michael@0: /* -*- Mode: C++; tab-width: 2; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
michael@0: /* This Source Code Form is subject to the terms of the Mozilla Public
michael@0:  * License, v. 2.0. If a copy of the MPL was not distributed with this
michael@0:  * file, You can obtain one at http://mozilla.org/MPL/2.0/. */
michael@0: 
michael@0: #include <stdio.h>
michael@0: #include "nsXPCOM.h"
michael@0: #include "nsIUnicodeNormalizer.h"
michael@0: #include "nsStringAPI.h"
michael@0: #include "nsCharTraits.h"
michael@0: #include "nsServiceManagerUtils.h"
michael@0: 
michael@0: struct testcaseLine {
michael@0:   wchar_t* c1;
michael@0:   wchar_t* c2;
michael@0:   wchar_t* c3;
michael@0:   wchar_t* c4;
michael@0:   wchar_t* c5;
michael@0:   char* description;
michael@0: };
michael@0: 
michael@0: #ifdef DEBUG_smontagu
michael@0: #define DEBUG_NAMED_TESTCASE(t, s) \
michael@0:   printf(t ": "); \
michael@0:   for (uint32_t i = 0; i < s.Length(); ++i) \
michael@0:     printf("%x ", s.CharAt(i)); \
michael@0:   printf("\n")
michael@0: #else
michael@0: #define DEBUG_NAMED_TESTCASE(t, s)
michael@0: #endif
michael@0: 
michael@0: #define DEBUG_TESTCASE(x) DEBUG_NAMED_TESTCASE(#x, x)
michael@0: 
michael@0: #define NORMALIZE_AND_COMPARE(base, comparison, form, description) \
michael@0:    normalized.Truncate();\
michael@0:    normalizer->NormalizeUnicode##form(comparison, normalized);\
michael@0:    DEBUG_NAMED_TESTCASE(#form "(" #comparison ")", normalized);\
michael@0:    if (!base.Equals(normalized)) {\
michael@0:      rv = false;\
michael@0:      showError(description, #base " != " #form "(" #comparison ")\n");\
michael@0:    }
michael@0: 
michael@0: NS_DEFINE_CID(kUnicodeNormalizerCID, NS_UNICODE_NORMALIZER_CID);
michael@0: 
michael@0: nsIUnicodeNormalizer *normalizer;
michael@0: bool verboseMode = false;
michael@0: 
michael@0: #include "NormalizationData.h"
michael@0: 
michael@0: void showError(const char* description, const char* errorText)
michael@0: {
michael@0:   if (verboseMode)
michael@0:     printf("%s failed: %s", description, errorText);
michael@0: }
michael@0: 
michael@0: bool TestInvariants(testcaseLine* testLine)
michael@0: {
michael@0:   nsAutoString c1, c2, c3, c4, c5, normalized;
michael@0:   c1 = nsDependentString((char16_t*)testLine->c1);
michael@0:   c2 = nsDependentString((char16_t*)testLine->c2);
michael@0:   c3 = nsDependentString((char16_t*)testLine->c3);
michael@0:   c4 = nsDependentString((char16_t*)testLine->c4);
michael@0:   c5 = nsDependentString((char16_t*)testLine->c5);
michael@0:   bool rv = true;
michael@0:  
michael@0:   /*
michael@0:     1. The following invariants must be true for all conformant implementations
michael@0: 
michael@0:     NFC
michael@0:       c2 ==  NFC(c1) ==  NFC(c2) ==  NFC(c3)
michael@0:   */
michael@0:   DEBUG_TESTCASE(c2);
michael@0:   NORMALIZE_AND_COMPARE(c2, c1, NFC, testLine->description);  
michael@0:   NORMALIZE_AND_COMPARE(c2, c2, NFC, testLine->description);
michael@0:   NORMALIZE_AND_COMPARE(c2, c3, NFC, testLine->description);
michael@0: 
michael@0:   /*
michael@0:       c4 ==  NFC(c4) ==  NFC(c5)
michael@0:   */
michael@0:   DEBUG_TESTCASE(c4);
michael@0:   NORMALIZE_AND_COMPARE(c4, c4, NFC, testLine->description);
michael@0:   NORMALIZE_AND_COMPARE(c4, c5, NFC, testLine->description);
michael@0: 
michael@0:   /*
michael@0:     NFD
michael@0:       c3 ==  NFD(c1) ==  NFD(c2) ==  NFD(c3)
michael@0:   */
michael@0:   DEBUG_TESTCASE(c3);
michael@0:   NORMALIZE_AND_COMPARE(c3, c1, NFD, testLine->description);
michael@0:   NORMALIZE_AND_COMPARE(c3, c2, NFD, testLine->description);
michael@0:   NORMALIZE_AND_COMPARE(c3, c3, NFD, testLine->description);
michael@0:   /*
michael@0:       c5 ==  NFD(c4) ==  NFD(c5)
michael@0:   */
