poll results

[ravi]

Michael Pollak wrote:
> On Tue, 19 Mar 2002, ravi wrote:
> 
>>yes, but it is trivially easy to set the archive up so only those users
>>who want their email addresses obscured will have that happen.
> 
> Is it trivially easy to do that retroactively, to documents that have
> already been archived?
> 

yes, that too. lets take a random shot (which will definitely not run
without mods and is probably not written too well either):

#!/opt/perl/bin/perl
# substitute location of perl above
#
# perl script to obscure email addresses in archives
# assumes each message is stored in a separate file
# de-obfuscated for readability
#

#--- configurable stuff ---

$archivedir="...";	# place where hypermail archives are
$invisiblefile="...";
# list of email addresses to anonymize

#--- end configurab stuff ---

open(INV, "$invisiblefile")

	|| die "Could not open inv members file: $invisiblefile.\n";

while( <INV> )
{

	chop;

	$inv{$_} = 1;
}
close(INV);

opendir(ARCHIVES, "$archivedir")

	|| die "Could not open archivedir: $archivedir.\n";

# who needs . and ..
readdir(ARCHIVES); readdir(ARCHIVES);

while( $file = readdir(ARCHIVES) )
{

	# there are much better ways to do the below! but as we know

	# there is no one right way to do anything in perl ;-)

	open(MSG, "$file")

		|| die "Could not open: $file.\n";

	open(TMP, "> /tmp/inv.$$")

		|| die "Could not open tmp file: /tmp/inv.$$\n";

	while( <MSG> )

	{

		# rather than be clever and look for hypermail HTML

		# tags for email addresses, lets just do brute force.

		# regexp below to get email address needs some work.

		# also email address should be replaced with something

		# more interesting than just "obscured"

		s/$1/ email obscured /g

			if( /([\w_\-\.]+\@[\w\-]+(\.[\w\-]+)+)/

				&& $inv{"$1"} );

		print TMP;

	}

	close(MSG);

	close(TMP);

	# definitely can do this a better way

	system("mv /tmp/inv.$$ $archivedir/$file");
}

closedir(ARCHIVES);

---------------------------

	--ravi