1  | 
     | 
     | 
    /*	$OpenBSD: heapsort.c,v 1.11 2017/05/20 12:48:56 millert Exp $ */  | 
    
    
    2  | 
     | 
     | 
    /*-  | 
    
    
    3  | 
     | 
     | 
     * Copyright (c) 1991, 1993  | 
    
    
    4  | 
     | 
     | 
     *	The Regents of the University of California.  All rights reserved.  | 
    
    
    5  | 
     | 
     | 
     *  | 
    
    
    6  | 
     | 
     | 
     * This code is derived from software contributed to Berkeley by  | 
    
    
    7  | 
     | 
     | 
     * Ronnie Kon at Mindcraft Inc., Kevin Lew and Elmer Yglesias.  | 
    
    
    8  | 
     | 
     | 
     *  | 
    
    
    9  | 
     | 
     | 
     * Redistribution and use in source and binary forms, with or without  | 
    
    
    10  | 
     | 
     | 
     * modification, are permitted provided that the following conditions  | 
    
    
    11  | 
     | 
     | 
     * are met:  | 
    
    
    12  | 
     | 
     | 
     * 1. Redistributions of source code must retain the above copyright  | 
    
    
    13  | 
     | 
     | 
     *    notice, this list of conditions and the following disclaimer.  | 
    
    
    14  | 
     | 
     | 
     * 2. Redistributions in binary form must reproduce the above copyright  | 
    
    
    15  | 
     | 
     | 
     *    notice, this list of conditions and the following disclaimer in the  | 
    
    
    16  | 
     | 
     | 
     *    documentation and/or other materials provided with the distribution.  | 
    
    
    17  | 
     | 
     | 
     * 3. Neither the name of the University nor the names of its contributors  | 
    
    
    18  | 
     | 
     | 
     *    may be used to endorse or promote products derived from this software  | 
    
    
    19  | 
     | 
     | 
     *    without specific prior written permission.  | 
    
    
    20  | 
     | 
     | 
     *  | 
    
    
    21  | 
     | 
     | 
     * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND  | 
    
    
    22  | 
     | 
     | 
     * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE  | 
    
    
    23  | 
     | 
     | 
     * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE  | 
    
    
    24  | 
     | 
     | 
     * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE  | 
    
    
    25  | 
     | 
     | 
     * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL  | 
    
    
    26  | 
     | 
     | 
     * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS  | 
    
    
    27  | 
     | 
     | 
     * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)  | 
    
    
    28  | 
     | 
     | 
     * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT  | 
    
    
    29  | 
     | 
     | 
     * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY  | 
    
    
    30  | 
     | 
     | 
     * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF  | 
    
    
    31  | 
     | 
     | 
     * SUCH DAMAGE.  | 
    
    
    32  | 
     | 
     | 
     */  | 
    
    
    33  | 
     | 
     | 
     | 
    
    
    34  | 
     | 
     | 
    #include <sys/types.h>  | 
    
    
    35  | 
     | 
     | 
    #include <errno.h>  | 
    
    
    36  | 
     | 
     | 
    #include <stdlib.h>  | 
    
    
    37  | 
     | 
     | 
     | 
    
    
    38  | 
     | 
     | 
    /*  | 
    
    
    39  | 
     | 
     | 
     * Swap two areas of size number of bytes.  Although qsort(3) permits random  | 
    
    
    40  | 
     | 
     | 
     * blocks of memory to be sorted, sorting pointers is almost certainly the  | 
    
    
    41  | 
     | 
     | 
     * common case (and, were it not, could easily be made so).  Regardless, it  | 
    
    
    42  | 
     | 
     | 
     * isn't worth optimizing; the SWAP's get sped up by the cache, and pointer  | 
    
    
    43  | 
     | 
     | 
     * arithmetic gets lost in the time required for comparison function calls.  | 
    
