BEGIN {
            use CGI::Carp qw(carpout);
            open(LOG, ">>/usr/local/cgi-logs/mycgi-log") or
              die("Unable to open mycgi-log: $!\n");
            carpout(LOG); #Correct way, according to tchrist, is carpo
+ut(\*LOG);
          }
[download]

For some reason, I cannot stand to see it not written as

BEGIN {
    use CGI::Carp qw(carpout);
    carpout(\*LOGGY) if open(LOGGY,'>>'.__FILE__.'.err.log');

}
[download]

I just felt I had to share that.

update: weeeel, S_Shrum, i'd either carpout(\*STDOUT); or make warnings fatal and stick with fatalsToBrowser, like (and this works a little weird)

#!/usr/bin/perl
use warnings FATAL => qw( all );
warn "hello there";
{ # line 4 (notice its in a block)
  BAREWORD;
  BAREWORD;
  BAREWORD;
}
[download]

or more portable (warnings pragma ain't available everywhere)

#!/usr/bin/perl
$SIG{__WARN__} = sub {die("warning: ",@_)};
warn "hello";
__END__
[download]

or better yet

#!/usr/bin/perl
use CGI::Carp qw( fatalsToBrowser );
use Fatal qw( warn );
use strict;
use warnings;
warn "hello";
[download]

and maybe

BEGIN {
   use vars qw( $WARR );
   $SIG{__WARN__} = sub {
                $WARR .= join ' ',(@_),caller(),"\n";
            };
}
warn "hello";
warn "hooyah";
die "aaah";

# ...
END {
    print "Content-Type: text-html\r\n\r\n <pre> $WARR </pre>";
}
[download]

Reccomended reading is warnings, perllexwarn, sigtrap, Perl's Warn and Die Signals.

 


______crazyinsomniac_____________________________ Of all the things I've lost, I miss my mind the most. perl -e "$q=$_;map({chr unpack qq;H*;,$_}split(q;;,q*H*));print;$q/$q;"

Your choice ofcourse :) the CGI::Carp POD mentioned the above example :)

Greetz
Beatnik
... Quidquid perl dictum sit, altum viditur.

This works great...

I'm finding that having to run my script, having it fail, going to the log file, openning it, scrolling to the bottom is just too time consuming. How can I carpout to the browser for faster debugging?

======================
Sean Shrum
http://www.shrum.net

If that open failed, how would S_Shrum be any more enlightened? :)

Perhaps something more along the lines of...

open(LOG, ">>/usr/local/cgi-logs/mycgi-log")
  or do {
    my $error = "Unable to open mycgi-log: $!\n";
    print "Content-Type: text/plain\n\n", $error; # to the screen
    die $error;                                   # & the logs
  };
[download]

...until you're sure that isn't going to kill the script, too (keeping in mind that you may reveal your log path to any visitors).

    --k.

Perl is functional as I did a simple "hello world" test.

The problem that I am finding *now* is that the "use lib" I have defined to point to my modules location is invalid. I have tried the following:

use lib "/lib";
use lib "/cgi-bin/lib";
use lib "$ENV{'DOCUMENT_ROOT'}/cgi-bin/lib";
use lib "/home/httpd/www.shrum.net/cgi-bin/lib";

...but none of these work. I have a email into the support guys at 1dollarhosting to get the correct prefix path to use in the "USE LIB".

Thanx

======================
Sean Shrum
http://www.shrum.net

With this much grief, I would have changed providers long before now ...

That said, you probably want use lib './lib'; if the lib directory is in the same place you keep your CGIs.

That should even work in a chrooted environment, but if it doesn't I suggest some careful probing of where your scripts think they are and what they can see.

Read more... (712 Bytes)

    --k.