cache: Introduce partial SourceVersion support

This is the first step that introduces a 1:1 mapping between version
and source version. In a future version this can use the fields
currently marked unavailable to deduplicate the SourceVersion
objects across the group.

The policy gains a member for storing pins for sourceversions.

Together, in the future we should be able to determine candidates
for source versions.

8 files changed, 99 insertions, 29 deletions

diff --git a/apt-pkg/cacheiterators.h b/apt-pkg/cacheiterators.h
index ceeb4b3a7..443912253 100644
--- a/apt-pkg/cacheiterators.h
+++ b/apt-pkg/cacheiterators.h
@@ -188,6 +188,31 @@ class APT_PUBLIC pkgCache::PkgIterator: public Iterator<Package, PkgIterator> {
 	inline PkgIterator() : Iterator<Package, PkgIterator>(), HashIndex(0) {}
 };
 									/*}}}*/
+// SourceVersion Iterator						/*{{{*/
+class APT_PUBLIC pkgCache::SrcVerIterator : public Iterator<SourceVersion, SrcVerIterator>
+{
+   public:
+   inline SourceVersion *OwnerPointer() const
+   {
+      return (Owner != 0) ? Owner->SrcVerP : 0;
+   }
+#if 0
+	// Iteration
+	inline SrcVerIterator& operator++() {if (S != Owner->SrcVerP) S = Owner->SrcVerP + S->NextSourceVersion; return *this;}
+	inline SrcVerIterator operator++(int) { SrcVerIterator const tmp(*this); operator++(); return tmp; }
+#endif
+   inline APT_PURE GrpIterator Group() const { return GrpIterator(*Owner, Owner->GrpP + S->Group); }
+   inline const char *VerStr() const { return S->VerStr == 0 ? 0 : Owner->StrP + S->VerStr; }
+
+   inline SrcVerIterator(pkgCache &Owner, SourceVersion *Trg = 0) : Iterator<SourceVersion, SrcVerIterator>(Owner, Trg)
+   {
+      if (S == 0)
+	 S = OwnerPointer();
+   }
+   inline SrcVerIterator() : Iterator<SourceVersion, SrcVerIterator>() {}
+};
+									/*}}}*/
+
 // Version Iterator							/*{{{*/
 class APT_PUBLIC pkgCache::VerIterator : public Iterator<Version, VerIterator> {
 	public:
@@ -219,12 +244,15 @@ class APT_PUBLIC pkgCache::VerIterator : public Iterator<Version, VerIterator> {
 	// Accessors
 	inline const char *VerStr() const {return S->VerStr == 0?0:Owner->StrP + S->VerStr;}
 	inline const char *Section() const {return S->Section == 0?0:Owner->StrP + S->Section;}
+	/** \brief source version this version comes from
+	   Always contains the version string, even if it is the same as the binary version */
+	SrcVerIterator SourceVersion() const { return SrcVerIterator(*Owner, Owner->SrcVerP + S->SourceVersion); }
 	/** \brief source package name this version comes from
 	   Always contains the name, even if it is the same as the binary name */
-	inline const char *SourcePkgName() const {return Owner->StrP + S->SourcePkgName;}
+	inline const char *SourcePkgName() const { return SourceVersion().Group().Name(); }
 	/** \brief source version this version comes from
 	   Always contains the version string, even if it is the same as the binary version */
-	inline const char *SourceVerStr() const {return Owner->StrP + S->SourceVerStr;}
+	inline const char *SourceVerStr() const { return SourceVersion().VerStr(); }
 	inline const char *Arch() const {
 		if ((S->MultiArch & pkgCache::Version::All) == pkgCache::Version::All)
 			return "all";
diff --git a/apt-pkg/deb/deblistparser.cc b/apt-pkg/deb/deblistparser.cc
index e3041f848..dfb07d2a6 100644
--- a/apt-pkg/deb/deblistparser.cc
+++ b/apt-pkg/deb/deblistparser.cc
@@ -169,8 +169,8 @@ bool debListParser::NewVersion(pkgCache::VerIterator &Ver)
    pkgCache::GrpIterator G = Ver.ParentPkg().Group();
 
    // Setup the defaults
-   Ver->SourcePkgName = G->Name;
-   Ver->SourceVerStr = Ver->VerStr;
+   Ver.SourceVersion()->Group = G.MapPointer();
+   Ver.SourceVersion()->VerStr = Ver->VerStr;
 
    // Parse the name and version str
    if (Section.Find(pkgTagSection::Key::Source,Start,Stop) == true)
@@ -191,7 +191,7 @@ bool debListParser::NewVersion(pkgCache::VerIterator &Ver)
 	       {
 		  map_stringitem_t const idx = StoreString(pkgCacheGenerator::VERSIONNUMBER, version);
 		  G = Ver.ParentPkg().Group();
-		  Ver->SourceVerStr = idx;
+		  Ver.SourceVersion()->VerStr = idx;
 	       }
 	    }
 	 }
@@ -208,7 +208,7 @@ bool debListParser::NewVersion(pkgCache::VerIterator &Ver)
    }
 
