index wrapping in tied FETCH and EXISTS

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

Replies are listed 'Best First'.
Re: index wrapping in tied FETCH and EXISTS by apl (Monsignor) on Feb 04, 2009 at 12:41 UTC
If I am doing some kind of subscript computation and my subscript happens to go negative, I want to know about it (by receiving undef) NOW How about `my $i = ... some calculation ... if ( $i >= 0 ) { $val = $tiedArray[$i]; } else { # report the problem! }` [download]	[reply] [d/l]
Re: index wrapping in tied FETCH and EXISTS by Anonymous Monk on Feb 04, 2009 at 08:17 UTC
Where does perl get off documenting features? http://search.cpan.org/~nwclark/perl-5.8.9/pod/perldata.pod#Subscripts The array indices start with 0. A negative subscript retrieves its value from the end. In our example...	[reply]
Re: index wrapping in tied FETCH and EXISTS by ikegami (Patriarch) on Feb 04, 2009 at 13:25 UTC
It's a very useful and commonly used feature. `sub peek { return $stack[-1]; }` [download]	[reply] [d/l]
Re: index wrapping in tied FETCH and EXISTS by cmac (Monk) on Feb 04, 2009 at 15:59 UTC
Oh. If non-tied arrays work that way too, the `our NEGATIVE_INDICES=1;`I put in my tied array class will come out. Also the underlying direct-call methods for array handling will have to change, in how they handle a negative index. The original note came about because I was adapting a test sequence for said direct-call methods to a tied array, and got irked when the tie structure broke a test for "outside the array" (below zero). Sorry, cmac	[reply] [d/l]
Re^2: index wrapping in tied FETCH and EXISTS by ikegami (Patriarch) on Feb 04, 2009 at 16:32 UTC
You can reach lower than the start of the array by using an index lower than `-@array`. `$ perl -wle'$\|=1; @a=qw(a b c); print "$_: $a[$_]" for -@a-1..$#a+1' Use of uninitialized value in concatenation (.) or string -4: -3: a -2: b -1: c 0: a 1: b 2: c Use of uninitialized value in concatenation (.) or string 3:` [download]	[reply] [d/l] [select]
Re^3: index wrapping in tied FETCH and EXISTS by cmac (Monk) on Feb 04, 2009 at 20:11 UTC
To me your test shows that you CAN'T reach lower than the start of the array by using an index lower than `-@array`. A return value of `undef` is the only way that perl has to tell you that you're doing something bad. OK, not the only way, but dieing/croaking seems too strong for an out-of-range array index. It occurred since my previous note that I shouldn't emulate this behavior in the direct-call routines that underlie my tied arrays, because then a user of a tied array would be faced with two layers of "wrapping around". Unless I can figure out how to differentiate a tied call from a direct one. Since tied calls come through a little "dereference wrapper" from my AUTOLOAD routine (its PM thread ended yesterday), this should be possible, but I'm not sure it's worth the overhead. cmac www.animalhead.com	[reply] [d/l] [select]
Re^4: index wrapping in tied FETCH and EXISTS by ikegami (Patriarch) on Feb 04, 2009 at 20:43 UTC
Re^5: index wrapping in tied FETCH and EXISTS by cmac (Monk) on Feb 05, 2009 at 06:15 UTC