    
    44  | 
     | 
     | 
     */  | 
    
    
    45  | 
     | 
     | 
    #define	SWAP(a, b, count, size, tmp) { \ | 
    
    
    46  | 
     | 
     | 
    	count = size; \  | 
    
    
    47  | 
     | 
     | 
    	do { \ | 
    
    
    48  | 
     | 
     | 
    		tmp = *a; \  | 
    
    
    49  | 
     | 
     | 
    		*a++ = *b; \  | 
    
    
    50  | 
     | 
     | 
    		*b++ = tmp; \  | 
    
    
    51  | 
     | 
     | 
    	} while (--count); \  | 
    
    
    52  | 
     | 
     | 
    }  | 
    
    
    53  | 
     | 
     | 
     | 
    
    
    54  | 
     | 
     | 
    /* Copy one block of size size to another. */  | 
    
    
    55  | 
     | 
     | 
    #define COPY(a, b, count, size, tmp1, tmp2) { \ | 
    
    
    56  | 
     | 
     | 
    	count = size; \  | 
    
    
    57  | 
     | 
     | 
    	tmp1 = a; \  | 
    
    
    58  | 
     | 
     | 
    	tmp2 = b; \  | 
    
    
    59  | 
     | 
     | 
    	do { \ | 
    
    
    60  | 
     | 
     | 
    		*tmp1++ = *tmp2++; \  | 
    
    
    61  | 
     | 
     | 
    	} while (--count); \  | 
    
    
    62  | 
     | 
     | 
    }  | 
    
    
    63  | 
     | 
     | 
     | 
    
    
    64  | 
     | 
     | 
    /*  | 
    
    
    65  | 
     | 
     | 
     * Build the list into a heap, where a heap is defined such that for  | 
    
    
    66  | 
     | 
     | 
     * the records K1 ... KN, Kj/2 >= Kj for 1 <= j/2 <= j <= N.  | 
    
    
    67  | 
     | 
     | 
     *  | 
    
    
    68  | 
     | 
     | 
     * There are two cases.  If j == nmemb, select largest of Ki and Kj.  If  | 
    
    
    69  | 
     | 
     | 
     * j < nmemb, select largest of Ki, Kj and Kj+1.  | 
    
    
    70  | 
     | 
     | 
     */  | 
    
    
    71  | 
     | 
     | 
    #define CREATE(initval, nmemb, par_i, child_i, par, child, size, count, tmp) { \ | 
    
    
    72  | 
     | 
     | 
    	for (par_i = initval; (child_i = par_i * 2) <= nmemb; \  | 
    
    
    73  | 
     | 
     | 
    	    par_i = child_i) { \ | 
    
    
    74  | 
     | 
     | 
    		child = base + child_i * size; \  | 
    
    
    75  | 
     | 
     | 
    		if (child_i < nmemb && compar(child, child + size) < 0) { \ | 
    
    
    76  | 
     | 
     | 
    			child += size; \  | 
    
    
    77  | 
     | 
     | 
    			++child_i; \  | 
    
    
    78  | 
     | 
     | 
    		} \  | 
    
    
    79  | 
     | 
     | 
    		par = base + par_i * size; \  | 
    
    
    80  | 
     | 
     | 
    		if (compar(child, par) <= 0) \  | 
    
    
    81  | 
     | 
     | 
    			break; \  | 
    
    
    82  | 
     | 
     | 
    		SWAP(par, child, count, size, tmp); \  | 
    
    
    83  | 
     | 
     | 
    	} \  | 
    
    
    84  | 
     | 
     | 
    }  | 
    
    
    85  | 
     | 
     | 
     | 
    
    
    86  | 
     | 
     | 
    /*  | 
    
    
    87  | 
     | 
     | 
     * Select the top of the heap and 'heapify'.  Since by far the most expensive  | 
    
    
    88  | 
     | 
     | 
     * action is the call to the compar function, a considerable optimization  | 
    
