Inconsistent core perl behaviour ?

[p5pRT]

Migrated from rt.perl.org#1832 (status was 'resolved')

Searchable as RT1832$

[p5pRT]

From rick.delaney@home.com

David Cassell wrote​:

geoff_gunner@​my-deja.com wrote​:

Perl grows arrays on demand​:
@​array=(); # a nice empty array
$array[4]='x'; # elements 0..3 are undef; 4 is
'x'

So scalar @​array would return 4, right? And an offset
of 8 is meaningless at this point...

It is no more meaningless than the offset of 4 was after the line

  @​array=();

So​: let's say that you now do​:
splice(@​array, 8, 0, 'a', 'b', 'c');
elements 8, 9 and 10 should now be a, b and c,
right ? Wrong - they're at 5,6 and 7. Even
though you explicitly asked for the data to go
here, perl hasn't created the extra elements.

So. You ask Perl to do something absurd, and it
does the most logical thing it can find. But you don't
like it. Still, I don't think that makes it a bug.

I agree that the behaviour is not a bug but it seems there is a tiny
error in the docs for splice.

=item splice ARRAY,OFFSET,LENGTH,LIST

=item splice ARRAY,OFFSET,LENGTH

=item splice ARRAY,OFFSET

Removes the elements designated by OFFSET and LENGTH from an array, and
replaces them with the elements of LIST, if any. In list context,
returns the elements removed from the array. In scalar context,
returns the last element removed, or C<undef> if no elements are
removed. The array grows or shrinks as necessary.
# ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

If OFFSET is negative then it start that far from the end of the array.
If LENGTH is omitted, removes everything from OFFSET onward.
If LENGTH is negative, leave that many elements off the end of the
array.
The following equivalences hold (assuming C<$[ == 0>)​:
# ^^^^^^^^^^^^

  push(@​a,$x,$y) splice(@​a,@​a,0,$x,$y)
  pop(@​a) splice(@​a,-1)
  shift(@​a) splice(@​a,0,1)
  unshift(@​a,$x,$y) splice(@​a,0,0,$x,$y)
  $a[$x] = $y splice(@​a,$x,1,$y)
# ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

These are not equivalent if $x &gt; $#a + 1.

May I suggest changing

If OFFSET is negative then it start that far from the end of the array.

to

If OFFSET is negative then it starts that far from the end of the array.
If OFFSET is greater than the last index of the array then it starts
from the last index plus one.

Please note that I've also changed "start" to "starts" so even if you
don't like it, at least the typo could be fixed. :-)

--
Rick Delaney
rick.delaney@​home.com

[p5pRT]

From [Unknown Contact. See original ticket]

This bug is still true. According to the docs the following two should be
equivalent.

$a[$x] = $y splice(@​a,$x,1,$y)

However, they are not. This simple case shows​:

$x[5]=5;
print $#x;
print "\n";
print join("​:", @​x);

# prints
5
:​::​::5

splice(@​x,5,1,7);
print $#x;
print "\n";
print join("​:", @​x);

# prints
0
5

-spp

[p5pRT]

From @simoncozens

On Mon, Dec 11, 2000 at 03​:31​:37PM -0500, Stephen P. Potter wrote​:

splice(@​x,5,1,7);
print $#x;
print "\n";
print join("​:", @​x);

# prints
0
5

This is even tested for​:

print "not " unless j(splice(@​a, 20, 0, 12, 13)) eq "" && j(@​a) eq j(0..13);
print "ok 6\n";

So what's right? Is it the documentation or the tests? If it's the
documentation​:

Inline Patch

--- pp.c~	Sun Dec 24 01:39:33 2000
+++ pp.c	Sun Dec 24 01:53:26 2000
@@ -3708,8 +3708,9 @@
 	length = AvMAX(ary) + 1;
     }
     if (offset > AvFILLp(ary) + 1)
-	offset = AvFILLp(ary) + 1;
-    after = AvFILLp(ary) + 1 - (offset + length);
+	av_fill(ary, offset + length);
+    
+    after = av_len(ary) + 1 - (offset + length);
     if (after < 0) {				/* not that much array */
 	length += after;			/* offset+length now in array */
 	after = 0;

And rewrite your own tests...

[p5pRT]

@rgs - Status changed from 'open' to 'resolved'