1package sort; 2 3use strict; 4use warnings; 5 6our $VERSION = '2.05'; 7 8sub import { 9 shift; 10 if (@_ == 0) { 11 require Carp; 12 Carp::croak("sort pragma requires arguments"); 13 } 14 $^H{sort} //= 0; 15 for my $subpragma (@_) { 16 next 17 if $subpragma eq 'stable' || $subpragma eq 'defaults'; 18 require Carp; 19 Carp::croak("sort: unknown subpragma '$_'"); 20 } 21} 22 23sub unimport { 24 shift; 25 if (@_ == 0) { 26 require Carp; 27 Carp::croak("sort pragma requires arguments"); 28 } 29 for my $subpragma (@_) { 30 next 31 if $subpragma eq 'stable'; 32 require Carp; 33 Carp::croak("sort: unknown subpragma '$_'"); 34 } 35} 36 37sub current { 38 warnings::warnif("deprecated", "sort::current is deprecated, and will always return 'stable'"); 39 return 'stable'; 40} 41 421; 43__END__ 44 45=head1 NAME 46 47sort - perl pragma to control sort() behaviour 48 49=head1 SYNOPSIS 50 51The sort pragma is now a no-op, and its use is discouraged. These three 52operations are valid, but have no effect: 53 54 use sort 'stable'; # guarantee stability 55 use sort 'defaults'; # revert to default behavior 56 no sort 'stable'; # stability not important 57 58=head1 DESCRIPTION 59 60Historically the C<sort> pragma you can control the behaviour of the builtin 61C<sort()> function. 62 63Prior to v5.28.0 there were two other options: 64 65 use sort '_mergesort'; 66 use sort '_qsort'; # or '_quicksort' 67 68If you try and specify either of these in v5.28+ it will croak. 69 70The default sort has been stable since v5.8.0, and given this consistent 71behaviour for almost two decades, everyone has come to assume stability. 72 73Stability will remain the default - hence there is no need for a pragma for 74code to opt into stability "just in case" this changes - it won't. 75 76We do not foresee going back to offering multiple implementations of general 77purpose sorting - hence there is no future need to offer a pragma to choose 78between them. 79 80If you know that you care that much about performance of your sorting, and 81that for your use case and your data, it was worth investigating 82alternatives, possible to identify an alternative from our default that was 83better, and the cost of switching was worth it, then you know more than we 84do. Likely whatever choices we can give are not as good as implementing your 85own. (For example, a Radix sort can be faster than O(n log n), but can't be 86used for all keys and has larger overheads.) 87 88We are not averse to B<changing> the sort algorithm, but we don't see the 89benefit in offering the choice of two general purpose implementations. 90 91=head1 CAVEATS 92 93The function C<sort::current()> was provided to report the current state of 94the sort pragmata. This function was not exported, and there is no code to 95call it on CPAN. It is now deprecated, and will warn by default. 96 97As we no longer store any sort "state", it can no longer return the correct 98value, so it will always return the string C<stable>, as this is consistent 99with what we actually have implemented. 100 101=cut 102