Skip to content

K
kicad-source-mirror
Commit b1d56ec1 authored by Vladimir Ur's avatar Vladimir Ur
Browse files
Options

Just a new draft of length units

parent 726d772c
Hide whitespace changes
Inline Side-by-side
Showing with 243 additions and 153 deletions
include/length.h
View file @ b1d56ec1
......@@ -3,148 +3,237 @@
* @file length.h
*/
/* sorry it is not styles correctly, i'll work on it further */
/* sorry it is not styled correctly, i'll work on it further */
#ifndef LENGTH_H_INCLUDED
#define LENGTH_H_INCLUDED 1
/* type to be used by length units by default */
typedef int DEF_LENGTH_VALUE;
template <typename T = DEF_LENGTH_VALUE, int P = 1> class LENGTH;
/**
* Length template class
* @param T actual type holding a value (be aware of precision and range!)
* @param P power of length unit: 1 - length, 2 - area, 3 - volume, -1 - lin. density etc...
* This class check length dimension in compile time. In runtime it behaves
* exactly like contained type t (which should be numeric type, like int or double)
* This class can be replaced with its contained type or simple stub.
* Check rules:
* - comparisons (< = etc.), addition, subtraction require values of same dimension
* e. g. length with length, area with area etc.
* - multiplication and division result have appropriate dimension (powers
* added and subtracted respectively)
* - sqrt and cbrt have appropriate dimensions (P/2 and P/3).
* Limitations:
* - functions which should not be applied to dimensioned values are not implemeted:
* they include algebraic (exp, log...), trigo (sin, cos...), hyperbolic (sinh, cosh..)
* - pow function is not implemented as it is require dimension check in runtime
* you should use multiplication, division, sqrt and cbrt functions instead.
* - sqrt and cbrt result type should be instantiated before they used
* Be aware when using them in complex formulae, e. g.
* LENGTH< double, 1 > len = cbrt(vol) - is ok, but
* LENGTH< double, 2 > vol = sqrt(area*area*area*area)/length - will fail
* if LENGTH<..., 4> is not instantiated
* - non-integer power values do not supported
* they should be implemented carefully using natural fractions, not floats, to be exact
* but they are very rare so you should not worry about.
* e. g. linear electric noise density should be in mV/sqrt(m)
* - automatic numeric type casts are not performed. You even have to manually
* cast LENGTH< short > to LENGTH< int > or LENGTH< float >
* to LENGTH< double >. Anyway it is not such trouble as progremmer should be
* very careful when mixing numeric types and avoid automatic casts.
*
*/
template < typename T = DEF_LENGTH_VALUE, int P = 1 > class LENGTH;
/**
* Length units contained in this class
*/
template <typename T> class LENGTH_UNITS;
template <typename T, int P> struct LENGTH_TRAITS {
typedef LENGTH<T, P> flat;
/**
* For internal needs
*/
template < typename T, int P > struct LENGTH_TRAITS
{
typedef LENGTH<T, P> flat;
};
template <typename T> struct LENGTH_TRAITS<T, 0> {
typedef T flat;
template < typename T > struct LENGTH_TRAITS< T, 0 >
{
/* length dimension to power 0 is just a number, so LENGTH<T, 0> should be automatically converted to T */
typedef T flat;
};
template<typename T, int P> class LENGTH {
friend class LENGTH_UNITS<T>;
friend class LENGTH_TRAITS<T, P>;
template <typename Y, int R> friend class LENGTH;
template< typename T, int P > class LENGTH
{
friend class LENGTH_UNITS< T >;
friend class LENGTH_TRAITS< T, P >;
template < typename Y, int R > friend class LENGTH;
protected:
T m_U;
LENGTH(T units) : m_U(units) {
}
static T RawValue(const LENGTH<T, P> &x) {
return x.m_U;
}
