Closes #2054: Fix iconv encoding for multi-byte characters and reenab…

…le tests (#2055)

* Fix iconv encoding for multi-byte characters
In the initial encoding implementation, the output buffer for
calculating the buffer length was a `bytes` buffer and that
was causing issues when an output character was a multi-byte
UTF-8 character. To fix this, the `bytes` buffer has been
switched to a `uint-8` array.

* Re-enable IDNA tests

src/Codecs.chpl

@@ -98,7 +98,7 @@ module Codecs {
         var inSize = (inBufSize): c_size_t;
 
         // TODO: this is probably worst way to allocate this
-        var chplRes:bytes = (" "*(inBufSize*4));
+        var chplRes: [0..#(inBufSize*4)] uint(8) = 0x00;
         var origSize = chplRes.size;
         var outSize = chplRes.size: c_size_t;
 

tests/string_test.py

@@ -639,7 +639,6 @@ def test_get_fixes(self):
         p = strings.get_suffixes(1, return_origins=False, proper=False)
         self.assertListEqual(["c", "d", "i"], p.to_list())
 
-    @pytest.mark.skip(reason="Skipping until issue resolved.")
     def test_encoding(self):
         idna_strings = ak.array(['Bücher.example','ドメイン.テスト', 'домен.испытание', 'Königsgäßchen'])
         expected = ak.array(['xn--bcher-kva.example','xn--eckwd4c7c.xn--zckzah', 'xn--d1acufc.xn--80akhbyknj4f', 'xn--knigsgchen-b4a3dun'])
@@ -666,7 +665,6 @@ def test_encoding(self):
         a3 = ak.random_strings_uniform(1, 10, UNIQUE, characters="printable")
         self.assertTrue((a3 == a3.encode('ascii').decode('ascii')).all())
 
-    @pytest.mark.skip(reason="Skipping until issue resolved.")
     def test_idna_utf16(self):
         s = ak.array(['xn--mnchen-3ya', 'xn--zrich-kva', 'example.com'])