Ro has asked for the wisdom of the Perl Monks concerning the following question:

Hi Monks,

This is what I am trying to do: Look for a string of characters //** in all .js files recursively in a directory, then print out the characters that come right after the string. I want to do this for each line that starts with the string mentioned above to a specified .html file.

My code works when grabbing one file, but looping through a directory is causing me issues. I am thinking I need File Find, but am stuck and don't know how to go about it.

If anyone could offer any help/ideas, that would allow me to do this while but with a bunch of files, that would be great.
My Code (That works for one specified file): 
# i have all the correct headers use filefind, /opt/bin/perl bla bla

$jscriptdir = "/home/javascript/";
$output = "/home/web/foo/index.htm";
$reading = "/home/javascript/jumper.js";

# Open file, match comment string, write to html
open(JS,"<$reading") or die "Can't Open $reading: $!";
open(NEW,">$output") or die "Can't Open $output: $!";
print NEW "<html><title>Comment Extractor</title></html>\n";
print NEW "<body>\n";
print NEW "<h1>Comment Extractor</h1>\n";
print NEW "<hr>\n";

    $script_title = rindex($reading, "/");
    print NEW substr($reading, $script_title+1);
    while (<JS>)
    {
        if ($_ =~ /\*\*/) {
            
            print NEW substr($_,4) . "<br>";
        }
    }
    
    print NEW "</body></html>\n";

close NEW;
close JS;

[download]
  • Comment on Find, read, write out contents of a certain file type recursively...
  • Download Code

Replies are listed 'Best First'.
Re: Find, read, write out contents of a certain file type recursively...
by gjb (Vicar) on Nov 14, 2002 at 19:44 UTC

    If all .js files are in a single directory, File::Find isn't the only option. 

    opendir(DIR, $jscriptdir) or die("Can't open $jscriptdir");
my @jsFiles = grep(/\.js$/, readdir(DIR));
closedir(DIR);
foreach my $jsFile (@jsFiles) {
    # do whatever
}

    [download]
    If you really want to recurse through a directory hierarcy, you can use the code below: 
    use File::Find;
use IO::File;

find(\&wanted, $jscriptdir);

sub wanted {
    if (/\.js$/) {
        my $fh = new IO::File($_) or die("Can't open $File::Find::dir/
+$_");
        while (<$fh>) {
            # your while here
        }
        $fh->close();
    }
}

    [download]
    Hope this helps, -gjb-

[reply]
[d/l]
[select]
Re: Find, read, write out contents of a certain file type recursively...
by jdporter (Paladin) on Nov 14, 2002 at 20:01 UTC
    Here's one way to do it. 
        use File::Find;
    my $jscriptdir = "/home/javascript/";
    my $pat = quotemeta '//**';
    find( \&proc, $jscriptdir );
    sub proc {
        /\.js$/ or return;
        my $n = $File::Find::name;
        open F, "< $n" or die "read $n: $!\n";
        my @l = map { s/^$pat// ? $_ : () } <F>;
        close F;
        @l or return;
        $n =~ s/\Q$jscriptdir\E.//;
        print "$n $_" for @l;
    }

    [download]
    Adding the HTML is left as an exercise. :-)

    jdporter
    ...porque es dificil estar guapo y blanco.

[reply]
[d/l]
Re: Find, read, write out contents of a certain file type recursively...
by Ro (Initiate) on Nov 14, 2002 at 20:21 UTC
    Thanks for the help guys. I got this thing working. I will post my modified code. Since I am new, I will look around and see, if I can post the source so someone else can benefit. Thanks! 
    
# Variables
$output = "/home/web/path/to/foo/foo.htm";
$jscriptdir = "/home/web/foo/javascript/";
$javadir = "/home/web/foo/servlet/";
$cgidir = "/home/web/cgi-bin/";

# Open the output file
open(NEW,">$output") or die "Can't Open $output: $!";
print NEW "<html><head><title>Comment Extractor</title></head>\n";
print NEW "<body>\n";
print NEW "<h1>Comment Extractor</h1>\n";
print NEW "<hr>\n";

# Let's find files. Call jsfind for the dirty work
find (\&jsfind, $jscriptdir);

print NEW "</body></html>\n";
close NEW;
exit;

sub jsfind {
    # get all file names that have last 2 characters "js"
    if($File::Find::name=~/\.js$/)
    {
        # ignore js files in those stupid Frontpage _vti* directories
        if ($File::Find::name=~/\_vti/) {} else 
        {
                # get name of script file and print it in red
            $script_title = $_;
            print NEW "<b><font style=\"color:red; size:16px; text-tra
+nsform: uppercase\">" . $script_title . "</font></b>";
            # placeholder for file
                # start opening files to read
                open(FILE, $File::Find::name) or die "could not open f
+ile - $_ - : $!";
                # iterate thru files to find match //**
                    while (<FILE>)
                {
                    if ($_ =~ /\*\*/) {
                        # print files to $output
                        print NEW substr($_,4) . "<br>";
                    }
                }
                close FILE;
        }
    } else {
    
    return;
        
    }
    
}

    [download]
[reply]
[d/l]
      I will post my modified code

      Thanks! (Um, be sure to post the whole thing... -- I didn't see a line saying "use File::Find" when I read this post.)

      Just one concept you should consider: doing a recursive directory search for files of a given type (e.g. *.js) is a very common facility that is handy for a wide range of particular needs -- that's why the GNU "find" utility (and the decades-old unix "find" that it's modeled on) is such a basic, essential component on so many systems (it's been ported to windows, etc).

      Even if you want to stick with File::Find (which happens to be a few times slower than GNU "find"), you should consider making it a separate utility by itself, and keep just the editing function in a simpler app that works on a single file, or on a list of files read from stdin -- e.g.: 

        my_find_utility '*.js' | my_comment_extractor

      [download]
      This way, when you come up with some other particular edit or summarization process to be done on all files of a given type in a directory tree, you don't have to re-write the part that recurses through the directories. Just write a script that will apply the new process to any list of file names on stdin, and use the same front-end program (or GNU "find") to feed it.
[reply]
[d/l]
Re: Find, read, write out contents of a certain file type recursively...
by dingus (Friar) on Nov 15, 2002 at 09:20 UTC
    This is my way of doing the recursive look for all files in a tree and process those that match a pattern. I'm using DosGlob because I do this on Windows PCs. I'm pretty sure that using regular glob works too but haven't tried it. 
    use File::DosGlob;
$dir=  '/home/javascript/*'; # note trailing * !!!
my @m = File::DosGlob::doglob(1,$dir);
for (@m) {
  push @m, File::DosGlob::doglob(1,$_.'/*') if -d($_);
  next unless (/\.js$/);
# put your code here to
# open file ($_) and search
  open(JS,"<$_") or die "Can't Open $_: $!";
# etc.
}

    [download]
    PS you probably want to open your output file outside the search LOOP!

    Dingus

    Enter any 47-digit prime number to continue.
[reply]
[d/l]

Back to Seekers of Perl Wisdom