Support cmaps with only CID characters, when building the ToUnicode-map (issue 9367)

In this particular case the `CMap`-data that we create contains only numbers, but no strings, which causes `PartialEvaluator.readToUnicode` to create a ToUnicode-map with only empty strings.

*Please note:* This is yet another case where I don't know if it's necessarily the best and most correct solution, but it does fix the referenced issue.

src/core/evaluator.js

@@ -3455,6 +3455,11 @@ class PartialEvaluator {
           // NOTE: cmap can be a sparse array, so use forEach instead of
           // `for(;;)` to iterate over all keys.
           cmap.forEach(function (charCode, token) {
+            // Some cmaps contain *only* CID characters (fixes issue9367.pdf).
+            if (typeof token === "number") {
+              map[charCode] = String.fromCodePoint(token);
+              return;
+            }
             const str = [];
             for (let k = 0; k < token.length; k += 2) {
               const w1 = (token.charCodeAt(k) << 8) | token.charCodeAt(k + 1);