static T RawValue(const T& x) {
return x;
}
T m_U;
LENGTH( T units ) : m_U( units )
{
}
static T RawValue( const LENGTH<T, P> &x )
{
return x.m_U;
}
static T RawValue( const T& x )
{
return x;
}
public:
typedef LENGTH<T, P> flat;
typedef T value_type;
enum {
dimension = P
};
LENGTH(const LENGTH <T, P> &orig) : m_U(orig.m_U) {
}
LENGTH( void ) : m_U() {
}
static LENGTH<T, P> zero (void) {
return T(0);
}
LENGTH<T, P> & operator = (const LENGTH<T, P> & y) {
this->m_U = y.m_U;
return *this;
}
template<typename Y>
operator LENGTH<Y, P> (void) {
return this->m_U;
}
/*************************/
/* comparisons and tests */
/*************************/
bool operator ==(const LENGTH <T, P> y) const {
return m_U == y.m_U;
}
bool operator !=(const LENGTH <T, P> y) const {
return m_U != y.m_U;
}
bool operator <(const LENGTH <T, P> y) const {
return m_U < y.m_U;
}
bool operator >=(const LENGTH <T, P> y) const {
return m_U >= y.m_U;
}
bool operator >(const LENGTH <T, P> y) const {
return m_U > y.m_U;
}
bool operator <=(const LENGTH <T, P> y) const {
return m_U <= y.m_U;
}
bool operator !( void ) const {
return !m_U;
}
/*************************/
/* basic arithmetic */
/*************************/
LENGTH<T, P> operator - (void) const {
return LENGTH<T, P>(-this->m_U);
}
LENGTH<T, P> operator - (const LENGTH<T, P> y) const {
return m_U - y.m_U;
}
LENGTH<T, P> operator + (const LENGTH<T, P> y) const {
return m_U + y.m_U;
}
template <int R>
typename LENGTH_TRAITS<T, P + R>::flat operator * (const LENGTH<T, R> &y) const {
return m_U * y.m_U;
}
LENGTH<T, P> operator * (const T &y) const {
return m_U * y;
}
LENGTH<T, P> friend operator * (const T &y, const LENGTH<T, P> &x) {
return x.m_U * y;
}
template <int R>
typename LENGTH_TRAITS<T, P - R>::flat operator / (const LENGTH<T, R> &y) const {
return m_U / y.m_U;
}
LENGTH<T, P> operator / (const T &y) const {
return m_U / y;
}
LENGTH<T, -P> friend operator / (const T &y, const LENGTH<T, P> &x) {
return y / x.m_U;
}
friend LENGTH<T, P> sqrt(LENGTH<T, P*2> y) {
return sqrt(y.m_U);
}
friend LENGTH<T, P> cbrt(LENGTH<T, P*3> y) {
return cbrt(y.m_U);
}
/*************************/
/* assignment arithmetic */
/*************************/
LENGTH<T, P>& operator -= (const LENGTH<T, P> y) {
return m_U -= y.m_U;
}
LENGTH<T, P>& operator += (const LENGTH<T, P> y) {
return m_U += y.m_U;
}
LENGTH<T, P>& operator *= (const T y) {
return m_U *= y;
}
LENGTH<T, P>& operator /= (const T y) {
return m_U /= y;
}
/*************************/
/* more arithmetic */
/*************************/
typedef T value_type;
enum
{
dimension = P
};
LENGTH( const LENGTH <T, P> &orig ) : m_U( orig.m_U )
{
}
LENGTH( void ) : m_U()
{
}
static LENGTH<T, P> zero ( void )
{
return T(0);
}
LENGTH<T, P> & operator = ( const LENGTH<T, P> & y )
{
this->m_U = y.m_U;
return *this;
}
template<typename Y> operator LENGTH< Y, P > ( void )
{
return this->m_U;
}
/*************************/
/* comparisons and tests */
/*************************/
bool operator ==( const LENGTH < T, P > y ) const
{
return m_U == y.m_U;
}
bool operator !=( const LENGTH < T, P > y ) const
{
return m_U != y.m_U;
}
bool operator <( const LENGTH < T, P > y ) const
{
return m_U < y.m_U;
}
bool operator >=( const LENGTH < T, P > y ) const
{
return m_U >= y.m_U;
}
bool operator >( const LENGTH < T, P > y ) const
{
return m_U > y.m_U;
}
bool operator <=( const LENGTH < T, P > y ) const
{
return m_U <= y.m_U;
}
bool operator !( void ) const
{
return !m_U;
}
/*************************/
/* basic arithmetic */
/*************************/
LENGTH< T, P > operator - ( void ) const
{
return LENGTH<T, P>(-this->m_U);
