Improve floating point scaling in UnitConverter.

When precision dictates that a number be zero padded, we cannot give sprintf()
a negative precision specifier.  This commit implements manual negative precision
printing of floats, taking into account common rounding errors with floating
point numbers.

Signed-off-by: Edward Z. Yang <edwardzyang@thewritingpot.com>

tests/HTMLPurifier/UnitConverterTest.php

@@ -94,6 +94,10 @@ class HTMLPurifier_UnitConverterTest extends HTMLPurifier_Harness
         $this->assertConversion('11.112pt', '0.15433in');
     }
     
+    function __onlytestRoundingBigNumber() {
+        $this->assertConversion('444400000000000000000000in', '42660000000000000000000000px');
+    }
+    
     protected function assertSigFig($n, $sigfigs) {
         $converter = new HTMLPurifier_UnitConverter();
         $result = $converter->getSigFigs($n);