Regular Expressions - UBB.Developers

Ok heres the deal, I am trying to parse the 'LIST' output of any unix directory. A file entry goes something like this:

drwx------ 5 anpatel students 512 Jan 13 2000 Desktop

Simple way of getting each field is:
my @properties = split(/s+/,$line);

But this code has one specific flaw. What if the file was "New Text Document.txt", there would be no way saying that $properties[9] was the full name of the file. I thought of improving the code:

my @properties = split( /s+/, $item );
my ( $perms, $lCount, $owner, $group, $size, $mod_mon, $mod_date, $mod_time ) = @properties;
my $file = join( "", @properties[8.. $#properties] );

But $file will be "NewTextDocument.txt" because the &join will compine all the rest of array elements with "".

I was looking for a perfect solution if some one could offer, of-course i am on the hunt. But lets see if some one figures out before me

hehehe

Thanks for reading this!

do you need all of the properties, or just a list of filenames?

Basically i need all the properties, but to be exact i need:

if the file is directory ($perms, $properties[0])
File Size ($size, $properties[4])
File Mod Time ( none, $properties[5... 7] )
File name ($file, $properties[8..*])

hmm i think you are suggesting me to use the simple 'ls' command, and find all the file names easily from that, I had this in mind, but dont want to use it (for many reasons)

hmm.. if we might locate somehow with Regexps, the char number of where the file name starts than bingo!.

But sadly i am still trying to find that way out.

Instead of trying to parse the text output of a Unix command, I'd recommend using readdir to scan through the files, and stat to get the file properties.

Thanks, but I am trying to parse the FTP list output, there is no way of readdir() in a FTP client, however there is a ftp->ls() command, that gives you only file name output, but there are no specific ways to merge that information to the ftp->dir() output.

The only choice that is left now is to effectively parse the LIST output.

Thanks for replying!

I did it!!!, i figured out that since we know at what field we need to stop, we just eliminate the fields before what we want. Sorry for bothering ya'all

my $str = "drwx------ 3 anpatel students 512 Apr 18 14:45 Really Bad Name Name for a file.";

$str =~ s/Ss+//; #1 perms-links
$str =~ s/Ss+//; #2 links-user
$str =~ s/Ss+//; #3 user-grp
$str =~ s/Ss+//; #4 grp-apr
$str =~ s/Ss+//; #5 apr-date
$str =~ s/Ss+//; #6 date-time
$str =~ s/Ss+//; #7 time-filename
$str =~ s/^(S+)s+(.*)/$2/; # remove all the things that were combined previously. result is we get the file name!
print $str;
exit;

Quote

Originally posted by Dave_L:

Instead of trying to parse the text output of a Unix command, I'd recommend using readdir to scan through the files, and stat to get the file properties.

How would stat be used after you've read the directory and would like to read/print file names, date, and size?

found some answers:

http://aspn.activestate.com/ASPN/docs/ActivePerl/lib/File/stat.html

so, you no longer have a question?

Still have questions, but that at least tells me the reasoning behind the answer

I can probably write it using the example posted, I'll write back when I have questions.

Basically we readdir then use stat to be able to print the file size ($size) and date ($mtime). $filename gives me the file's name?

Would something like this be in the right direction?

Code

use File::stat;

($dev,$ino,$mode,$nlink,$uid,$gid,$rdev,$size, $atime,$mtime,$ctime,$blksize,$blocks) = stat($filename);

then use $filename, $mtime, and $size for the data I'm looking for?

File::stat gives you a by-name access to stat() values (actually it returns an object). If you'll not use them, just use the CORE stat function (which returns an array)... Else, something like this will work:

Code

my @files = <*.txt>; # all txt in current dir
my $stat;
foreach my $file (@files) {
   $stat = stat $file or die "Error: $!";
   printf "%st%st%sn", $file, format_size($stat->size), format_time($stat->mtime);
}

thank you, that should do it

So in the final print line $file will give me the filename?

yes $file includes file name or full path to the file... depending on your usage...

How does this look?

