view lib-src/qsort.c @ 814:a634e3b7acc8

[xemacs-hg @ 2002-04-14 12:41:59 by ben] latest changes TODO.ben-mule-21-5: Update. make-docfile.c: Add basic support for handling ISO 2022 doc strings -- we parse the basic charset designation sequences so we know whether we're in ASCII and have to pay attention to end quotes and such. Reformat code according to coding standards. abbrev.el: Add `global-abbrev-mode', which turns on or off abbrev-mode in all buffers. Added `defining-abbrev-turns-on-abbrev-mode' -- if non-nil, defining an abbrev through an interactive function will automatically turn on abbrev-mode, either globally or locally depending on the command. This is the "what you'd expect" behavior. indent.el: general function for indenting a balanced expression in a mode-correct way. Works similar to indent-region in that a mode can specify a specific command to do the whole operation; if not, figure out the region using forward-sexp and indent each line using indent-according-to-mode. keydefs.el: Removed. Modify M-C-backslash to do indent-region-or-balanced-expression. Make S-Tab just insert a TAB char, like it's meant to do. make-docfile.el: Now that we're using the call-process-in-lisp, we need to load an extra file win32-native.el because we're running a bare temacs. menubar-items.el: Totally redo the Cmds menu so that most used commands appear directly on the menu and less used commands appear in submenus. The old way may have been very pretty, but rather impractical. process.el: Under Windows, don't ever use old-call-process-internal, even in batch mode. We can do processes in batch mode. subr.el: Someone recoded truncate-string-to-width, saying "the FSF version is too complicated and does lots of hard-to-understand stuff" but the resulting recoded version was *totally* wrong! it misunderstood the basic point of this function, which is work in *columns* not chars. i dumped ours and copied the version from FSF 21.1. Also added truncate-string-with-continuation-dots, since this idiom is used often. config.inc.samp, xemacs.mak: Separate out debug and optimize flags. Remove all vestiges of USE_MINIMAL_TAGBITS, USE_INDEXED_LRECORD_IMPLEMENTATION, and GUNG_HO, since those ifdefs have long been removed. Make error-checking support actually work. Some rearrangement of config.inc.samp to make it more logical. Remove callproc.c and ntproc.c from xemacs.mak, no longer used. Make pdump the default. lisp.h: Add support for strong type-checking of Bytecount, Bytebpos, Charcount, Charbpos, and others, by making them classes, overloading the operators to provide integer-like operation and carefully controlling what operations are allowed. Not currently enabled in C++ builds because there are still a number of compile errors, and it won't really work till we merge in my "8-bit-Mule" workspace, in which I make use of the new types Charxpos, Bytexpos, Memxpos, representing a "position" either in a buffer or a string. (This is especially important in the extent code.) abbrev.c, alloc.c, eval.c, buffer.c, buffer.h, editfns.c, fns.c, text.h: Warning fixes, some of them related to new C++ strict type checking of Bytecount, Charbpos, etc. dired.c: Caught an actual error due to strong type checking -- char len being passed when should be byte len. alloc.c, backtrace.h, bytecode.c, bytecode.h, eval.c, sysdep.c: Further optimize Ffuncall: -- process arg list at compiled-function creation time, converting into an array for extra-quick access at funcall time. -- rewrite funcall_compiled_function to use it, and inline this function. -- change the order of check for magic stuff in SPECBIND_FAST_UNSAFE to be faster. -- move the check for need to garbage collect into the allocation code, so only a single flag needs to be checked in funcall. buffer.c, symbols.c: add debug funs to check on mule optimization info in buffers and strings. eval.c, emacs.c, text.c, regex.c, scrollbar-msw.c, search.c: Fix evil crashes due to eistrings not properly reinitialized under pdump. Redo a bit some of the init routines; convert some complex_vars_of() into simple vars_of(), because they didn't need complex processing. callproc.c, emacs.c, event-stream.c, nt.c, process.c, process.h, sysdep.c, sysdep.h, syssignal.h, syswindows.h, ntproc.c: Delete. Hallelujah, praise the Lord, there is no god but Allah!!! fix so that processes can be invoked in bare temacs -- thereby eliminating any need for callproc.c. (currently only eliminated under NT.) remove all crufty and unnecessary old process code in ntproc.c and elsewhere. move non-callproc-specific stuff (mostly environment) into process.c, so callproc.c can be left out under NT. console-tty.c, doc.c, file-coding.c, file-coding.h, lstream.c, lstream.h: fix doc string handling so it works with Japanese, etc docs. change handling of "character mode" so callers don't have to manually set it (quite error-prone). event-msw.c: spacing fixes. lread.c: eliminate unused crufty vintage-19 "FSF defun hack" code. lrecord.h: improve pdump description docs. buffer.c, ntheap.c, unexnt.c, win32.c, emacs.c: Mule-ize some unexec and startup code. It was pseudo-Mule-ized before by simply always calling the ...A versions of functions, but that won't cut it -- eventually we want to be able to run properly even if XEmacs has been installed in a Japanese directory. (The current problem is the timing of the loading of the Unicode tables; this will eventually be fixed.) Go through and fix various other places where the code was not Mule-clean. Provide a function mswindows_get_module_file_name() to get our own name without resort to PATH_MAX and such. Add a big comment in main() about the problem with Unicode table load timing that I just alluded to. emacs.c: When error-checking is enabled (interpreted as "user is developing XEmacs"), don't ask user to "pause to read messages" when a fatal error has occurred, because it will wedge if we are in an inner modal loop (typically when a menu is popped up) and make us unable to get a useful stack trace in the debugger. text.c: Correct update_entirely_ascii_p_flag to actually work. lisp.h, symsinit.h: declarations for above changes.
author ben
date Sun, 14 Apr 2002 12:43:31 +0000
parents 576fb035e263
children 061f4f90f874
line wrap: on
line source

