Hi Sluggers:
I have been looking at an efficient way to move away from having to
constantly have paths specified in scripts repeatedly. So here is the full
story in great detail so that I can hopefully get an informed decision.
This is not a problem but merely looking for an efficient way to do this.
When I write scripts (Perl) I normally specify the full paths to the base
root directory (base html dir) and the path to cgi-bin specified. With
these two paths I can access other scripts interfaces after pushing the
path in @INC and use 'require', and also process text files etc ....
Now I was normally specifying these two paths in all scripts at the top in
the past. But recently have been thinking about moving away from this. So
here are two things I did but is not quite to what I want to achieve but
close.
Attempt 1.
===========
I basically have a text file where I have these two full paths specified.
The file is located in the same directory as all the scripts. If there are
scripts in another directory, then in that directory I have a sim link to
the file that holds the actual path data. Now in all scripts, I just do
this before anything else to set the paths:
my @paths;
open (FILE, "path.dat") || die "error blah blah ...";
while(<FILE>) {
chomp;
push @paths, $_;
}
close (FILE);
As I know the order of the paths in the file, I have these two below as
globals:
my $homedir = $paths[0];
my $cgibin = $paths[1];
This work greats with browser called scripts, but I hit a problem with
scripts that runs via cron. The problem with cron scripts is that it
cannot open the "path.dat" file despite that it's in the same directory as
the cron script itself. I think where cron executes (don't know where)
it's not in reference with the same directory where the script and file is
located, so cannot see it.
So I moved away from this solution and went to attempt 2.
Attempt 2.
===========
I create a 'path.pl' script where I specify $homedir, $cgibin, and other
other common used stuff by all scripts via a routine called "set_paths()".
Then with "Exporter::Lite", I export these two variables and the others.
In other scripts the problem is that I have to tell it from this 'path.pl'
script is. So I am forced to have one path specified. i.e I have to define
$cgibin = "/path_to_where_path.pl_is_located";
Then I do this
push($cgibin, @INC);
require 'path.pl';
&set_paths();
This now has all common stuff accessible. But I still have to specify one
hardcoded path in all scripts which is no way as good as attempt one. With
attempt 2 cron scripts also works fine.
I have been looking at a way to have @INC permanently have the path to
where this 'path.pl' is located so that all I need to do is just call
"&set_paths()". I read about this from this url
http://perl.apache.org/docs/1.0/guide/porting.html#_INC_and_mod_perl
However I'm not sure what configuration file they are talking about and
also what is this startup.pl script located. It also appears that only the
server administrator can do this. Is this right ?
So for now I am with attempt 2 as I can run cron and browser called scripts.
If anyone have some thoughts or a better solution on this please share
them with me.
Cheers.
--
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug
