Replace hot Tokenizer loop with a mathematical expression

This piece of code is very hot, executed thousands, possibly millions of times.

I found this when I benchmarked our custom sniffs. The replaceTabsInToken() method is consistently reported as one of the absolute hottest ones. I looked at the method and was wondering why this particular piece is a loop when it can be a mathematical expression. According to my micro-benchmark (only measuring the changed piece) it's about twice as fast now. This will probably not have much of an effect in a real-world scenario. Still I find it worth it.

Author: thiemowmde

src/Tokenizers/Tokenizer.php

@@ -642,21 +642,9 @@ public function replaceTabsInToken(&$token, $prefix=' ', $padding=' ', $tabWidth
                 $tabNum++;
 
                 // Move the pointer to the next tab stop.
-                if (($currColumn % $tabWidth) === 0) {
-                    // This is the first tab, and we are already at a
-                    // tab stop, so this tab counts as a single space.
-                    $currColumn++;
-                } else {
-                    $currColumn++;
-                    while (($currColumn % $tabWidth) !== 0) {
-                        $currColumn++;
-                    }
-
-                    $currColumn++;
-                }
-
-                $length     += ($currColumn - $lastCurrColumn);
-                $newContent .= $prefix.str_repeat($padding, ($currColumn - $lastCurrColumn - 1));
+                $length      = $tabWidth - ($currColumn + $tabWidth - 1) % $tabWidth;
+	            $currColumn += $length;
+                $newContent .= $prefix.str_repeat($padding, ($length - 1));
             }//end foreach
         }//end if