The behavior of array and hash OPs

llancet has asked for the wisdom of the Perl Monks concerning the following question:

In perl, array get and set, hash get and set are value-based. When you set/get a scalar from an array/hash, its value is copied:

my @array;
my $i = "abcd";
$array[0] = $i;

# $array[0] won't change
$i = "efgh";

my %hash;
my $i2 = 1234;
$hash{'key'} = $i2;

# $hash{'key'} won't change
$i2 = 5678;
[download]

However, in the perl's C interface, calls like av_store, av_fetch, hv_store, hv_fetch, av_push, av_pop, the scalar itself is really being stored to the container, or fetched from the container, instead of its copy. So, if I'm going to make a general-purpose C++ wrapper for perl, should I imitate the behavior of perl, or just follow the behavior of perlapi?

Comment on The behavior of array and hash OPs Download Code

Replies are listed 'Best First'.
Re: The behavior of array and hash OPs by Anonymous Monk on Jun 22, 2011 at 07:07 UTC
the scalar itself is really being stored to the container, or fetched from the container, instead of its copy Its got two parts. `$array[0] = $i;` [download] is really something like `$array[0] = sv_mortalcopy($i);` [download] There is no av_store_copy, there is av_store and sv_mortalcopy. Seems somebody already thought about this some https://github.com/Leont/libperl--/blob/master/perl++/headers/array.h https://github.com/Leont/libperl--/blob/master/perl++/doc/Array.pod	[reply] [d/l] [select]
Re: The behavior of array and hash OPs by 7stud (Deacon) on Jun 22, 2011 at 01:27 UTC
When you set/get a scalar from an array/hash, its value is copied: Not always: `use strict; use warnings; use 5.010; my @arr = (1, 2, 3); for (@arr) { $_ *= 2; } print "@arr\n"; --output:-- 2 4 6` [download]	[reply] [d/l]
Re^2: The behavior of array and hash OPs by ikegami (Patriarch) on Jun 22, 2011 at 04:05 UTC
Yes, always. Unless you claim you can change the result of the multiplication after you assigned it, and that doing so changes the array.	[reply]
Re^2: The behavior of array and hash OPs by 7stud (Deacon) on Jun 22, 2011 at 06:01 UTC
How about if I claim that when perl gets the thing from the array, it is not a copy of the value because when I change it, the array changes?	[reply]
Re^3: The behavior of array and hash OPs by frieduck (Hermit) on Jun 22, 2011 at 16:05 UTC
How about you claim that "the foreach loop index variable is an implicit alias for each item in the list that you're looping over. " as mentioned in the perlsyn section on foreach?	[reply]
A reply falls below the community's threshold of quality. You may see it by logging in.
Re^2: The behavior of array and hash OPs by llancet (Friar) on Jun 23, 2011 at 08:35 UTC
The array iterator is specially treated. If you use more 'common' way, it is copied by value: `$ perl -E 'my @array=qw/1 2 3/; my $first=$array[0]; $first=0; say "ar +ray: @array"; say "fetched changed: $first";' array: 1 2 3 fetched changed: 0` [download]	[reply] [d/l]
Re: The behavior of array and hash OPs by locked_user sundialsvc4 (Abbot) on Jun 22, 2011 at 15:58 UTC
The magic is “references,” which are scalars too. Perhaps you could elaborate on this “general-purpose wrapper,” and describe in more detail exactly how you intend for it to be used by C++ programmers ...
Re^2: The behavior of array and hash OPs by llancet (Friar) on Jun 23, 2011 at 08:39 UTC
Actually, not too "general". I attempt to design and use it as a scripting engine in a game. So, the typical usage would be: calling a perl sub with wrapped game object as argument.	[reply]