Ok, let's hope i can shed some light on this situation.
What's going wrong is that you are using a reference to an array slice...
not sure why that's a big problem but it is apparently.
in your case, there are 2 solutions:
1) do something like the following if you insist on using references
@foo = (1,2,3,4,5,6,7,8,9);
@baz = @foo[1..5];
bar(\@baz);
sub bar {
my $aref = shift;
for (@$aref) {print}
}
#which will nicely print '23456'
2) Seeing your passing an array slice to the sub, there is no real need for
references, so you could do something like this:
@foo = (1,2,3,4,5,6,7,8,9);
bar(@foo[1..5]);
sub bar {
my @a = @_;
for (@a) {print}
}
#which will nicely print '23456'
I hope this will solve your problem for now...
Regards,
Jos Boumans
Paul wrote:
> --- Gary Stainburn <[EMAIL PROTECTED]> wrote:
> > Hi all,
>
> Hi, Gary. =o)
>
> > I have an array that holds the source code to a cobol program. I
> > then have a number of integers that hold the subscript for section
> and
> > division changes.
> > I then have subs that parse sections of that array. e.g. a sub called
> > file_control will parse the file-control section and extract
> > information from the file select statements.
>
> Actually, that sounds like a pretty cool arrangement.
>
> > At the moment my code works because I store the array as a global
> > variable and pass to the sub the start and end array elements.
>
> Ah. A reasonable way to deal with the circumstance, but globals are
> commonly considered "bad", lol....
>
> > I want to now make my code a bit more flexible and simply pass a
> > reference to the section I require. The code I've tried may give a
> > better idea of what i mean. when I run my code I get the following
> > error on the 'foreach (@$lines)' line
> >
> > Not an ARRAY reference at ./s line 109.
>
> Hmm....looking below, the cause is not immediately apparent to me.
> In fact, what you've done was the first solution that popped into my
> mind. (Somebody please explain the problem with that for me, too?)
>
> > I could simply pass the array sections to the subroutines but that
> > would be very inefficient - the procedure divisions of some the
> > programs are 5k lines - hence the attempt to reference.
>
> I think this is a good candidate for an object.
> Try this:
>
> my $obj = {};
> $obj->{CODE} = [ @lines ];
>
> now you can undef @lines to recoup that memory (though Perl probably
> won't give it back.....)
>
> Then you can do neat things like putting the section indexes on the
> object also:
>
> $obj->{FILECTL} = $fc_ndx;
>
> Then all you have to do is pass the object to each function.
> file_control($obj);
>
> and they can grab the pieces they need off it.
>
> sub func {
> my $obj = shift;
> for my $ndx ($obj->{SECTION1} .. ($obj->{SECTION2}-1)) {
> # code for section
> }
> }
>
> try looking at the perldocs for
> perlref
> perlreftut
> perlobj
> perlboot
> perltoot
> perltootc
>
> Paul
>
> (Original code left below for reader's convenience)
>
> > my @lines=( # simulate read cobol file
> > ........
> > " file-control.\r\n",
> > " select jobfile\r\n",
> > " assign external das1\r\n",
> > " organization is indexed\r\n",
> > " access dynamic\r\n",
> > " record key is job-jf-key\r\n",
> > " file status is w01-jobf-stat\r\n",
> > " lock mode is automatic with rollback.\r\n",
> > .......
> >
> > &file_control(\@lines[10..36]); # pass pointer to array elements
> > 10-36
> >
> > sub file_control() {
> > my ($lines)=@_;
> > foreach (@$lines) { # process each line in array section
> > next if (/^ [\/\*]/);
> > ........
>
> __________________________________________________
> Do You Yahoo!?
> Yahoo! Auctions - buy the things you want at great prices
> http://auctions.yahoo.com/