    
    89  | 
     | 
     | 
     * in the average case can be achieved due to the fact that k, the displaced  | 
    
    
    90  | 
     | 
     | 
     * element, is usually quite small, so it would be preferable to first  | 
    
    
    91  | 
     | 
     | 
     * heapify, always maintaining the invariant that the larger child is copied  | 
    
    
    92  | 
     | 
     | 
     * over its parent's record.  | 
    
    
    93  | 
     | 
     | 
     *  | 
    
    
    94  | 
     | 
     | 
     * Then, starting from the *bottom* of the heap, finding k's correct place,  | 
    
    
    95  | 
     | 
     | 
     * again maintaining the invariant.  As a result of the invariant no element  | 
    
    
    96  | 
     | 
     | 
     * is 'lost' when k is assigned its correct place in the heap.  | 
    
    
    97  | 
     | 
     | 
     *  | 
    
    
    98  | 
     | 
     | 
     * The time savings from this optimization are on the order of 15-20% for the  | 
    
    
    99  | 
     | 
     | 
     * average case. See Knuth, Vol. 3, page 158, problem 18.  | 
    
    
    100  | 
     | 
     | 
     *  | 
    
    
    101  | 
     | 
     | 
     * XXX Don't break the #define SELECT line, below.  Reiser cpp gets upset.  | 
    
    
    102  | 
     | 
     | 
     */  | 
    
    
    103  | 
     | 
     | 
    #define SELECT(par_i, child_i, nmemb, par, child, size, k, count, tmp1, tmp2) { \ | 
    
    
    104  | 
     | 
     | 
    	for (par_i = 1; (child_i = par_i * 2) <= nmemb; par_i = child_i) { \ | 
    
    
    105  | 
     | 
     | 
    		child = base + child_i * size; \  | 
    
    
    106  | 
     | 
     | 
    		if (child_i < nmemb && compar(child, child + size) < 0) { \ | 
    
    
    107  | 
     | 
     | 
    			child += size; \  | 
    
    
    108  | 
     | 
     | 
    			++child_i; \  | 
    
    
    109  | 
     | 
     | 
    		} \  | 
    
    
    110  | 
     | 
     | 
    		par = base + par_i * size; \  | 
    
    
    111  | 
     | 
     | 
    		COPY(par, child, count, size, tmp1, tmp2); \  | 
    
    
    112  | 
     | 
     | 
    	} \  | 
    
    
    113  | 
     | 
     | 
    	for (;;) { \ | 
    
    
    114  | 
     | 
     | 
    		child_i = par_i; \  | 
    
    
    115  | 
     | 
     | 
    		par_i = child_i / 2; \  | 
    
    
    116  | 
     | 
     | 
    		child = base + child_i * size; \  | 
    
    
    117  | 
     | 
     | 
    		par = base + par_i * size; \  | 
    
    
    118  | 
     | 
     | 
    		if (child_i == 1 || compar(k, par) < 0) { \ | 
    
    
    119  | 
     | 
     | 
    			COPY(child, k, count, size, tmp1, tmp2); \  | 
    
    
    120  | 
     | 
     | 
    			break; \  | 
    
    
    121  | 
     | 
     | 
    		} \  | 
    
    
    122  | 
     | 
     | 
    		COPY(child, par, count, size, tmp1, tmp2); \  | 
    
    
    123  | 
     | 
     | 
    	} \  | 
    
    
    124  | 
     | 
     | 
    }  | 
    
    
    125  | 
     | 
     | 
     | 
    
    
    126  | 
     | 
     | 
    /*  | 
    
    
    127  | 
     | 
     | 
     * Heapsort -- Knuth, Vol. 3, page 145.  Runs in O (N lg N), both average  | 
    
    
    128  | 
     | 
     | 
     * and worst.  While heapsort is faster than the worst case of quicksort,  | 
    
    
    129  | 
     | 
     | 
     * the BSD quicksort does median selection so that the chance of finding  | 
    