/* Plug-compatible replacement for UNIX qsort.
   Copyright (C) 1989 Free Software Foundation, Inc.
   Written by Douglas C. Schmidt (schmidt@ics.uci.edu)

This file is part of GNU CC.

GNU QSORT 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, or (at your option)
any later version.

GNU QSORT 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 GNU QSORT; see the file COPYING.  If not, write to
the Free the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
Boston, MA 02111-1307, USA. */

/* Synched up with: FSF 19.28. */

#ifdef sparc
#include <alloca.h>
#endif

/* Invoke the comparison function, returns either 0, < 0, or > 0. */
#define CMP(A,B) ((*cmp)((A),(B)))

/* Byte-wise swap two items of size SIZE. */
#define SWAP(A,B,SIZE) do {int sz = (SIZE); char *a = (A); char *b = (B); \
    do { char _temp = *a;*a++ = *b;*b++ = _temp;} while (--sz);} while (0)

/* Copy SIZE bytes from item B to item A. */
#define COPY(A,B,SIZE) {int sz = (SIZE); do { *(A)++ = *(B)++; } while (--sz); }

/* This should be replaced by a standard ANSI macro. */
#define BYTES_PER_WORD 8

/* The next 4 #defines implement a very fast in-line stack abstraction. */
#define STACK_SIZE (BYTES_PER_WORD * sizeof (long))
#define PUSH(LOW,HIGH) do {top->lo = LOW;top++->hi = HIGH;} while (0)
#define POP(LOW,HIGH)  do {LOW = (--top)->lo;HIGH = top->hi;} while (0)
#define STACK_NOT_EMPTY (stack < top)                

/* Discontinue quicksort algorithm when partition gets below this size.
   This particular magic number was chosen to work best on a Sun 4/260. */
#define MAX_THRESH 4

/* Stack node declarations used to store unfulfilled partition obligations. */
typedef struct 
{
  char *lo;
  char *hi;
} stack_node;

/* Order size using quicksort.  This implementation incorporates
   four optimizations discussed in Sedgewick:
   
   1. Non-recursive, using an explicit stack of pointer that store the 
      next array partition to sort.  To save time, this maximum amount 
      of space required to store an array of MAX_INT is allocated on the 
      stack.  Assuming a 32-bit integer, this needs only 32 * 
      sizeof (stack_node) == 136 bits.  Pretty cheap, actually.

   2. Choose the pivot element using a median-of-three decision tree.
      This reduces the probability of selecting a bad pivot value and 
      eliminates certain extraneous comparisons.

   3. Only quicksorts TOTAL_ELEMS / MAX_THRESH partitions, leaving
      insertion sort to order the MAX_THRESH items within each partition.  
      This is a big win, since insertion sort is faster for small, mostly
      sorted array segments.
   
