/* Copyright (C) 1989, 2000 artofcode LLC.  All rights reserved.
  
  This program is free software; you can redistribute it and/or modify it
  under the terms of the GNU General Public License as published by the
  Free Software Foundation; either version 2 of the License, or (at your
  option) any later version.

  This program is distributed in the hope that it will be useful, but
  WITHOUT ANY WARRANTY; without even the implied warranty of
  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  General Public License for more details.

  You should have received a copy of the GNU General Public License along
  with this program; if not, write to the Free Software Foundation, Inc.,
  59 Temple Place, Suite 330, Boston, MA, 02111-1307.

*/

/*$Id: gxmatrix.h,v 1.4.6.1.2.1 2003/01/17 00:49:04 giles Exp $ */
/* Internal matrix routines for Ghostscript library */

#ifndef gxmatrix_INCLUDED
#  define gxmatrix_INCLUDED

#include "gsmatrix.h"

/*
 * Define a matrix with a cached fixed-point copy of the translation.
 * This is only used by a few routines in gscoord.c; they are responsible
 * for ensuring the validity of the cache.  Note that the floating point
 * tx/ty values may be too large to fit in a fixed values; txy_fixed_valid
 * is false if this is the case, and true otherwise.
 */
typedef struct gs_matrix_fixed_s {
    _matrix_body;
    fixed tx_fixed, ty_fixed;
    bool txy_fixed_valid;
} gs_matrix_fixed;

/* Make a gs_matrix_fixed from a gs_matrix. */
int gs_matrix_fixed_from_matrix(P2(gs_matrix_fixed *, const gs_matrix *));

/* Coordinate transformations to fixed point. */
int gs_point_transform2fixed(P4(const gs_matrix_fixed *, floatp, floatp,
				gs_fixed_point *));
int gs_distance_transform2fixed(P4(const gs_matrix_fixed *, floatp, floatp,
				   gs_fixed_point *));

/*
 * Define the fixed-point coefficient structure for avoiding
 * floating point in coordinate transformations.
 * Currently this is used only by the Type 1 font interpreter.
 * The setup is in gscoord.c.
 */
typedef struct {
    long xx, xy, yx, yy;
    int skewed;
    int shift;			/* see m_fixed */
    int max_bits;		/* max bits of coefficient */
    fixed round;		/* ditto */
} fixed_coeff;

/*
 * Multiply a fixed point value by a coefficient.  The coefficient has two
 * parts: a value (long) and a shift factor (int), The result is (fixed *
 * coef_value + round_value) >> (shift + _fixed_shift)) where the shift
 * factor and the round value are picked from the fixed_coeff structure, and
 * the coefficient value (from one of the coeff1 members) is passed
 * explicitly.  The last parameter specifies the number of bits available to
 * prevent overflow for integer arithmetic.  (This is a very custom
 * routine.)  The intermediate value may exceed the size of a long integer.
 */
fixed fixed_coeff_mult(P4(fixed, long, const fixed_coeff *, int));

/*
 * Multiply a fixed whose integer part usually does not exceed max_bits
 * in magnitude by a coefficient from a fixed_coeff.
 * We can use a faster algorithm if the fixed is an integer within
 * a range that doesn't cause the multiplication to overflow an int.
 */
#define m_fixed(v, c, fc, maxb)\
  (((v) + (fixed_1 << (maxb - 1))) &\
    ((-fixed_1 << maxb) | _fixed_fraction_v) ?	/* out of range, or has fraction */\
    fixed_coeff_mult((v), (fc).c, &(fc), maxb) : \
   arith_rshift(fixed2int_var(v) * (fc).c + (fc).round, (fc).shift))

#endif /* gxmatrix_INCLUDED */