find / grep to list files with php's short_open_tag

[Reindl Harald]

am i stupid or grep? :-)

find /data/www/ -type f -name *.php -exec grep -l -i -n -F "<?" '{}' ;
lists any file which contains  "<?" under the path /data/www/

find /data/www/ -type f -name *.php -exec grep -l -i -n -F "<?\n" '{}' ;
is expected to find any file with "<?" followed directly by a linebreak
but that does not work and if someone finds google empty it was me :-(

[Patrick O'Callaghan]

On Tue, 2013-05-07 at 19:41 +0200, Reindl Harald wrote:
> am i stupid or grep? :-)

Well I know you aren't grep ...
> find /data/www/ -type f -name *.php -exec grep -l -i -n -F "<?" '{}' ;
> lists any file which contains  "<?" under the path /data/www/
>
> find /data/www/ -type f -name *.php -exec grep -l -i -n -F "<?\n" '{}' ;
> is expected to find any file with "<?" followed directly by a linebreak
> but that does not work and if someone finds google empty it was me :-(

Not sure if "\n" is allowed, but why not use $ for end of line as in
most regular expressions?

poc

[Suvayu Ali]

On Wed, May 08, 2013 at 11:23:34PM -0430, Patrick O'Callaghan wrote:
> On Tue, 2013-05-07 at 19:41 +0200, Reindl Harald wrote:
> > find /data/www/ -type f -name *.php -exec grep -l -i -n -F "<?\n" '{}' ;
> > is expected to find any file with "<?" followed directly by a linebreak
> > but that does not work and if someone finds google empty it was me :-(
>
> Not sure if "\n" is allowed, but why not use $ for end of line as in
> most regular expressions?

Reindl is using fgrep (implied by the -F flag), so I would think \n is
interpreted literally.  I would agree with your suggestion to use

$ grep -E "<?$"

Another recommendation would be to use -print0 and pipe to xargs -0.
Using the -exec flag for something like this might be significantly less
efficient.

[Reindl Harald]

Am 09.05.2013 05:53, schrieb Patrick O'Callaghan:
> On Tue, 2013-05-07 at 19:41 +0200, Reindl Harald wrote:
> > find /data/www/ -type f -name *.php -exec grep -l -i -n -F "<?" '{}' ;
> > lists any file which contains  "<?" under the path /data/www/
> >
> > find /data/www/ -type f -name *.php -exec grep -l -i -n -F "<?\n" '{}' ;
> > is expected to find any file with "<?" followed directly by a linebreak
> > but that does not work and if someone finds google empty it was me :-(
>
> Not sure if "\n" is allowed, but why not use $ for end of line as in
> most regular expressions?

because it does not work and i have no idea how to "mask" the LF for grep

[harry@srv-rhsoft:~]$ find /mnt/data/www/ -type f -name *.php -exec grep -l -i -n -F "<?$" '{}' ;
[harry@srv-rhsoft:~]$

as you see nothing found

[harry@srv-rhsoft:~]$ cat /mnt/data/www/short-open.php
<?
 echo 'TEST';
?>

as you see there is at least one file

[Ed Greshko]

On 05/09/13 21:05, Reindl Harald wrote:
> Am 09.05.2013 05:53, schrieb Patrick O'Callaghan:
> > On Tue, 2013-05-07 at 19:41 +0200, Reindl Harald wrote:
> > > find /data/www/ -type f -name *.php -exec grep -l -i -n -F "<?" '{}' ;
> > > lists any file which contains  "<?" under the path /data/www/
> > >
> > > find /data/www/ -type f -name *.php -exec grep -l -i -n -F "<?\n" '{}' ;
> > > is expected to find any file with "<?" followed directly by a linebreak
> > > but that does not work and if someone finds google empty it was me :-(
> >
> > Not sure if "\n" is allowed, but why not use $ for end of line as in
> > most regular expressions?
>
> because it does not work and i have no idea how to "mask" the LF for grep
>
> [harry@srv-rhsoft:~]$ find /mnt/data/www/ -type f -name *.php -exec grep -l -i -n -F "<?$" '{}' ;
> [harry@srv-rhsoft:~]$
>
> as you see nothing found
>
> [harry@srv-rhsoft:~]$ cat /mnt/data/www/short-open.php
>
> <?
>  echo 'TEST';
> ?>
>
> as you see there is at least one file

-G seems to work....  Got a splitting headache, so not thinking of the implications....

