Update to Unicode 16.0.0

README.md

@@ -1,6 +1,6 @@
 # Upa IDNA
 
-Upa IDNA is the [Unicode IDNA Compatibility Processing (UTS #46)](https://www.unicode.org/reports/tr46/) C++ library. It is compliant with the latest 15.1.0 version of the Unicode standard.
+Upa IDNA is the [Unicode IDNA Compatibility Processing (UTS #46)](https://www.unicode.org/reports/tr46/) C++ library. It is compliant with the 16.0.0 version of the Unicode standard.
 
 This library implements two functions from [UTS #46](https://www.unicode.org/reports/tr46/): [`to_ascii`](https://www.unicode.org/reports/tr46/#ToASCII) and [`to_unicode`](https://www.unicode.org/reports/tr46/#ToUnicode), and two functions from the [WHATWG URL Standard](https://url.spec.whatwg.org/): [`domain_to_ascii`](https://url.spec.whatwg.org/#concept-domain-to-ascii) and [`domain_to_unicode`](https://url.spec.whatwg.org/#concept-domain-to-unicode). It has no dependencies and requires C++11 or later.
 

include/upa/idna/idna.h

@@ -9,7 +9,6 @@
 #include "idna_table.h"
 #include "iterate_utf.h"
 #include "nfc.h"
-#include <algorithm>
 #include <string>
 #include <type_traits> // std::make_unsigned
 
@@ -119,13 +118,11 @@ inline bool map(std::u32string& mapped, const CharT* input, const CharT* input_e
                 }
                 break;
             default:
-                // CP_DISALLOWED
-                // CP_NO_STD3_MAPPED, CP_NO_STD3_VALID if Option::UseSTD3ASCIIRules
-                // Starting with Unicode 15.1.0 - don't record an error
+                // CP_DISALLOWED or
+                // CP_NO_STD3_VALID if Option::UseSTD3ASCIIRules
+                // Starting with Unicode 15.1.0 don't record an error
                 if (is_to_ascii && // to_ascii optimization
-                    ((value & util::CP_DISALLOWED_STD3) == 0
-                    ? !std::binary_search(std::begin(util::comp_disallowed), std::end(util::comp_disallowed), cp)
-                    : !std::binary_search(std::begin(util::comp_disallowed_std3), std::end(util::comp_disallowed_std3), cp)))
+                    ((value & util::CP_DISALLOWED_STD3) == 0 || cp > 0x3E || cp < 0x3C))
                     return false;
                 mapped.push_back(cp);
                 break;

include/upa/idna/idna_table.h

@@ -24,7 +24,6 @@ const std::uint32_t CP_MAPPED = 0x0002 << 16;
 const std::uint32_t CP_DEVIATION = CP_VALID | CP_MAPPED; // 0x0003 << 16
 const std::uint32_t CP_DISALLOWED_STD3 = 0x0004 << 16;
 const std::uint32_t CP_NO_STD3_VALID = CP_VALID | CP_DISALLOWED_STD3;
-const std::uint32_t CP_NO_STD3_MAPPED = CP_MAPPED | CP_DISALLOWED_STD3;
 const std::uint32_t MAP_TO_ONE = 0x0008 << 16;
 // General_Category=Mark
 const std::uint32_t CAT_MARK = 0x0010 << 16;
@@ -57,8 +56,7 @@ extern const std::uint32_t blockData[];
 extern const std::uint16_t blockIndex[];
 extern const char32_t allCharsTo[];
 
-extern const std::uint32_t comp_disallowed[5];
-extern const std::uint32_t comp_disallowed_std3[21];
+extern const std::uint8_t comp_disallowed_std3[3];
 
 extern const std::uint8_t asciiData[128];
 // END-GENERATED

src/idna.cpp

@@ -54,9 +54,11 @@ bool processing_mapped(std::u32string* pdecoded, const std::u32string& mapped, O
         // P4 - Convert/Validate
         if (label_end - label >= 4 && label[0] == 'x' && label[1] == 'n' && label[2] == '-' && label[3] == '-') {
             if (*(label_end - 1) == '-' && label_end - label != 5) {
-                // For compatibility with ICU, report errors on "xn--" or "xn--ascii-" labels.
-                // Ignore "xn---", it will fail punycode::decode.
-                // More info: https://github.com/whatwg/url/issues/760#issuecomment-1462706617
+                // > 4. Processing - 4. - 3. If (after Punycode decode) the label is empty, or if the label
+                // > contains only ASCII code points, record that there was an error.
+                // 1) "xn--" is decoded to empty label
+                // 2) "xn--ascii-" is decoded to "ascii"
+                // Note: "xn---" is ignored here, because it will fail punycode::decode
                 error = true;
                 // Decode "xn--ascii-" to "ascii" for to_unicode:
                 if (pdecoded && label_end - label > 5) {
@@ -303,7 +305,6 @@ bool to_ascii_mapped(std::string& domain, const std::u32string& mapped, Option o
 
     // A2 - Break the result into labels at U+002E FULL STOP
     if (mapped.length() == 0) {
-        // to simplify root label detection
         if (detail::has(options, Option::VerifyDnsLength))
             ok = false;
     } else {
@@ -312,9 +313,6 @@ bool to_ascii_mapped(std::string& domain, const std::u32string& mapped, Option o
         std::size_t domain_len = domain.length() + static_cast<std::size_t>(-1);
         bool first_label = true;
         split(first, last, 0x002E, [&](const char32_t* label, const char32_t* label_end) {
-            // root is ending empty label
-            const bool is_root = (label == last);
-
             // join
             if (first_label) {
                 first_label = false;
@@ -342,7 +340,7 @@ bool to_ascii_mapped(std::string& domain, const std::u32string& mapped, Option o
             }
 
             // A4 - DNS length restrictions
-            if (detail::has(options, Option::VerifyDnsLength) && !is_root) {
+            if (detail::has(options, Option::VerifyDnsLength)) {
                 const std::size_t label_length = domain.length() - label_start_ind;
                 // A4_2
                 if (label_length < 1 || label_length > 63)