expairseq.h File Reference

Interface to sequences of expression pairs. More...

#include "expair.h"
#include "indexed.h"
#include <algorithm>
#include <list>
#include <memory>
#include <vector>

Go to the source code of this file.

Classes
class	GiNaC::expairseq
	A sequence of class expair. More...
class	GiNaC::make_flat_inserter
	Class to handle the renaming of dummy indices. More...
Namespaces
namespace	GiNaC
	CRC32 hash function.
Defines
#define	EXPAIRSEQ_USE_HASHTAB   0
	Using hash tables can potentially enhance the asymptotic behaviour of combining n terms into one large sum (or n terms into one large product) from O(n*log(n)) to about O(n).
Typedefs
typedef std::vector< expair >	GiNaC::epvector
	expair-vector
typedef epvector::iterator	GiNaC::epp
	expair-vector pointer
typedef std::list< epp >	GiNaC::epplist
	list of expair-vector pointers
typedef std::vector< epplist >	GiNaC::epplistvector
	vector of epplist
Functions
epvector *	GiNaC::conjugateepvector (const epvector &)
	Complex conjugate every element of an epvector.

---

Detailed Description

Interface to sequences of expression pairs.

Definition in file expairseq.h.

---

Define Documentation

#define EXPAIRSEQ_USE_HASHTAB   0

Using hash tables can potentially enhance the asymptotic behaviour of combining n terms into one large sum (or n terms into one large product) from O(n*log(n)) to about O(n).

There are, however, several drawbacks. The constant in front of O(n) is quite large, when copying such an object one also has to copy the has table, comparison is quite expensive because there is no ordering any more, it doesn't help at all when combining two expairseqs because due to the presorted nature the behaviour would be O(n) anyways, the code is quite messy, etc, etc. The code is here as an example for following generations to tinker with.

Definition at line 46 of file expairseq.h.