pangram: fix test for duplicate mixed-case chars (#852)

From what I can see, the last test case is intended to catch incorrect
programs such as this, which don't take into account different-cased
duplicates:

```haskell
isPangram = (== 26) . length . nub . filter isAlpha
```

However the above program passed the final test as the count of distinct
upper and lower case characters is 27, rather than 26.

This commit changes the test text to a non-pangram that has exactly 26
distinct upper and lower case characters, which causes the above
incorrect program to fail.

exercises/pangram/canonical-data.json

@@ -3,7 +3,7 @@
   "comments": [
     "A pangram is a sentence using every letter of the alphabet at least once."
   ],
-  "version": "1.0.0",
+  "version": "1.1.0",
   "cases": [
     {
       "description": "Check if the given string is an pangram",
@@ -62,7 +62,7 @@
         {
           "description": "upper and lower case versions of the same character should not be counted separately",
           "property": "isPangram",
-          "input": "the quick brown fox jumped over the lazy FOX",
+          "input": "the quick brown fox jumps over with lazy FX",
           "expected": false
         }
       ]