I am looking for a command line solution to replace a potentially dangerous string in php files by a safer one, on a Linux server. I'd like to conduct the replace in all hosted sites, and then have it run as a daily cronjob to be on the safer side. Bash or PHP would be perfect I guess, but I'm ok with Python if that's more efficient. I could find many examples but could not adapt them to my case, which is
scan all /httpdocs/ subdirectories in /home/vhosts/(sitename)/ but scanning all files in /home/vhosts will do too.
determine if there's a file containing a multiple-lines code, ie
$authsites = array ( 'flickr.com', 'picasa.com', 'blah.com', );
and if so replace it (sed ?) with
$authsites = array ();
and log where it replaced it.
Help greatly appreciated, thanks.
i use command line perl for bulk search and replace
perl -pi -e 's/FIND IT/REPLACE IT' /home/name/public_html/html/*.html
How about this awk one-liner. You can use it with find or for loop to do mass substitution.
awk '/\$authsites/{sub(/\(.*\)/,"( )");print;next}1' INPUTFILE
Test:
[jaypal:~/Temp] cat file00
$authsites = array ( 'flickr.com', 'picasa.com', 'blah.com', );
$authsites = array ( 'flickr.com', 'aaaa.com', 'blah.com', );
fefe
efef
$authsites = array ( 'flickr.com', 'picasa.com', );
[jaypal:~/Temp] awk '/\$authsites/{sub(/\(.*\)/,"( )");print;next}1' file00
$authsites = array ( );
$authsites = array ( );
fefe
efef
$authsites = array ( );
A for loop like this could work (though there may be better ways)
You can change the name according to your files
for i in $(find . -type f -name "file*" -exec grep -H "\$authsites" {} \;| cut -d":" -f1 | uniq); do awk '/\$authsites/{sub(/\(.*\)/,"( )");print;next}1' $i > $i.new; done