[egreshko@meimei try]$ cat one.php
<?
 echo 'TEST';
?>
[egreshko@meimei try]$ cat two.php
<?xxxx
 echo 'TEST';
?>xxxx

[egreshko@meimei try]$ find /tmp/try -type f -name *.php -exec grep -l -i -n -G "<?$" '{}' ;
/tmp/try/one.php

[David Beveridge]

> [egreshko@meimei try]$ find /tmp/try -type f -name *.php -exec grep -l -i -n -G "<?$" '{}' ;
> /tmp/try/one.php
>
> --
> The only thing worse than a poorly asked question is a cryptic answer.

So you want all the <? but not the <?php, <?xml, <?= in text files?

grep -r "<?" * | grep -v "Binary file" | grep -v "<?php" | grep -v
"<?xml" | grep -v "<?="

[Suvayu Ali]

Hi Reindl,

On Thu, May 09, 2013 at 03:05:01PM +0200, Reindl Harald wrote:
> because it does not work and i have no idea how to "mask" the LF for grep
>
> [harry@srv-rhsoft:~]$ find /mnt/data/www/ -type f -name *.php -exec grep -l -i -n -F "<?$" '{}' ;

Did you look at my response earlier?

http://mid.gmane.org/[e-mail]

Hope this helps,

[Reindl Harald]

Am 10.05.2013 01:52, schrieb David Beveridge:
> > [egreshko@meimei try]$ find /tmp/try -type f -name *.php -exec grep -l -i -n -G "<?$" '{}' ;
> > /tmp/try/one.php
> >
> > The only thing worse than a poorly asked question is a cryptic answer.
>
> So you want all the <? but not the <?php, <?xml, <?= in text files?
>
> grep -r "<?" * | grep -v "Binary file" | grep -v "<?php" | grep -v
> "<?xml" | grep -v "<?="

i finally solved it with a php-script itself using a years ago written
PHP-SPL-wrapper for recursive listing of files started in the dir the
php-cli-script is living

relevant is <? followed by space, LF, CR, TAB to catch also files with
windows or OSX line-breaks

you must not list here files containing value="<?=(int)$id?>" because this
short echo-variant is starting with PHP 5.4 always supported and no longer
in context with "short_open_tag" from php.ini

well, i found 7 files out of some thousand written in the last 10 years :-)

script and 3 test-files below
__________________________________________________________________________________

#!/usr/bin/php
<?php
 if(PHP_SAPI != 'cli')
 {
  exit('Forbidden');
 }
 require('global.inc.php');
 $fs = new rh_filesystem();
 $liste = $fs->ListFilesRecursive(/**$path*/__DIR__, /**$details*/false, /**$excludes*/array());
 foreach($liste as $file)
 {
  if($file != __FILE__ && strtolower(pathinfo($file, PATHINFO_EXTENSION)) == 'php' && is_readable($file))
  {
   $content = file_get_contents($file);
   if(strpos($content, '<? ') !== false || strpos($content, "<?\r") !== false || strpos($content, "<?\n") !== false
|| strpos($content, "<?\t") !== false)
   {
    echo $file . MY_LE;
    flush();
   }
  }
 }
?>

__________________________________________________________________________________

$fs->ListFilesRecursive() is basically this:

$spl_objects = new RecursiveIteratorIterator(new RecursiveDirectoryIterator($path));
     try
     {
      foreach($spl_objects as $filename=>$spl_object)
      {
       if(!$spl_object->isDir())
       {
        if(DIRECTORY_SEPARATOR === "\")
        {
         $filename = str_replace(DIRECTORY_SEPARATOR, '/', $filename);
        }
        if(substr($filename, 0, 2) === './')
        {
         $filename = substr($filename, 2);
        }
        if(empty($pattern) || !preg_match($pattern, $filename))
        {
         if(!$ignore_svn || strpos($filename, '/.svn') === false)
         {
          $enumerate_list[] = $filename;
         }
        }
       }
      }
     }
     catch(UnexpectedValueException $e)
     {
      error_log('Directory "' . $path . '" contained a directory we can not recurse into');
      return false;
     }
     sort($enumerate_list);

__________________________________________________________________________________

[harry@srv-rhsoft:/www]$ cat test-short-open-1.php
<?
 echo 'TEST';
?>
__________________________________________________________________________________

[harry@srv-rhsoft:/www]$ cat test-short-open-2.php
<?php
 echo 'TEST';
?>
<form action="<? echo "TEST"?>"></form>
__________________________________________________________________________________

[harry@srv-rhsoft:/www]$ cat test-short-open-3.php
<?
 echo 'TEST';
?>
<td>
 <?
  echo 'TEST';
 ?>
</td>
__________________________________________________________________________________