Thanks for the suggestion. It is true that the XML tags and data within them is present if i check "view source" in my browser. I finally got that working by replace "<" by "&lt;"before displaying it in browser.
But that is not the only problem. Other than the XML part, there is a large amount of data missing from the output which I normally get if i log onto my web server and run the same command on the command line, i.e, a large part of the output seems to be truncated when I run the command through the web page. What should I do to capture all the output into the perl variable?

I'm still not convinced that the variable isn't capturing all of the output. Perhaps the output is different for different users and the CGI user's output is only a subset of what you get as yourself. It's also possible that the browser times out in downloading a large amount of data though I kind of doubt that. Change your CGI to print the output directly to a file on the server and examine that file to see what the variable is actually capturing.

Are you sure you're missing data in the output? When I do this sort of thing, I include some size information in some debugging output:

my $data = ` ... `;
print "Got data length [$data]\n";
[download]

However, if you are expecting a lot of data, you don't need to store it all in memory (unless you need it for something else). You can open a pipe and read from it. When you get a line from the pipe, you send it to the browser.

open PIPE, " ... | ";
while( <PIPE> )
    {
    ... do stuff ...
    print;
    }
close PIPE or die "Got error [$!|$?]\n";
[download]

While debugging these things, I tend to output everything as "text/plain" so I see exactly what's going on, or I use a perl script (perhaps using WWW::Mechanize) to grab and save the output so I can inspect it later.

Good luck :)

"text/plain" is great for debugging stuff like this, just be careful of things like internet explorer (and others? are there any), that don't follow the standards, and may or may not decide that it's text, and may instead try to display it as html, or whatever else it feels like.

Only part of the output of the command is captured in the variable when the output is large (i.e, when output size is greater than 0.1Mb).

my $bigstuff = `$^X -e'print 1x(1024*1024)' `;
print length $bigstuff,$/;
print $bigstuff,$/;
__END__
[download]

You've received more intelligent answers, but in the spirit of TIMTOWTDI:

print "<pre>";  
for (split("\n", qx(cat somefile))) { # or other output producing comm
+and
    s/</&lt;/g;
    print; 
}
print "</pre>";
[download]

my $var = `property <dataid>`;
print "<td align="middle">$var</td>";
[download]

This command (property) sometimes produces an output of size almost 0.5Mb to 1Mb for some values of <dataid>(I checked the size of the output by executing the "property" command on the web server CLI).

So, your plan is to up upwards of 1MB of data into a single "<td>" element in the html that you output? Okay, I guess.

Regarding the presence of XML-tagged stuff in the middle of a plain-text stream: are you trying to show that verbatim, making all the XML tags visible?

And regarding your testing of the "property" program on the command line: are you sure that all the data ouput from this program is text? (Could there be non-printing control characters or other non-text, binary content? If so, what sort of operating system are you using, and do you need to specify "binary mode" for reading the data?)

It might be better (or at least, easier to debug) with a pipeline open statement:

my $var;
open( PROG, "property $dataid |" );
# binmode PROG;  # might need to do this?
{
    local $/;  # set input rec. separator to undef
    $var = <PROG>;  # slurp output from property
}
close PROG;

$var =~ s/</&lt;/g; # make sure browsers don't see XML tags as tags

# if there is binary data, there's more you'll need to do
# to make it presentable to a browser...
[download]

Hi, Thanks for the suggestion. I got the html formatting working. Regarding the incomplete output, I tried using pipe open like you suggested but I am still not able to get the entire output(just the same incomplete output). Also I tried to redirect the output from the "property" command to a file on my account on the webserver instead of directly sending it to the browser and still the file didn't contain the entire output. It just contained the incomplete output that I used to get earlier on the browser. But if I log onto my account on the webserver and run the "property" command and redirect it to a file, I get the COMPLETE output in the file. I also observed that the file was about 400Kb in size which is not too big. Any suggestions on what else I could try? Thanks, vinay

Hi guys, Thanks for all your suggestions. I did learn some good debugging techniques in perl. I finally got the script working to retrieve the large output. The problem was that the web admin had put a cap on the apache memlimit directive and hence I wasn't able to retrieve the complete output. Once we increased that limit observing the memory requirements of the "property" command, everything started working!! :) Thanks a lot :) Regards, Vinay