}
LENGTH< T, P > operator - ( const LENGTH< T, P > y ) const
{
return m_U - y.m_U;
}
LENGTH< T, P > operator + ( const LENGTH< T, P > y ) const
{
return m_U + y.m_U;
}
template < int R >
typename LENGTH_TRAITS< T, P + R >::flat operator * ( const LENGTH<T, R> &y ) const
{
return m_U * y.m_U;
}
LENGTH< T, P > operator * ( const T & y) const
{
return m_U * y;
}
LENGTH< T, P > friend operator * ( const T &y, const LENGTH<T, P> &x )
{
return x.m_U * y;
}
template < int R >
typename LENGTH_TRAITS< T, P - R >::flat operator / ( const LENGTH<T, R> &y ) const
{
return m_U / y.m_U;
}
LENGTH< T, P > operator / ( const T &y ) const
{
return m_U / y;
}
LENGTH< T, -P > friend operator / ( const T &y, const LENGTH< T, P > &x )
{
return y / x.m_U;
}
friend LENGTH< T, P > sqrt( LENGTH< T, P*2 > y )
{
return sqrt( y.m_U );
}
friend LENGTH< T, P > cbrt( LENGTH< T, P*3 > y )
{
return cbrt( y.m_U );
}
/*************************/
/* assignment arithmetic */
/*************************/
LENGTH< T, P >& operator -= ( const LENGTH< T, P > y )
{
return m_U -= y.m_U;
}
LENGTH< T, P >& operator += ( const LENGTH< T, P > y )
{
return m_U += y.m_U;
}
LENGTH< T, P >& operator *= ( const T y )
{
return m_U *= y;
}
LENGTH< T, P >& operator /= ( const T y )
{
return m_U /= y;
}
/*************************/
/* more arithmetic */
/*************************/
};
template <typename T = DEF_LENGTH_VALUE> class LENGTH_UNITS {
/**
* Units of length
*
* How to use them:
* there are several functions, named LENGTH_UNITS< T >::METRE, which return
* named unit (1 meter in example) which have type LENGTH< T, P >.
* to get specific length you should use a multiplication:
* 3*LENGTH_UNITS::metre() gives 3 metres
* 0.01*LENGTH_UNITS::metre() gives 0.01 inch
* to get numeric value of length in specific units you should use a division
* length/LENGTH_UNITS::metre() gives number of metres in length
* legnth/LENGTH_UNITS::foot() gives number of feet in length
*/
template < typename T = DEF_LENGTH_VALUE > class LENGTH_UNITS {
protected:
enum
{
......@@ -152,37 +241,38 @@ protected:
INCH = METRE / 10000 * 254
};
public:
static LENGTH<T, 1> metre( void ) {
return T(METRE);
}
static LENGTH<T, 1> decimetre( void ) {
return T(METRE / 10);
}
static LENGTH<T, 1> centimetre( void ) {
return T(METRE / 100);
}
static LENGTH<T, 1> millimetre( void ) {
return T(METRE / 1000);
}
static LENGTH<T, 1> micrometre( void ) {
return T(METRE / 1000000);
}
static LENGTH<T, 1> foot( void ) { /* do not think this will ever need */
return T(INCH * 12);
}
static LENGTH<T, 1> inch( void ) {
return T(INCH);
}
static LENGTH<T, 1> mil( void ) {
return T(INCH / 1000);
}
public:
static LENGTH< T, 1 > metre( void ) {
return T( METRE );
}
static LENGTH< T, 1 > decimetre( void ) {
return T( METRE / 10 );
}
static LENGTH< T, 1 > centimetre( void ) {
return T( METRE / 100 );
}
static LENGTH< T, 1 > millimetre( void ) {
return T( METRE / 1000 );
}
static LENGTH< T, 1 > micrometre( void ) {
return T( METRE / 1000000 );
}
static LENGTH< T, 1 > foot( void ) { /* do not think this will ever need */
return T( INCH * 12 );
}
static LENGTH< T, 1 > inch( void ) {
return T( INCH );
}
static LENGTH< T, 1 > mil( void ) {
return T( INCH / 1000 );
}
};
/* shortcut */
template <typename T, int D> class LENGTH_UNITS<LENGTH<T, D> >: public LENGTH_UNITS<T> {
/**
* shortcut to get units of given length type
*/
template < typename T, int D > class LENGTH_UNITS< LENGTH< T, D > >: public LENGTH_UNITS< T >
{
};
/* TODO: argument promotion (but is this need? explicit casts would be enough) */
#endif
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment