (HashIndex+1) < (signed)_count(Owner->HeaderP->PkgHashTable))
diff --git a/apt-pkg/pkgcache.h b/apt-pkg/pkgcache.h
index 22dc6218c..55f0187f9 100644
--- a/apt-pkg/pkgcache.h
+++ b/apt-pkg/pkgcache.h
@@ -316,8 +316,8 @@ struct pkgCache::Header
        these packages are stored as a sequence in the list.
 
        Beware: The Hashmethod assumes that the hash table sizes are equal */
-   map_ptrloc PkgHashTable[2*1048];
-   map_ptrloc GrpHashTable[2*1048];
+   map_ptrloc PkgHashTable[64*1048];
+   map_ptrloc GrpHashTable[64*1048];
 
    /** \brief Size of the complete cache file */
    unsigned long  CacheFileSize;
@@ -390,7 +390,7 @@ struct pkgCache::Package
 
    // Linked list
    /** \brief Link to the next package in the same bucket */
-   map_ptrloc NextPackage;       // Package
+   map_ptrloc Next;       // Package
    /** \brief List of all dependencies on this package */
    map_ptrloc RevDepends;        // Dependency
    /** \brief List of all "packages" this package provide */
diff --git a/apt-pkg/pkgcachegen.cc b/apt-pkg/pkgcachegen.cc
index ac1cea0eb..9615b4c22 100644
--- a/apt-pkg/pkgcachegen.cc
+++ b/apt-pkg/pkgcachegen.cc
@@ -656,16 +656,16 @@ bool pkgCacheGenerator::NewPackage(pkgCache::PkgIterator &Pkg,const string &Name
       unsigned long const Hash = Cache.Hash(Name);
       map_ptrloc *insertAt = &Cache.HeaderP->PkgHashTable[Hash];
       while (*insertAt != 0 && strcasecmp(Name.c_str(), Cache.StrP + (Cache.PkgP + *insertAt)->Name) > 0)
-	 insertAt = &(Cache.PkgP + *insertAt)->NextPackage;
-      Pkg->NextPackage = *insertAt;
+	 insertAt = &(Cache.PkgP + *insertAt)->Next;
+      Pkg->Next = *insertAt;
       *insertAt = Package;
    }
    else // Group the Packages together
    {
       // this package is the new last package
       pkgCache::PkgIterator LastPkg(Cache, Cache.PkgP + Grp->LastPackage);
-      Pkg->NextPackage = LastPkg->NextPackage;
-      LastPkg->NextPackage = Package;
+      Pkg->Next = LastPkg->Next;
+      LastPkg->Next = Package;
    }
    Grp->LastPackage = Package;
 
diff --git a/cmdline/apt-cache.cc b/cmdline/apt-cache.cc
index 1414617eb..2ed1bf5d4 100644
--- a/cmdline/apt-cache.cc
+++ b/cmdline/apt-cache.cc
@@ -264,6 +264,44 @@ static bool DumpPackage(CommandLine &CmdL)
    return true;
 }
 									/*}}}*/
+// ShowHashTableStats - Show stats about a hashtable			/*{{{*/
+// ---------------------------------------------------------------------
+/* */
+template<class T>
+static void ShowHashTableStats(std::string Type,
+                               T *StartP,
+                               map_ptrloc *Hashtable,
+                               unsigned long Size)
+{
+   // hashtable stats for the HashTable
+   long NumBuckets = Size;
+   long UsedBuckets = 0;
+   long UnusedBuckets = 0;
+   long LongestBucket = 0;
+   long ShortestBucket = NumBuckets;
+   for (unsigned int i=0; i < NumBuckets; ++i)
+   {
+      T *P = StartP + Hashtable[i];
+      if(P == 0 || P == StartP)
+      {
+         UnusedBuckets++;
+         continue;
+      }
+      long ThisBucketSize = 0;
+      for (; P != StartP; P = StartP + P->Next)
+         ThisBucketSize++;
+      LongestBucket = std::max(ThisBucketSize, LongestBucket);
+      ShortestBucket = std::min(ThisBucketSize, ShortestBucket);
+      UsedBuckets += ThisBucketSize;
+   }
+   cout << "Total buckets " << Type << ": " << SizeToStr(NumBuckets) << std::endl;
+   cout << "  Unused: " << SizeToStr(UnusedBuckets) << std::endl;
+   cout << "  Used: " << UsedBuckets  << std::endl;
+   cout << "  Average entries: " << UsedBuckets/(double)NumBuckets << std::endl;
+   cout << "  Longest: " << LongestBucket << std::endl;
+   cout << "  Shortest: " << ShortestBucket << std::endl;
+}
+									/*}}}*/
 // Stats - Dump some nice statistics					/*{{{*/
 // ---------------------------------------------------------------------
 /* */
@@ -373,7 +411,13 @@ static bool Stats(CommandLine &)
            Cache->Head().VerFileCount*Cache->Head().VerFileSz +
            Cache->Head().ProvidesCount*Cache->Head().ProvidesSz;
    cout << _("Total space accounted for: ") << SizeToStr(Total) << endl;
-   
+
+   // hashtable stats
+   int HashTableSize = sizeof(Cache->HeaderP->PkgHashTable)/sizeof(map_ptrloc);
+   ShowHashTableStats<pkgCache::Package>("PkgHashTable", Cache->PkgP, Cache->HeaderP->PkgHashTable, HashTableSize);
+   HashTableSize = sizeof(Cache->HeaderP->GrpHashTable)/sizeof(map_ptrloc);
+   ShowHashTableStats<pkgCache::Group>("GrpHashTable", Cache->GrpP, Cache->HeaderP->GrpHashTable, HashTableSize);
+
    return true;
 }
 									/*}}}*/