Question

I have a huge java codebase (more than 10,000 java classes) that makes extensive use of CORBA (no documentation available on its usage though).

As first step to figure out the CORBA usage, I decided to scan entire codebase and extract/print unique lines which contain the pattern “org.omg.CORBA“. These are usually in the import statements (e.g. import org.omg.CORBA.x.y.z).

I am newbie to Perl and want to know if there is a way I can extract these details on Windows OS. I need to be able to scan all folders (and sub-folders) that have java classes.

Answer 1

You can use File::Find in a one-liner:

perl -MFile::Find -lwe "
    find(sub { if (-f && /\.java$/) { push @ARGV,$File::Find::name } },'.'); 
    while(<>) { /org.omg.CORBA/ && $seen{$_}++; }; 
    print for keys %seen;"

Note that this one-liner is using the double quotes required for Windows.

This will search the current directory recursively for files with extension .java and add them to the @ARGV array. Then we use the diamond operator to open the files and search for the string org.omg.CORBA, and if it is found, that line is added as a key to the %seen hash, which will effectively remove duplicates. The last statement prints out all the unique keys in the hash.

In script form it looks like this:

use strict;
use warnings;
use File::Find;

find(sub { if (-f && /\.java$/) { push @ARGV,$File::Find::name } },'.');
my %seen;
while(<>) {
    /org.omg.CORBA/ && $seen{$_}++; 
}
print "$_\n" for keys %seen;"