michael@0:   DEBUG_TESTCASE(c5);
michael@0:   NORMALIZE_AND_COMPARE(c5, c4, NFD, testLine->description);
michael@0:   NORMALIZE_AND_COMPARE(c5, c5, NFD, testLine->description);
michael@0: 
michael@0:   /*
michael@0:     NFKC
michael@0:       c4 == NFKC(c1) == NFKC(c2) == NFKC(c3) == NFKC(c4) == NFKC(c5)
michael@0:   */
michael@0:   DEBUG_TESTCASE(c4);
michael@0:   NORMALIZE_AND_COMPARE(c4, c1, NFKC, testLine->description);
michael@0:   NORMALIZE_AND_COMPARE(c4, c2, NFKC, testLine->description);
michael@0:   NORMALIZE_AND_COMPARE(c4, c3, NFKC, testLine->description);
michael@0:   NORMALIZE_AND_COMPARE(c4, c4, NFKC, testLine->description);
michael@0:   NORMALIZE_AND_COMPARE(c4, c5, NFKC, testLine->description);
michael@0: 
michael@0:   /*
michael@0:     NFKD
michael@0:       c5 == NFKD(c1) == NFKD(c2) == NFKD(c3) == NFKD(c4) == NFKD(c5)
michael@0:   */
michael@0:   DEBUG_TESTCASE(c5);
michael@0:   NORMALIZE_AND_COMPARE(c5, c1, NFKD, testLine->description);
michael@0:   NORMALIZE_AND_COMPARE(c5, c2, NFKD, testLine->description);
michael@0:   NORMALIZE_AND_COMPARE(c5, c3, NFKD, testLine->description);
michael@0:   NORMALIZE_AND_COMPARE(c5, c4, NFKD, testLine->description);
michael@0:   NORMALIZE_AND_COMPARE(c5, c5, NFKD, testLine->description);
michael@0: 
michael@0:   return rv;
michael@0: }
michael@0: 
michael@0: uint32_t UTF32CodepointFromTestcase(testcaseLine* testLine)
michael@0: {
michael@0:   if (!IS_SURROGATE(testLine->c1[0]))
michael@0:     return testLine->c1[0];
michael@0: 
michael@0:   NS_ASSERTION(NS_IS_HIGH_SURROGATE(testLine->c1[0]) &&
michael@0:                NS_IS_LOW_SURROGATE(testLine->c1[1]),
michael@0:                "Test data neither in BMP nor legal surrogate pair");
michael@0:   return SURROGATE_TO_UCS4(testLine->c1[0], testLine->c1[1]);
michael@0: }
michael@0: 
michael@0: bool TestUnspecifiedCodepoint(uint32_t codepoint)
michael@0: {
michael@0:   bool rv = true;
michael@0:   char16_t unicharArray[3];
michael@0:   nsAutoString X, normalized;
michael@0:   char description[9];
michael@0: 
michael@0:   if (IS_IN_BMP(codepoint)) {
michael@0:     unicharArray[0] = codepoint;
michael@0:     unicharArray[1] = 0;
michael@0:     X = nsDependentString(unicharArray);
michael@0:   }
michael@0:   else {
michael@0:     unicharArray[0] = H_SURROGATE(codepoint);
michael@0:     unicharArray[1] = L_SURROGATE(codepoint);
michael@0:     unicharArray[2] = 0;
michael@0:     X = nsDependentString(unicharArray);
michael@0:   }
michael@0: 
michael@0:   /*
michael@0:  2. For every code point X assigned in this version of Unicode that is not specifically
michael@0:     listed in Part 1, the following invariants must be true for all conformant
michael@0:     implementations:
michael@0: 
michael@0:       X == NFC(X) == NFD(X) == NFKC(X) == NFKD(X)
michael@0:   */
michael@0:   DEBUG_TESTCASE(X);
michael@0:   sprintf(description, "U+%04X", codepoint);
michael@0:   NORMALIZE_AND_COMPARE(X, X, NFC, description);
michael@0:   NORMALIZE_AND_COMPARE(X, X, NFD, description);
michael@0:   NORMALIZE_AND_COMPARE(X, X, NFKC, description);
michael@0:   NORMALIZE_AND_COMPARE(X, X, NFKD, description);
michael@0:   return rv;
michael@0: }
michael@0: 
michael@0: void TestPart0()
michael@0: {
michael@0:   printf("Test Part0: Specific cases\n");
michael@0: 
michael@0:   uint32_t i = 0;
michael@0:   uint32_t numFailed = 0;
michael@0:   uint32_t numPassed = 0;
michael@0: 
michael@0:   while (Part0TestData[i].c1[0] != 0) {
michael@0:     if (TestInvariants(&Part0TestData[i++]))
michael@0:       ++numPassed;
michael@0:     else
michael@0:       ++numFailed;
michael@0:   }
michael@0:   printf(" %d cases passed, %d failed\n\n", numPassed, numFailed);
michael@0: }
michael@0: 
michael@0: void TestPart1()
michael@0: {
michael@0:   printf("Test Part1: Character by character test\n");
