Here's some code to detect a few code smells in a PHP file.  It can be easily adapted to detect code smells in C, C++, Java, etc. code.



 $MAX_INDENTATION_LEVEL = 3;
 $MAX_FUNCTION_LINE_COUNT = 20;
 $MAX_NUM_PARAMETERS = 4;



 function print_warning( $filename, $line, $line_number, $warning )
   {
   print( "<code>$filename</code>: <strong>$warning</strong><br />\n" .
          "&nbsp; &nbsp; Line $line_number: <code>$line</code><br />\n" );
   }



 function test_line_for_indentation( $line, $max_indentation_level )
   {
   $char_posit = 0;
   while( substr( $line, $char_posit, 1 ) == "\t" )
     $char_posit++;
   if( $char_posit > $max_indentation_level )
     return false;
   return true;
   }



 function test_line_for_function_definition( $line, $max_num_parameters )
   {
   if( substr( $line, 0, 9 ) == "function " )
     {
     if( substr_count( $line, ',' ) >= $max_num_parameters )
       return false;
     }
     return true;
   }



 function test_file_for_smells( $filename )
   {
   global $MAX_FUNCTION_LINE_COUNT, $MAX_INDENTATION_LEVEL;
   $function_line_count = 0;
   $found_continued_line = 0;
   $line_number = 0;
   $fp = fopen( $filename, "r" );
   if( ! $fp )
     return;
   while( ! feof( $fp ) )
     {
     $line = chop( fgets( $fp, 4096 ) );
     $line_number++;
     while( substr( $line, 0, 9 ) == "function "  &&
            substr( $line, strlen($line) - 1, 1 ) != ')' )
       {
       $line = $line . chop( fgets( $fp, 4096 ) )
       $found_continued_line++;
       }
     if( substr( $line, 0, 9 ) == "function " )
       {
       $function_line_count = 1;
       $function_name = substr( $line, 9, strpos($line, '(') - 9 );
       }
     if( $function_line_count > 0 )
       {
       if( $line == "\t}" )
         {
         $function_line_count = $function_line_count - 3;
         if( $function_line_count > $MAX_FUNCTION_LINE_COUNT )
           print_warning( $filename, $line, $line_number, "$function_name() too long ($function_line_count lines)" );
         $function_line_count = 0;
         }
       else
         $function_line_count++;
       $result = test_line_for_indentation( $line, $MAX_INDENTATION_LEVEL );
       if( ! $result )
         print_warning( $filename, $line, $line_number, "Too much nesting." );
       $result = test_line_for_function_definition( $line, $MAX_NUM_PARAMETERS );
       if( ! $result )
         print_warning( $filename, $line, $line_number, "Too many parameters." );
       if( $found_continued_line > 0 )
         {
         $line_number += $found_continued_line;
         $found_continued_line = 0;
         }
       }
     }
   }

Please feel free to ReFactor this code.  test_file_for_smells() would fail its own tests! -- BrentNewhall

''Doesn't this assume tabs are always used for indentation?''

This code probably needs to be refactored to use the tokenizer extension.  I have a "CodeSmeller" object written in PHP5, I will post it soon.  -- JonathanArkell
----
CategoryPhp