    // Link into by source package group.
-   Ver->SourcePkgName = G->Name;
+   Ver.SourceVersion()->Group = G.MapPointer();
    Ver->NextInSource = G->VersionsInSource;
    G->VersionsInSource = Ver.MapPointer();
 
diff --git a/apt-pkg/edsp/edsplistparser.cc b/apt-pkg/edsp/edsplistparser.cc
index d3e761be4..f591a6a54 100644
--- a/apt-pkg/edsp/edsplistparser.cc
+++ b/apt-pkg/edsp/edsplistparser.cc
@@ -52,7 +52,7 @@ bool edspLikeListParser::NewVersion(pkgCache::VerIterator &Ver)
       if (version != Ver.VerStr())
       {
 	 map_stringitem_t const idx = StoreString(pkgCacheGenerator::VERSIONNUMBER, version);
-	 Ver->SourceVerStr = idx;
+	 Ver.SourceVersion()->VerStr = idx;
       }
    }
 
diff --git a/apt-pkg/pkgcache.cc b/apt-pkg/pkgcache.cc
index 2beef0e74..a27b1b054 100644
--- a/apt-pkg/pkgcache.cc
+++ b/apt-pkg/pkgcache.cc
@@ -67,6 +67,7 @@ pkgCache::Header::Header()
    APT_HEADER_SET(PackageSz, sizeof(pkgCache::Package));
    APT_HEADER_SET(ReleaseFileSz, sizeof(pkgCache::ReleaseFile));
    APT_HEADER_SET(PackageFileSz, sizeof(pkgCache::PackageFile));
+   APT_HEADER_SET(SourceVersionSz, sizeof(pkgCache::SourceVersion));
    APT_HEADER_SET(VersionSz, sizeof(pkgCache::Version) + sizeof(pkgCache::Version::Extra));
    APT_HEADER_SET(DescriptionSz, sizeof(pkgCache::Description));
    APT_HEADER_SET(DependencySz, sizeof(pkgCache::Dependency));
@@ -79,6 +80,7 @@ pkgCache::Header::Header()
    GroupCount = 0;
    PackageCount = 0;
    VersionCount = 0;
+   SourceVersionCount = 0;
    DescriptionCount = 0;
    DependsCount = 0;
    DependsDataCount = 0;
@@ -112,6 +114,7 @@ bool pkgCache::Header::CheckSizes(Header &Against) const
        ReleaseFileSz == Against.ReleaseFileSz &&
        PackageFileSz == Against.PackageFileSz &&
        VersionSz == Against.VersionSz &&
+       SourceVersionSz == Against.SourceVersionSz &&
        DescriptionSz == Against.DescriptionSz &&
        DependencySz == Against.DependencySz &&
        DependencyDataSz == Against.DependencyDataSz &&
@@ -151,6 +154,7 @@ bool pkgCache::ReMap(bool const &Errorchecks)
    RlsFileP = (ReleaseFile *)Map.Data();
    PkgFileP = (PackageFile *)Map.Data();
    VerP = (Version *)Map.Data();
+   SrcVerP = (SourceVersion *)Map.Data();
    DescP = (Description *)Map.Data();
    ProvideP = (Provides *)Map.Data();
    DepP = (Dependency *)Map.Data();
diff --git a/apt-pkg/pkgcache.h b/apt-pkg/pkgcache.h
index 22eea3f6c..22e13e009 100644
--- a/apt-pkg/pkgcache.h
+++ b/apt-pkg/pkgcache.h
@@ -130,6 +130,7 @@ class APT_PUBLIC pkgCache								/*{{{*/
    struct Package;
    struct ReleaseFile;
    struct PackageFile;
+   struct SourceVersion;
    struct Version;
    struct Description;
    struct Provides;
@@ -144,6 +145,7 @@ class APT_PUBLIC pkgCache								/*{{{*/
    class GrpIterator;
    class PkgIterator;
    class VerIterator;
+   class SrcVerIterator;
    class DescIterator;
    class DepIterator;
    class PrvIterator;
@@ -224,13 +226,14 @@ class APT_PUBLIC pkgCache								/*{{{*/
    DescFile *DescFileP;
    ReleaseFile *RlsFileP;
    PackageFile *PkgFileP;
+   SourceVersion *SrcVerP; // reserved for SourceVersion objects
    Version *VerP;
    Description *DescP;
    Provides *ProvideP;
    Dependency *DepP;
    DependencyData *DepDataP;
    char *StrP;
-   void *reserved[12];
+   void *reserved[13];
 