   4. The larger of the two sub-partitions is always pushed onto the
      stack first, with the algorithm then concentrating on the
      smaller partition.  This *guarantees* no more than log (n)
      stack size is needed (actually O(1) in this case)! */
      
int 
qsort (base_ptr, total_elems, size, cmp)
     char *base_ptr;
     int total_elems;
     int size;
     int (*cmp)();
{
  /* Allocating SIZE bytes for a pivot buffer facilitates a better 
     algorithm below since we can do comparisons directly on the pivot. */
  char *pivot_buffer = (char *) alloca (size);
  int   max_thresh   = MAX_THRESH * size;

  if (total_elems > MAX_THRESH)
    {
      char       *lo = base_ptr;
      char       *hi = lo + size * (total_elems - 1);
      stack_node stack[STACK_SIZE]; /* Largest size needed for 32-bit int!!! */
      stack_node *top = stack + 1;

      while (STACK_NOT_EMPTY)
        {
          char *left_ptr;
          char *right_ptr;
          {
            char *pivot = pivot_buffer;
            {
              /* Select median value from among LO, MID, and HI. Rearrange
                 LO and HI so the three values are sorted. This lowers the 
                 probability of picking a pathological pivot value and 
                 skips a comparison for both the LEFT_PTR and RIGHT_PTR. */

              char *mid = lo + size * ((hi - lo) / size >> 1);

              if (CMP (mid, lo) < 0)
                SWAP (mid, lo, size);
              if (CMP (hi, mid) < 0)
                SWAP (mid, hi, size);
              else 
                goto jump_over;
              if (CMP (mid, lo) < 0)
                SWAP (mid, lo, size);
            jump_over:
              COPY (pivot, mid, size);
              pivot = pivot_buffer;
            }
            left_ptr  = lo + size;
            right_ptr = hi - size; 

            /* Here's the famous ``collapse the walls'' section of quicksort.  
               Gotta like those tight inner loops!  They are the main reason 
               that this algorithm runs much faster than others. */
            do 
              {
                while (CMP (left_ptr, pivot) < 0)
                  left_ptr += size;

                while (CMP (pivot, right_ptr) < 0)
                  right_ptr -= size;

                if (left_ptr < right_ptr) 
                  {
                    SWAP (left_ptr, right_ptr, size);
                    left_ptr += size;
                    right_ptr -= size;
                  }
                else if (left_ptr == right_ptr) 
                  {
                    left_ptr += size;
                    right_ptr -= size;
                    break;
                  }
              } 
            while (left_ptr <= right_ptr);

          }

          /* Set up pointers for next iteration.  First determine whether
             left and right partitions are below the threshold size. If so, 
             ignore one or both.  Otherwise, push the larger partition's
             bounds on the stack and continue sorting the smaller one. */

          if ((right_ptr - lo) <= max_thresh)
            {
              if ((hi - left_ptr) <= max_thresh) /* Ignore both small partitions. */
                POP (lo, hi); 
              else              /* Ignore small left partition. */  
                lo = left_ptr;
            }
          else if ((hi - left_ptr) <= max_thresh) /* Ignore small right partition. */
            hi = right_ptr;
          else if ((right_ptr - lo) > (hi - left_ptr)) /* Push larger left partition indices. */
            {                   
              PUSH (lo, right_ptr);
              lo = left_ptr;
            }
          else                  /* Push larger right partition indices. */
            {                   
              PUSH (left_ptr, hi);
              hi = right_ptr;
            }
        }
    }

  /* Once the BASE_PTR array is partially sorted by quicksort the rest
     is completely sorted using insertion sort, since this is efficient 
     for partitions below MAX_THRESH size. BASE_PTR points to the beginning 
     of the array to sort, and END_PTR points at the very last element in
     the array (*not* one beyond it!). */

#define MIN(X,Y) ((X) < (Y) ? (X) : (Y))

  {
    char *end_ptr = base_ptr + size * (total_elems - 1);
    char *run_ptr;
    char *tmp_ptr = base_ptr;
    char *thresh  = MIN (end_ptr, base_ptr + max_thresh);

    /* Find smallest element in first threshold and place it at the
       array's beginning.  This is the smallest array element,
       and the operation speeds up insertion sort's inner loop. */

    for (run_ptr = tmp_ptr + size; run_ptr <= thresh; run_ptr += size)
      if (CMP (run_ptr, tmp_ptr) < 0)
        tmp_ptr = run_ptr;

    if (tmp_ptr != base_ptr)
      SWAP (tmp_ptr, base_ptr, size);

    /* Insertion sort, running from left-hand-side up to `right-hand-side.' 
       Pretty much straight out of the original GNU qsort routine. */

    for (run_ptr = base_ptr + size; (tmp_ptr = run_ptr += size) <= end_ptr; )
      {

        while (CMP (run_ptr, tmp_ptr -= size) < 0)
          ;

        if ((tmp_ptr += size) != run_ptr)
          {
            char *trav;

            for (trav = run_ptr + size; --trav >= run_ptr;)
              {
                char c = *trav;
                char *hi, *lo;

                for (hi = lo = trav; (lo -= size) >= tmp_ptr; hi = lo)
                  *hi = *lo;
                *hi = c;
              }
          }

      }
  }
  return 1;
}