new layout of symop table to work around compiler alignment issues

include/cif++/Structure.hpp

@@ -67,7 +67,7 @@ class Atom
 	Atom(const Atom& rhs);
 
 	// a special constructor to create symmetry copies
-	Atom(const Atom& rhs, const Point& symmmetry_location);
+	Atom(const Atom& rhs, const Point& symmmetry_location, const std::string& symmetry_operation);
 
 	~Atom();
 
@@ -89,8 +89,8 @@ class Atom
 	const cif::Row getRowAniso() const;
 
 	// Atom symmetryCopy(const Point& d, const clipper::RTop_orth& rt);
-	// bool isSymmetryCopy() const;
-	// std::string symmetry() const;
+	bool isSymmetryCopy() const;
+	std::string symmetry() const;
 	// const clipper::RTop_orth& symop() const;
 	
 	const Compound& comp() const;

include/cif++/Symmetry.hpp

@@ -26,9 +26,89 @@
 
 #pragma once
 
+#include <cstdint>
+#include <array>
+
 namespace mmcif
 {
 
+// --------------------------------------------------------------------
+
+struct Spacegroup
+{
+	const char* name;
+	const char* xHM;
+	const char* Hall;
+	int nr;
+};
+
+extern const Spacegroup kSpaceGroups[];
+extern const std::size_t kNrOfSpaceGroups;
+
+// --------------------------------------------------------------------
+
+struct SymopData
+{
+	constexpr SymopData(const std::array<int,15>& data)
+		: m_packed((static_cast<uint64_t>(data[ 0]) & 0x03) << 34 bitor
+				   (static_cast<uint64_t>(data[ 1]) & 0x03) << 32 bitor
+				   (static_cast<uint64_t>(data[ 2]) & 0x03) << 30 bitor
+				   (static_cast<uint64_t>(data[ 3]) & 0x03) << 28 bitor
+				   (static_cast<uint64_t>(data[ 4]) & 0x03) << 26 bitor
+				   (static_cast<uint64_t>(data[ 5]) & 0x03) << 24 bitor
+				   (static_cast<uint64_t>(data[ 6]) & 0x03) << 22 bitor
+				   (static_cast<uint64_t>(data[ 7]) & 0x03) << 20 bitor
+				   (static_cast<uint64_t>(data[ 8]) & 0x03) << 18 bitor
+				   (static_cast<uint64_t>(data[ 9]) & 0x07) << 15 bitor
+				   (static_cast<uint64_t>(data[10]) & 0x07) << 12 bitor
+				   (static_cast<uint64_t>(data[11]) & 0x07) <<  9 bitor
+				   (static_cast<uint64_t>(data[12]) & 0x07) <<  6 bitor
+				   (static_cast<uint64_t>(data[13]) & 0x07) <<  3 bitor
+				   (static_cast<uint64_t>(data[14]) & 0x07) <<  0)
+	{
+	}
+
+	bool operator==(const SymopData& rhs) const
+	{
+		return m_packed == rhs.m_packed;
+	}
+
+  private:
+
+	friend struct SymopDataBlock;
+
+	const uint64_t kPackMask = (~0UL >> (64-36));
+
+	SymopData(uint64_t v)
+		: m_packed(v bitand kPackMask) {}
+
+	uint64_t m_packed;
+};
+
+struct SymopDataBlock
+{
+	constexpr SymopDataBlock(int spacegroup, int rotational_number, const std::array<int,15>& rt_data)
+		: m_v((spacegroup & 0xffffULL) << 48 bitor
+			  (rotational_number & 0xffULL) << 40 bitor
+			  SymopData(rt_data).m_packed)
+	{
+	}
+
+	uint16_t spacegroup() const			{ return m_v >> 48; }
+	SymopData symop() const				{ return SymopData(m_v); }
+	uint8_t rotational_number() const	{ return (m_v >> 40) bitand 0xff; }
+
+  private:
+	uint64_t m_v;
+};
+
+static_assert(sizeof(SymopDataBlock) == sizeof(uint64_t), "Size of SymopData is wrong");
+
+extern const SymopDataBlock kSymopNrTable[];
+extern const std::size_t kSymopNrTableSize;
+
+// --------------------------------------------------------------------
+
 int GetSpacegroupNumber(std::string spacegroup);	// alternative for clipper's parsing code
 
 }

src/PDB2Cif.cpp

@@ -1229,6 +1229,9 @@ void PDBFileParser::PreParseInput(std::istream& is)
 	{
 		std::cerr << "Dropped unsupported records: " << ba::join(dropped, ", ") << std::endl;
 	}
+
+	if (mData == nullptr)
+		throw std::runtime_error("Empty file?");
 	
 	mRec = mData;
 }
@@ -1247,6 +1250,7 @@ void PDBFileParser::GetNextRecord()
 
 void PDBFileParser::Match(const std::string& expected, bool throwIfMissing)
 {
+	assert(mRec);
 	if (mRec->mName != expected)
 	{
 		if (throwIfMissing)

src/Structure.cpp

@@ -216,13 +216,13 @@ struct AtomImpl
 	// 	mLocation -= d;
 	// }
 
-	AtomImpl(const AtomImpl& impl, const Point& loc)
+	AtomImpl(const AtomImpl& impl, const Point& loc, const std::string& sym_op)
 		: mFile(impl.mFile), mID(impl.mID), mType(impl.mType), mAtomID(impl.mAtomID)
 		, mCompID(impl.mCompID), mAsymID(impl.mAsymID), mSeqID(impl.mSeqID)
 		, mAltID(impl.mAltID), mLocation(loc), mRefcount(1)
 		, mRow(impl.mRow), mCompound(impl.mCompound), mRadius(impl.mRadius)
 		, mCachedProperties(impl.mCachedProperties)
-		, mSymmetryCopy(true)
+		, mSymmetryCopy(true), mSymmetryOperator(sym_op)
 	{
 	}
 