    virtual bool ReMap(bool const &Errorchecks = true);
    inline bool Sync() {return Map.Sync();}
@@ -320,6 +323,7 @@ struct pkgCache::Header
    map_number_t ReleaseFileSz;
    map_number_t PackageFileSz;
    map_number_t VersionSz;
+   map_number_t SourceVersionSz;
    map_number_t DescriptionSz;
    map_number_t DependencySz;
    map_number_t DependencyDataSz;
@@ -335,6 +339,7 @@ struct pkgCache::Header
    map_id_t GroupCount;
    map_id_t PackageCount;
    map_id_t VersionCount;
+   map_id_t SourceVersionCount;
    map_id_t DescriptionCount;
    map_id_t DependsCount;
    map_id_t DependsDataCount;
@@ -371,7 +376,7 @@ struct pkgCache::Header
        twice the number of pools as there are non-private structure types. The generator
        stores this information so future additions can make use of any unused pool
        blocks. */
-   DynamicMMap::Pool Pools[2 * 12];
+   DynamicMMap::Pool Pools[2 * 13];
 
    /** \brief hash tables providing rapid group/package name lookup
 
@@ -431,6 +436,9 @@ struct pkgCache::Group
    /** \brief List of binary produces by source package with this name. */
    map_pointer<Version> VersionsInSource;
 
+   /** \brief SourceVersionList */
+   map_pointer<SourceVersion> SourceVersionList;
+
    /** \brief Private pointer */
    map_pointer<void> d;
 };
@@ -612,6 +620,27 @@ struct pkgCache::DescFile
    map_filesize_t Offset;         // File offset
 };
 									/*}}}*/
+// SourceVersion structure					  	/*{{{*/
+/** \brief information for a single version of a source package
+
+    The version list is always sorted from highest version to lowest
+    version by the generator. Equal version numbers are either merged
+    or handled as separate versions based on the Hash value. */
+struct pkgCache::SourceVersion
+{
+   /** \brief unique sequel ID */
+   map_id_t ID;
+   /** \brief Group the source package belongs too */
+   map_pointer<pkgCache::Group> Group;
+   /** \brief complete version string */
+   map_stringitem_t VerStr;
+   map_pointer<Version> VersionList [[gnu::unavailable("not yet available")]];
+   map_pointer<SourceVersion> NextSourceVersion [[gnu::unavailable("not yet available")]];
+
+   /** \brief Private pointer */
+   map_pointer<void> d;
+};
+									/*}}}*/
 // Version structure							/*{{{*/
 /** \brief information for a single version of a package
 
@@ -626,12 +655,11 @@ struct pkgCache::Version
    map_stringitem_t VerStr;
    /** \brief section this version is filled in */
    map_stringitem_t Section;
-   /** \brief source package name this version comes from
-      Always contains the name, even if it is the same as the binary name */
-   map_stringitem_t SourcePkgName;
-   /** \brief source version this version comes from
-      Always contains the version string, even if it is the same as the binary version */
-   map_stringitem_t SourceVerStr;
+
+   /** \brief the source version object */
+   map_pointer<pkgCache::SourceVersion> SourceVersion;
+   /** \brief next version in the source package (might be different binary) */
+   map_pointer<Version> NextInSourceVersion;
 
    /** \brief Multi-Arch capabilities of a package version */
    enum VerMultiArch { No = 0, /*!< is the default and doesn't trigger special behaviour */
@@ -825,6 +853,7 @@ class pkgCache::Namespace						/*{{{*/
    typedef pkgCache::GrpIterator GrpIterator;
    typedef pkgCache::PkgIterator PkgIterator;
    typedef pkgCache::VerIterator VerIterator;
+   typedef pkgCache::SrcVerIterator SrcVerIterator;
    typedef pkgCache::DescIterator DescIterator;
    typedef pkgCache::DepIterator DepIterator;
    typedef pkgCache::PrvIterator PrvIterator;
diff --git a/apt-pkg/pkgcachegen.cc b/apt-pkg/pkgcachegen.cc
index fec931cf2..6e115e4f4 100644
--- a/apt-pkg/pkgcachegen.cc
+++ b/apt-pkg/pkgcachegen.cc
@@ -894,6 +894,11 @@ map_pointer<pkgCache::Version> pkgCacheGenerator::NewVersion(pkgCache::VerIterat
    Ver->d = d;
    if (not Ver.PhasedUpdatePercentage(100))
       abort();
+   auto SourceVersion = AllocateInMap<pkgCache::SourceVersion>(); // sequence point so Ver can be moved if needed
+   Ver->SourceVersion = SourceVersion;
+   Ver.SourceVersion()->ID = Cache.HeaderP->SourceVersionCount++;
+   if (not Ver.SourceVersion())
+      abort();
 