    
    130  | 
     | 
     | 
     * a data set that will trigger the worst case is nonexistent.  Heapsort's  | 
    
    
    131  | 
     | 
     | 
     * only advantage over quicksort is that it requires little additional memory.  | 
    
    
    132  | 
     | 
     | 
     */  | 
    
    
    133  | 
     | 
     | 
    int  | 
    
    
    134  | 
     | 
     | 
    heapsort(void *vbase, size_t nmemb, size_t size,  | 
    
    
    135  | 
     | 
     | 
        int (*compar)(const void *, const void *))  | 
    
    
    136  | 
     | 
     | 
    { | 
    
    
    137  | 
     | 
     | 
    	size_t cnt, i, j, l;  | 
    
    
    138  | 
     | 
     | 
    	char tmp, *tmp1, *tmp2;  | 
    
    
    139  | 
     | 
     | 
    	char *base, *k, *p, *t;  | 
    
    
    140  | 
     | 
     | 
     | 
    
    
    141  | 
     | 
     | 
    	if (nmemb <= 1)  | 
    
    
    142  | 
     | 
     | 
    		return (0);  | 
    
    
    143  | 
     | 
     | 
     | 
    
    
    144  | 
     | 
     | 
    	if (!size) { | 
    
    
    145  | 
     | 
     | 
    		errno = EINVAL;  | 
    
    
    146  | 
     | 
     | 
    		return (-1);  | 
    
    
    147  | 
     | 
     | 
    	}  | 
    
    
    148  | 
     | 
     | 
     | 
    
    
    149  | 
     | 
     | 
    	if ((k = malloc(size)) == NULL)  | 
    
    
    150  | 
     | 
     | 
    		return (-1);  | 
    
    
    151  | 
     | 
     | 
     | 
    
    
    152  | 
     | 
     | 
    	/*  | 
    
    
    153  | 
     | 
     | 
    	 * Items are numbered from 1 to nmemb, so offset from size bytes  | 
    
    
    154  | 
     | 
     | 
    	 * below the starting address.  | 
    
    
    155  | 
     | 
     | 
    	 */  | 
    
    
    156  | 
     | 
     | 
    	base = (char *)vbase - size;  | 
    
    
    157  | 
     | 
     | 
     | 
    
    
    158  | 
     | 
     | 
    	for (l = nmemb / 2 + 1; --l;)  | 
    
    
    159  | 
     | 
     | 
    		CREATE(l, nmemb, i, j, t, p, size, cnt, tmp);  | 
    
    
    160  | 
     | 
     | 
     | 
    
    
    161  | 
     | 
     | 
    	/*  | 
    
    
    162  | 
     | 
     | 
    	 * For each element of the heap, save the largest element into its  | 
    
    
    163  | 
     | 
     | 
    	 * final slot, save the displaced element (k), then recreate the  | 
    
    
    164  | 
     | 
     | 
    	 * heap.  | 
    
    
    165  | 
     | 
     | 
    	 */  | 
    
    
    166  | 
     | 
     | 
    	while (nmemb > 1) { | 
    
    
    167  | 
     | 
     | 
    		COPY(k, base + nmemb * size, cnt, size, tmp1, tmp2);  | 
    
    
    168  | 
     | 
     | 
    		COPY(base + nmemb * size, base + size, cnt, size, tmp1, tmp2);  | 
    
    
    169  | 
     | 
     | 
    		--nmemb;  | 
    
    
    170  | 
     | 
     | 
    		SELECT(i, j, nmemb, t, p, size, k, cnt, tmp1, tmp2);  | 
    
    
    171  | 
     | 
     | 
    	}  | 
    
    
    172  | 
     | 
     | 
    	free(k);  | 
    
    
    173  | 
     | 
     | 
    	return (0);  | 
    
    
    174  | 
     | 
     | 
    }  | 
    
    
    175  | 
     | 
     | 
    DEF_WEAK(heapsort);  |