Code

opendir(UPLOADS, "$opt{dir}/$user_number");
my @uploads = readdir(UPLOADS);
closedir(UPLOADS);
my @files = <*.txt>; # all txt in current dir
my $stat;
foreach my $file (@files) {
	$stat = stat $file or die "Error: $!";

	##output it

	print qq(
	<tr valign="top">
	<td bgcolor="$vars_style{AltColumnColor2}" align="center" width="20"><input type="checkbox" name="$key" value="include" checked="checked" /></td>
	<td bgcolor="$vars_style{AltColumnColor2}"><b><font size="$vars_style{FDTextSize}">$filename</font></b></td>
	<td bgcolor="$vars_style{AltColumnColor2}" align="center" valign="middle">$size</td>
	<td bgcolor="$vars_style{AltColumnColor2}" align="center" valign="middle">$mtime</td>
	</tr>
	);
}

It's not throwing an error, but then it's not reading/printing anything either

How would I fix the section to view all files, not just .txt?

Thank you again for your time

txt one was an example (or demo) on the usage

use this:

Code

opendir(UPLOADS, "$opt{dir}/$user_number") or die "Can not opendir($opt{dir}/$user_number): $!";
my @uploads = readdir(UPLOADS);
closedir(UPLOADS);
my($size, $mtime);
foreach my $filename (@uploads) {
        next if $filename =~ /^./; # ignore anything that starts with a dot. like "." ".." ".htaccess"
	($size, $mtime) = (stat "$opt{dir}/$user_number/$filename")[7,9] or die "Error: $!";

	##output it

	print qq(
	<tr valign="top">
	<td bgcolor="$vars_style{AltColumnColor2}" align="center" width="20"><input type="checkbox" name="$key" value="include" checked="checked" /></td>
	<td bgcolor="$vars_style{AltColumnColor2}"><b><font size="$vars_style{FDTextSize}">$filename</font></b></td>
	<td bgcolor="$vars_style{AltColumnColor2}" align="center" valign="middle">$size</td>
	<td bgcolor="$vars_style{AltColumnColor2}" align="center" valign="middle">$mtime</td>
	</tr>
	);
}

it is always better to see the actual code

yay, we're getting there:

https://www.ubbdev.com/ubbcgi/ubb_fileman.cgi

no problem... but you need to swap "Date:" "Size:" texts

and... I can give you a byte converter

Code

sub fsize {
   my $size = shift or return '0 byte';
   return sprintf("%1.2f %s",$size / 1024 / 1024 / 1024, " GB"  ) if($size >= 1024*1024*1024);
   return sprintf("%1.2f %s",$size / 1024 / 1024       , " MB"  ) if($size >= 1024*1024     );
   return sprintf("%1.2f %s",$size / 1024              , " KB"  ) if($size >= 1024          );
   return sprintf("%1.2f %s",$size                     , " Byte") if($size <  1024          );
   return "$size Byte";
}

so, you can use it like:

Code

opendir(UPLOADS, "$opt{dir}/$user_number") or die "Can not opendir($opt{dir}/$user_number): $!";
my @uploads = readdir(UPLOADS);
closedir(UPLOADS);
my($size, $mtime);
foreach my $filename (@uploads) {
        next if $filename =~ /^./; # ignore anything that starts with a dot. like "." ".." ".htaccess"
        next if -d "$opt{dir}/$user_number/$filename"; # ignore directories
	($size, $mtime) = (stat "$opt{dir}/$user_number/$filename")[7,9] or die "Error: $!";

	##output it
	$size = fsize($size);

	print qq(
	<tr valign="top">
	<td bgcolor="$vars_style{AltColumnColor2}" align="center" width="20"><input type="checkbox" name="$key" value="include" checked="checked" /></td>
	<td bgcolor="$vars_style{AltColumnColor2}"><b><font size="$vars_style{FDTextSize}">$filename</font></b></td>
	<td bgcolor="$vars_style{AltColumnColor2}" align="center" valign="middle">$size</td>
	<td bgcolor="$vars_style{AltColumnColor2}" align="center" valign="middle">$mtime</td>
	</tr>
	);
}

hey! I've just noticed the cake!

Happy Birthday!

grazi, I'm not getting older, I'm getting... well, ok, I'm getting older

May be this looks a bit more appropriate

my $str = "drwx------ 3 anpatel students 512 Apr 18 14:45 Really Bad Name Name for a file.";

$str =~ s/(S)s/$1/; #1 perms-links
$str =~ s/(S)s/$1/; #2 links-user
$str =~ s/(S)s/$1/; #3 user-grp
$str =~ s/(S)s/$1/; #4 grp-apr
$str =~ s/(S)s/$1/; #5 apr-date
$str =~ s/(S)s/$1/; #6 date-time
$str =~ s/(S)s/$1/; #7 time-filename
$str =~ s/^(S+)s+(.*)/$2/; # remove all the things that were combined previously. result is we get the file name!
print $str,"n";