    //Dynamic<pkgCache::VerIterator> DynV(Ver); // caller MergeListVersion already takes care of it
    Ver->NextVer = Next;
diff --git a/apt-pkg/policy.h b/apt-pkg/policy.h
index 36c32b644..3f126a7c0 100644
--- a/apt-pkg/policy.h
+++ b/apt-pkg/policy.h
@@ -58,6 +58,7 @@ class APT_PUBLIC pkgPolicy : public pkgDepCache::Policy
       explicit PkgPin(std::string const &Pkg) : Pin(), Pkg(Pkg) {};
    };
 
+   std::unique_ptr<Pin[]> SrcVerPins;
    std::unique_ptr<Pin[]> VerPins;
    std::unique_ptr<signed short[]> PFPriority;
    std::vector<Pin> Defaults;
diff --git a/cmdline/apt-cache.cc b/cmdline/apt-cache.cc
index 67dde67a6..d8e9e7e4b 100644
--- a/cmdline/apt-cache.cc
+++ b/cmdline/apt-cache.cc
@@ -239,6 +239,8 @@ static bool Stats(CommandLine &CmdL)
    cout << _("  Mixed virtual packages: ") << NVirt << endl;
    cout << _("  Missing: ") << Missing << endl;
 
+   cout << _("Total distinct source versions: ") << Cache->Head().SourceVersionCount << " (" << SizeToStr(Cache->Head().SourceVersionCount * Cache->Head().SourceVersionSz) << ')' << endl;
+
    cout << _("Total distinct versions: ") << Cache->Head().VersionCount << " (" <<
       SizeToStr(Cache->Head().VersionCount*Cache->Head().VersionSz) << ')' << endl;
    cout << _("Total distinct descriptions: ") << Cache->Head().DescriptionCount << " (" <<
@@ -266,8 +268,8 @@ static bool Stats(CommandLine &CmdL)
 	    stritems.insert(V->VerStr);
 	 if (V->Section != 0)
 	    stritems.insert(V->Section);
-	 stritems.insert(V->SourcePkgName);
-	 stritems.insert(V->SourceVerStr);
+	 // FIXME: Count the source versions
+	 stritems.insert(V.SourceVersion()->VerStr);
 	 for (pkgCache::DepIterator D = V.DependsList(); D.end() == false; ++D)
 	 {
 	    if (D->Version != 0)
@@ -317,18 +319,19 @@ static bool Stats(CommandLine &CmdL)
    unsigned long Total = 0;
 #define APT_CACHESIZE(X,Y) (Cache->Head().X * Cache->Head().Y)
    Total = Slack + Size +
-      APT_CACHESIZE(GroupCount, GroupSz) +
-      APT_CACHESIZE(PackageCount, PackageSz) +
-      APT_CACHESIZE(VersionCount, VersionSz) +
-      APT_CACHESIZE(DescriptionCount, DescriptionSz) +
-      APT_CACHESIZE(DependsCount, DependencySz) +
-      APT_CACHESIZE(DependsDataCount, DependencyDataSz) +
-      APT_CACHESIZE(ReleaseFileCount, ReleaseFileSz) +
-      APT_CACHESIZE(PackageFileCount, PackageFileSz) +
-      APT_CACHESIZE(VerFileCount, VerFileSz) +
-      APT_CACHESIZE(DescFileCount, DescFileSz) +
-      APT_CACHESIZE(ProvidesCount, ProvidesSz) +
-      (2 * Cache->Head().GetHashTableSize() * sizeof(map_id_t));
+	   APT_CACHESIZE(GroupCount, GroupSz) +
+	   APT_CACHESIZE(PackageCount, PackageSz) +
+	   APT_CACHESIZE(VersionCount, VersionSz) +
+	   APT_CACHESIZE(SourceVersionCount, SourceVersionSz) +
+	   APT_CACHESIZE(DescriptionCount, DescriptionSz) +
+	   APT_CACHESIZE(DependsCount, DependencySz) +
+	   APT_CACHESIZE(DependsDataCount, DependencyDataSz) +
+	   APT_CACHESIZE(ReleaseFileCount, ReleaseFileSz) +
+	   APT_CACHESIZE(PackageFileCount, PackageFileSz) +
+	   APT_CACHESIZE(VerFileCount, VerFileSz) +
+	   APT_CACHESIZE(DescFileCount, DescFileSz) +
+	   APT_CACHESIZE(ProvidesCount, ProvidesSz) +
+	   (2 * Cache->Head().GetHashTableSize() * sizeof(map_id_t));
    cout << _("Total space accounted for: ") << SizeToStr(Total) << endl;
 #undef APT_CACHESIZE