michael@0: 
michael@0:   uint32_t i = 0;
michael@0:   uint32_t numFailed = 0;
michael@0:   uint32_t numPassed = 0;
michael@0:   uint32_t codepoint;
michael@0:   uint32_t testDataCodepoint = UTF32CodepointFromTestcase(&Part1TestData[i]);
michael@0: 
michael@0:   for (codepoint = 1; codepoint < 0x110000; ++codepoint) {
michael@0:     if (testDataCodepoint == codepoint) {
michael@0:       if (TestInvariants(&Part1TestData[i]))
michael@0:         ++numPassed;
michael@0:       else
michael@0:         ++numFailed;
michael@0:       testDataCodepoint = UTF32CodepointFromTestcase(&Part1TestData[++i]);
michael@0:     } else {
michael@0:       if (TestUnspecifiedCodepoint(codepoint))
michael@0:         ++numPassed;
michael@0:       else
michael@0:         ++numFailed;
michael@0:     }
michael@0:   }
michael@0:   printf(" %d cases passed, %d failed\n\n", numPassed, numFailed);
michael@0: }
michael@0: 
michael@0: void TestPart2()
michael@0: {
michael@0:   printf("Test Part2: Canonical Order Test\n");
michael@0: 
michael@0:   uint32_t i = 0;
michael@0:   uint32_t numFailed = 0;
michael@0:   uint32_t numPassed = 0;
michael@0: 
michael@0:   while (Part2TestData[i].c1[0] != 0) {
michael@0:     if (TestInvariants(&Part2TestData[i++]))
michael@0:       ++numPassed;
michael@0:     else
michael@0:       ++numFailed;
michael@0:   }
michael@0:   printf(" %d cases passed, %d failed\n\n", numPassed, numFailed);
michael@0: }
michael@0: 
michael@0: void TestPart3()
michael@0: {
michael@0:   printf("Test Part3: PRI #29 Test\n");
michael@0: 
michael@0:   uint32_t i = 0;
michael@0:   uint32_t numFailed = 0;
michael@0:   uint32_t numPassed = 0;
michael@0: 
michael@0:   while (Part3TestData[i].c1[0] != 0) {
michael@0:     if (TestInvariants(&Part3TestData[i++]))
michael@0:       ++numPassed;
michael@0:     else
michael@0:       ++numFailed;
michael@0:   }
michael@0:   printf(" %d cases passed, %d failed\n\n", numPassed, numFailed);
michael@0: }
michael@0: 
michael@0: int main(int argc, char** argv) {
michael@0:   if (sizeof(wchar_t) != 2) {
michael@0:     printf("This test can only be run where sizeof(wchar_t) == 2\n");
michael@0:     return 1;
michael@0:   }
michael@0:   if (strlen(versionText) == 0) {
michael@0:     printf("No testcases: to run the tests generate the header file using\n");
michael@0:     printf(" perl genNormalizationData.pl\n");
michael@0:     printf("in intl/unichar/tools and rebuild\n");
michael@0:     return 1;
michael@0:   }
michael@0: 
michael@0:   printf("NormalizationTest: test nsIUnicodeNormalizer. UCD version: %s\n", 
michael@0:          versionText); 
michael@0:   if (argc <= 1)
michael@0:     verboseMode = false;
michael@0:   else if ((argc == 2) && (!strcmp(argv[1], "-v")))
michael@0:     verboseMode = true;
michael@0:   else {
michael@0:     printf("                   Usage: NormalizationTest [OPTION]..\n");
michael@0:     printf("Options:\n");
michael@0:     printf("        -v   Verbose mode\n");
michael@0:     return 1;
michael@0:   }
michael@0: 
michael@0:   nsresult rv = NS_InitXPCOM2(nullptr, nullptr, nullptr);
michael@0:   if (NS_FAILED(rv)) {
michael@0:     printf("NS_InitXPCOM2 failed\n");
michael@0:     return 1;
michael@0:   }
michael@0:   
michael@0:   normalizer = nullptr;
michael@0:   nsresult res;
michael@0:   res = CallGetService(kUnicodeNormalizerCID, &normalizer);
michael@0:   
michael@0:  if(NS_FAILED(res) || !normalizer) {
michael@0:     printf("GetService failed\n");
michael@0:     return 1;
michael@0:   }
michael@0: 
michael@0:   TestPart0();
michael@0:   TestPart1();
michael@0:   TestPart2();
michael@0:   TestPart3();
michael@0:   
michael@0:   NS_RELEASE(normalizer);
michael@0: 
michael@0:   printf("Test finished \n");
michael@0:   return 0;
michael@0: }