use Data::Dumper;
use SVN::Log;
my $revs = SVN::Log::retrieve($repoURL, $rev1, $rev2);
print Dumper($revs);
[download]

Also look at the test code for t/02basics.t in the CPAN files for more examples on how to descend the data structure.

UPDATE: Here is an actual example that works on my svn repo:

use warnings;
use strict;
use SVN::Log;
use Data::Dumper;

my $revs = SVN::Log::retrieve('my://repo', 1, 5);
#print Dumper($revs);

print $revs->[0]{revision}, "\n";
print $revs->[0]{date    }, "\n";
print $revs->[0]{author  }, "\n";

__END__

1
2012-11-13 14:16:18
root
[download]

Here is a debugger session, based on a made-up structure, similar to yours:
I have interjected comments starting with >>

$ perl -dE 'my @revs=([{author=>"myid",date=>"2013-06-21"},2,3]); say 
+$revs[0]{date}'

Loading DB routines from perl5db.pl version 1.33
Editor support available.

Enter h or `h h' for help, or `man perldebug' for more help.

main::(-e:1):    my @revs=([{author=>"myid",date=>"2013-06-21"},2,3]);
+ say $revs[0]{date}
>>>> Step into the initialization of the structure ....
  DB<1> s
main::(-e:1):    my @revs=([{author=>"myid",date=>"2013-06-21"},2,3]);
+ say $revs[0]{date}

>>>>>*** Looking at the whole structure (Not a good idea if it is larg
+e)
  DB<1> x @revs
0  ARRAY(0x1765270)
   0  HASH(0x1765210)
      'author' => 'myid'
      'date' => '2013-06-21'
   1  2
   2  3

>>>>> Slow dive - We know @revs is an ARRAY - let us see what TYPE the
+ first (0) element is :
  DB<2> x ref $revs[0]
0  'ARRAY'

>>>> OK - so it is an 'ARRAY' - lets examine the TYPE of the first (0)
+ index.
  DB<7> x ref $revs[0][0]
0  'HASH'

>>> Aha - so $revs[0][0] is a HASHREF - lets see what KEYS are availab
+le...
  DB<9> x keys %{ $revs[0][0] }
0  'date'
1  'author'

>>>> Now we could do another "ref', but assuming this will result in a
+ target value, we simply try a key:
  DB<10> x $revs[0][0]{date}
0  '2013-06-21'

>>> end the debug session with 'quit'
  DB<11> q
[download]

Because I came across my own question when I searched on this topic (having apparently given up the previous time) and I am now working on the same problem. I thought I would post my own solution (which I expect deserves to be savaged). So the poor schmuck wrestling with the same problem, has something to reflect on.

For the record I read the suggestions (thanking contributors very much for your assistance). It must have been a bit above my ability to grok apparently. Yes, I may be a dope. For me it's always the syntax that baffles me and I confess I do not completely understand what I am doing even now, yet have to do it anyway.

So here is my solution in part.

use strict;
use warnings;
use SVN::Log;
my $SVNrevs=SVN::Log::retrieve($branchURL, $firstTagRev, $secondTagRev
+);
$i=0; #need this because part the data structure demands it.
foreach @$SVNrevs){
# the easy part
    print $SVNrevs->[$i]{revision}, ",";
    print $SVNrevs->[$i]{date    }, ",";
    print $SVNrevs->[$i]{author  }, "\n";

# This is the pissy bit. Getting the path and the action which are chi
+ldren of the path element (the embedded hash)
    
   foreach my $recordpath ( keys $SVNrevs->[$i] ) {
      for my $path ( keys %{ $SVNrevs->[$i]{$recordpath} } ) {
         foreach my $action ($SVNrevs->[$i]{$recordpath}{$path}{'actio
+n'}) {
      print "\t$action\t$path\n";
      }
      }
    }
    my $message=$SVNrevs->[$i]{message  };
    $_=$message;
    s/\n/ /gi;   #cleaning up random newlines in the message
    $message = $_;
    print "\n\t $message \n\n";
#  }
$i++; #iterate for the next record
}
[download]

Incidentally much of the example code (in the module doc and elsewhere including O'Reilly's Perl books for general concepts and examples, which I rely heavily upon) caused unexpected results (the way I applied it) such as duplicating the array (@SVNrevs) which was unnecessary, and I found the while loop to be tricky to figure out how many times to iterate. I made it work. It may not be elegant, but it's doing the job.