@@ -378,6 +378,7 @@ struct AtomImpl
 	bool				mSymmetryCopy = false;
 	bool				mClone = false;
 
+	std::string			mSymmetryOperator;
 	// clipper::RTop_orth	mRTop;
 	// Point				mD;
 	// int32_t				mRTix;
@@ -418,8 +419,8 @@ Atom Atom::clone() const
 	return Atom(mImpl_ ? new AtomImpl(*mImpl_) : nullptr);
 }
 
-Atom::Atom(const Atom& rhs, const Point& loc)
-	: mImpl_(new AtomImpl(*rhs.mImpl_, loc))
+Atom::Atom(const Atom& rhs, const Point& loc, const std::string& sym_op)
+	: mImpl_(new AtomImpl(*rhs.mImpl_, loc, sym_op))
 {
 }
 
@@ -629,6 +630,16 @@ void Atom::location(Point p)
 // 	return clipper::Symop(impl()->mRTop).format() + "\n" + impl()->mRTop.format();
 // }
 
+bool Atom::isSymmetryCopy() const
+{
+	return mImpl_->mSymmetryCopy;
+}
+
+std::string Atom::symmetry() const
+{
+	return mImpl_->mSymmetryOperator;
+}
+
 // const clipper::RTop_orth& Atom::symop() const
 // {
 // 	return impl()->mRTop;

src/Symmetry.cpp

@@ -53,7 +53,7 @@ int GetSpacegroupNumber(std::string spacegroup)
 
 	int result = 0;
 
-	const size_t N = sizeof(kSpaceGroups) / sizeof(Spacegroup);
+	const size_t N = kNrOfSpaceGroups;
 	int32_t L = 0, R = static_cast<int32_t>(N - 1);
 	while (L <= R)
 	{
@@ -75,8 +75,9 @@ int GetSpacegroupNumber(std::string spacegroup)
 	// not found, see if we can find a match based on xHM name
 	if (result == 0)
 	{
-		for (auto& sp: kSpaceGroups)
+		for (size_t i = 0; i < kNrOfSpaceGroups; ++i)
 		{
+			auto& sp = kSpaceGroups[i];
 			if (sp.xHM == spacegroup)
 			{
 				result = sp.nr;

tools/symop-map-generator.cpp

@@ -350,13 +350,9 @@ int main()
 // and $CLIBD/syminfo.lib using symop-map-generator,
 // part of the PDB-REDO suite of programs.
 
-struct Spacegroup
-{
-	const char* name;
-	const char* xHM;
-	const char* Hall;
-	int nr;
-} kSpaceGroups[] =
+#include "cif++/Symmetry.hpp"
+
+const Spacegroup kSpaceGroups[] =
 {
 )";
 
@@ -390,35 +386,9 @@ struct Spacegroup
 cout << R"(
 };
 
-union SymopData
-{
-	struct
-	{
-		int rot_0_0:2;
-		int rot_0_1:2;
-		int rot_0_2:2;
-		int rot_1_0:2;
-		int rot_1_1:2;
-		int rot_1_2:2;
-		int rot_2_0:2;
-		int rot_2_1:2;
-		int rot_2_2:2;
-		unsigned int trn_0_0:3;
-		unsigned int trn_0_1:3;
-		unsigned int trn_1_0:3;
-		unsigned int trn_1_1:3;
-		unsigned int trn_2_0:3;
-		unsigned int trn_2_1:3;
-	};
-	uint64_t iv:36;
-};
+const size_t kNrOfSpaceGroups = sizeof(kSpaceGroups) / sizeof(Spacegroup);
 
-struct SymopDataBlock
-{
-	uint16_t spacegroupNr;
-	uint8_t rotationalNr;
-	SymopData rt;
-} kSymopNrTable[] = {
+const SymopDataBlock kSymopNrTable[] = {
 )" << endl;
 
 		int spacegroupNr = 0;
@@ -432,13 +402,17 @@ struct SymopDataBlock
 			spacegroupNr = sp;
 
 			cout << "    { " << setw(3) << sp
-					<< ", " << setw(3) << o << ", { { ";
+					<< ", " << setw(3) << o << ", { ";
 			for (auto i: get<2>(sd))
 				cout << setw(2) << i << ',';
-			cout << " } } }," << endl;
+			cout << " } }," << endl;
 		}
 
-		cout << "};" << endl;
+		cout << R"(};
+
+const size_t kSymopNrTableSize = sizeof(kSymopNrTable) / sizeof(SymopDataBlock);
+
+)" << endl;
 	}
 	catch (const exception& ex)
 	{