Fixes/comments for guessBestEncoding (#304)

* tests: add case for getRecommendedEncoding

* vmime: avoid integer multiply wraparound in wordEncoder::guessBestEncoding

If the input string is 42949673 characters long or larger, there will
be integer overflow on 32-bit platforms when multiplying by 100.
Switch that one computation to floating point.

* vmime: update comment in wordEncoder::guessBestEncoding

3 files changed, 27 insertions, 4 deletions

diff --git a/src/vmime/wordEncoder.cpp b/src/vmime/wordEncoder.cpp
index 4f47d047..d1632022 100644
--- a/src/vmime/wordEncoder.cpp
+++ b/src/vmime/wordEncoder.cpp
@@ -302,14 +302,16 @@ wordEncoder::Encoding wordEncoder::guessBestEncoding(
 		}
 	}
 
-	// Use Base64 if more than 40% non-ASCII, or Quoted-Printable else (default)
+	// Base64 would be more space-efficient when the ASCII content is
+	// below 83.33%, but QP has a legibility arugment going for it, so we
+	// picked 60%.
 	const size_t asciiCount =
 		utility::stringUtils::countASCIIchars(buffer.begin(), buffer.end());
 
-	const size_t asciiPercent =
-		buffer.length() == 0 ? 100 : (100 * asciiCount) / buffer.length();
+	const double asciiPercent =
+		buffer.length() == 0 ? 100 : static_cast<double>(asciiCount) / buffer.length();
 
-	if (asciiPercent < 60) {
+	if (asciiPercent < 0.60) {
 		return ENCODING_B64;
 	} else {
 		return ENCODING_QP;
diff --git a/tests/parser/textTest.cpp b/tests/parser/textTest.cpp
index e9ddc262..f5111e79 100644
--- a/tests/parser/textTest.cpp
+++ b/tests/parser/textTest.cpp
@@ -32,6 +32,7 @@ VMIME_TEST_SUITE_BEGIN(textTest)
 	VMIME_TEST_LIST_BEGIN
 		VMIME_TEST(testConstructors)
 		VMIME_TEST(testCopy)
+		VMIME_TEST(testRecommendedEncoding)
 		VMIME_TEST(testNewFromString)
 		VMIME_TEST(testDisplayForm)
 		VMIME_TEST(testParse)
@@ -155,6 +156,17 @@ VMIME_TEST_SUITE_BEGIN(textTest)
 		VASSERT("copyFrom", t1 == t2);
 	}
 
+	void testRecommendedEncoding() {
+
+		vmime::encoding enc;
+		VASSERT_TRUE("1.1", vmime::charset("iso8859-1").getRecommendedEncoding(enc));
+		VASSERT_TRUE("1.2", enc.getName() == vmime::encodingTypes::QUOTED_PRINTABLE);
+		VASSERT_TRUE("1.2", vmime::charset("iso8859-2").getRecommendedEncoding(enc));
+		VASSERT_TRUE("1.3", enc.getName() == vmime::encodingTypes::QUOTED_PRINTABLE);
+		VASSERT_FALSE("1.3", vmime::charset(vmime::charsets::UTF_8).getRecommendedEncoding(enc));
+
+	}
+
 	void testNewFromString() {
 
 		vmime::string s1 = "only ASCII characters";
diff --git a/tests/parser/wordEncoderTest.cpp b/tests/parser/wordEncoderTest.cpp
index 08d34aa4..0c827bff 100644
--- a/tests/parser/wordEncoderTest.cpp
+++ b/tests/parser/wordEncoderTest.cpp
@@ -36,6 +36,7 @@ VMIME_TEST_SUITE_BEGIN(wordEncoderTest)
 		VMIME_TEST(testIsEncodingNeeded_specialChars)
 		VMIME_TEST(testGuessBestEncoding_QP)
 		VMIME_TEST(testGuessBestEncoding_B64)
+		VMIME_TEST(testGuessBestEncoding_size)
 		VMIME_TEST(testEncodeQP_RFC2047)
 	VMIME_TEST_LIST_END
 
@@ -159,6 +160,14 @@ VMIME_TEST_SUITE_BEGIN(wordEncoderTest)
 		);
 	}
 
+	void testGuessBestEncoding_size() {
+
+		std::string i(42949673, 'a');
+		auto enc = vmime::wordEncoder::guessBestEncoding(i, vmime::charset("utf-8"));
+		VASSERT_EQ("1", enc, vmime::wordEncoder::ENCODING_QP);
+
+	}
+
 	void testEncodeQP_RFC2047() {
 
 		// When Quoted-Printable is used, it should be RFC-2047 QP encoding