$sort::mergesort_bit = 0x00000002;
$sort::sort_bits = 0x000000FF; # allow 256 different ones
$sort::stable_bit = 0x00000100;
-$sort::insensitive_bit = 0x00000200;
-$sort::safe_bits = 0x00000300;
-$sort::fast_bit = 0x00000400;
use strict;
}
$^H |= $sort::hint_bits;
local $_;
+ no warnings 'uninitialized'; # $^H{SORT} bitops would warn
while ($_ = shift(@_)) {
- if (/^q(?:uick)?sort$/) {
+ if (/^_q(?:uick)?sort$/) {
$^H{SORT} &= ~$sort::sort_bits;
$^H{SORT} |= $sort::quicksort_bit;
- return;
- } elsif ($_ eq 'mergesort') {
+ } elsif ($_ eq '_mergesort') {
$^H{SORT} &= ~$sort::sort_bits;
$^H{SORT} |= $sort::mergesort_bit;
- return;
- } elsif ($_ eq 'safe') {
- $^H{SORT} &= ~$sort::fast_bit;
- $^H{SORT} |= $sort::safe_bits;
- $_ = 'mergesort';
- redo;
- } elsif ($_ eq 'fast') {
- $^H{SORT} &= ~$sort::safe_bits;
- $^H{SORT} |= $sort::fast_bit;
- $_ = 'quicksort';
- redo;
+ } elsif ($_ eq 'stable') {
+ $^H{SORT} |= $sort::stable_bit;
} else {
require Carp;
- Carp::croak("sort: unknown subpragma '@_'");
+ Carp::croak("sort: unknown subpragma '$_'");
}
}
}
if ($^H{SORT}) {
push @sort, 'quicksort' if $^H{SORT} & $sort::quicksort_bit;
push @sort, 'mergesort' if $^H{SORT} & $sort::mergesort_bit;
- push @sort, 'safe' if $^H{SORT} & $sort::safe_bits;
- push @sort, 'fast' if $^H{SORT} & $sort::fast_bit;
+ push @sort, 'stable' if $^H{SORT} & $sort::stable_bit;
}
push @sort, 'mergesort' unless @sort;
join(' ', @sort);
=head1 SYNOPSIS
- use sort 'quicksort';
- use sort 'mergesort';
+ use sort 'stable'; # guarantee stability
+ use sort '_quicksort'; # use a quicksort algorithm
+ use sort '_mergesort'; # use a mergesort algorithm
- use sort 'qsort'; # alias for quicksort
+ use sort '_qsort'; # alias for quicksort
- # alias for mergesort: insenstive and stable
- use sort 'safe';
-
- # alias for raw quicksort: sensitive and nonstable
- use sort 'fast';
-
- my $current = sort::current();
+ my $current = sort::current(); # identify prevailing algorithm
=head1 DESCRIPTION
sort() function.
In Perl versions 5.6 and earlier the quicksort algorithm was used to
-implement sort(), but in Perl 5.8 the algorithm was changed to mergesort,
-mainly to guarantee insensitiveness to sort input: the worst case of
-quicksort is O(N**2), while mergesort is always O(N log N).
-
-On the other hand, for same cases (especially for shorter inputs)
-quicksort is faster.
-
-In Perl 5.8 and later by default quicksort is wrapped into a
-stabilizing layer. A stable sort means that for records that compare
-equal, the original input ordering is preserved. Mergesort is stable;
-quicksort is not.
-
-The metapragmas 'fast' and 'safe' select quicksort without the
-stabilizing layer and mergesort, respectively. In other words,
-'safe' is the default.
-
-Finally, the sort performance is also dependent on the platform
-(smaller CPU caches favour quicksort).
+implement sort(), but in Perl 5.8 a mergesort algorithm was also made
+available, mainly to guarantee worst case O(N log N) behaviour:
+the worst case of quicksort is O(N**2). In Perl 5.8 and later,
+quicksort defends against quadratic behaviour by shuffling large
+arrays before sorting.
+
+A stable sort means that for records that compare equal, the original
+input ordering is preserved. Mergesort is stable, quicksort is not.
+Stability will matter only if elements that compare equal can be
+distinguished in some other way. That means that simple numerical
+and lexical sorts do not profit from stability, since equal elements
+are indistinguishable. However, with a comparison such as
+
+ { substr($a, 0, 3) cmp substr($b, 0, 3) }
+
+stability might matter because elements that compare equal on the
+first 3 characters may be distinguished based on subsequent characters.
+In Perl 5.8 and later, quicksort can be stabilized, but doing so will
+add overhead, so it should only be done if it matters.
+
+The best algorithm depends on many things. On average, mergesort
+does fewer comparisons than quicksort, so it may be better when
+complicated comparison routines are used. Mergesort also takes
+advantage of pre-existing order, so it would be favored for using
+sort to merge several sorted arrays. On the other hand, quicksort
+is often faster for small arrays, and on platforms with small memory
+caches that are much faster than main memory. You can force the
+choice of algorithm with this pragma, but this feels heavy-handed,
+so the subpragmas beginning with a C<_> may not persist beyond Perl 5.8.
+
+=head1 CAVEATS
+
+This pragma is not lexically scoped : its effect is global to the program
+it appears in. This may change in future versions.
=cut
https://perl5.git.perl.org / perl5.git / blobdiff