1 // © 2016 and later: Unicode, Inc. and others.
2 // License & terms of use: http://www.unicode.org/copyright.html
3 /********************************************************************
4 * COPYRIGHT:
5 * Copyright (c) 2002-2014, International Business Machines Corporation and
6 * others. All Rights Reserved.
7 ********************************************************************/
8
9 //
10 // dcfmtest.cpp
11 //
12 // Decimal Formatter tests, data driven.
13 //
14
15 #include "intltest.h"
16
17 #if !UCONFIG_NO_FORMATTING && !UCONFIG_NO_REGULAR_EXPRESSIONS
18
19 #include "unicode/regex.h"
20 #include "unicode/uchar.h"
21 #include "unicode/ustring.h"
22 #include "unicode/unistr.h"
23 #include "unicode/dcfmtsym.h"
24 #include "unicode/decimfmt.h"
25 #include "unicode/locid.h"
26 #include "cmemory.h"
27 #include "dcfmtest.h"
28 #include "util.h"
29 #include "cstring.h"
30 #include <stdlib.h>
31 #include <string.h>
32 #include <stdio.h>
33
34 #if defined(__GLIBCXX__)
35 namespace std { class type_info; } // WORKAROUND: http://llvm.org/bugs/show_bug.cgi?id=13364
36 #endif
37
38 #include <string>
39 #include <iostream>
40
41 //---------------------------------------------------------------------------
42 //
43 // Test class boilerplate
44 //
45 //---------------------------------------------------------------------------
DecimalFormatTest()46 DecimalFormatTest::DecimalFormatTest()
47 {
48 }
49
50
~DecimalFormatTest()51 DecimalFormatTest::~DecimalFormatTest()
52 {
53 }
54
55
56
runIndexedTest(int32_t index,UBool exec,const char * & name,char *)57 void DecimalFormatTest::runIndexedTest( int32_t index, UBool exec, const char* &name, char* /*par*/ )
58 {
59 if (exec) logln("TestSuite DecimalFormatTest: ");
60 switch (index) {
61
62 #if !UCONFIG_NO_FILE_IO
63 case 0: name = "DataDrivenTests";
64 if (exec) DataDrivenTests();
65 break;
66 #else
67 case 0: name = "skip";
68 break;
69 #endif
70
71 default: name = "";
72 break; //needed to end loop
73 }
74 }
75
76
77 //---------------------------------------------------------------------------
78 //
79 // Error Checking / Reporting macros used in all of the tests.
80 //
81 //---------------------------------------------------------------------------
82 #define DF_CHECK_STATUS UPRV_BLOCK_MACRO_BEGIN { \
83 if (U_FAILURE(status)) { \
84 dataerrln("DecimalFormatTest failure at line %d. status=%s", \
85 __LINE__, u_errorName(status)); \
86 return 0; \
87 } \
88 } UPRV_BLOCK_MACRO_END
89
90 #define DF_ASSERT(expr) UPRV_BLOCK_MACRO_BEGIN { \
91 if ((expr)==FALSE) { \
92 errln("DecimalFormatTest failure at line %d.\n", __LINE__); \
93 } \
94 } UPRV_BLOCK_MACRO_END
95
96 #define DF_ASSERT_FAIL(expr, errcode) UPRV_BLOCK_MACRO_BEGIN { \
97 UErrorCode status=U_ZERO_ERROR; \
98 (expr); \
99 if (status!=errcode) { \
100 dataerrln("DecimalFormatTest failure at line %d. Expected status=%s, got %s", \
101 __LINE__, u_errorName(errcode), u_errorName(status)); \
102 } \
103 } UPRV_BLOCK_MACRO_END
104
105 #define DF_CHECK_STATUS_L(line) UPRV_BLOCK_MACRO_BEGIN { \
106 if (U_FAILURE(status)) { \
107 errln("DecimalFormatTest failure at line %d, from %d. status=%d\n",__LINE__, (line), status); \
108 } \
109 } UPRV_BLOCK_MACRO_END
110
111 #define DF_ASSERT_L(expr, line) UPRV_BLOCK_MACRO_BEGIN { \
112 if ((expr)==FALSE) { \
113 errln("DecimalFormatTest failure at line %d, from %d.", __LINE__, (line)); \
114 return; \
115 } \
116 } UPRV_BLOCK_MACRO_END
117
118
119
120 //
121 // InvariantStringPiece
122 // Wrap a StringPiece around the extracted invariant data of a UnicodeString.
123 // The data is guaranteed to be nul terminated. (This is not true of StringPiece
124 // in general, but is true of InvariantStringPiece)
125 //
126 class InvariantStringPiece: public StringPiece {
127 public:
128 InvariantStringPiece(const UnicodeString &s);
~InvariantStringPiece()129 ~InvariantStringPiece() {}
130 private:
131 MaybeStackArray<char, 20> buf;
132 };
133
InvariantStringPiece(const UnicodeString & s)134 InvariantStringPiece::InvariantStringPiece(const UnicodeString &s) {
135 int32_t len = s.length();
136 if (len+1 > buf.getCapacity()) {
137 buf.resize(len+1);
138 }
139 // Buffer size is len+1 so that s.extract() will nul-terminate the string.
140 s.extract(0, len, buf.getAlias(), len+1, US_INV);
141 this->set(buf.getAlias(), len);
142 }
143
144
145 // UnicodeStringPiece
146 // Wrap a StringPiece around the extracted (to the default charset) data of
147 // a UnicodeString. The extracted data is guaranteed to be nul terminated.
148 // (This is not true of StringPiece in general, but is true of UnicodeStringPiece)
149 //
150 class UnicodeStringPiece: public StringPiece {
151 public:
152 UnicodeStringPiece(const UnicodeString &s);
~UnicodeStringPiece()153 ~UnicodeStringPiece() {}
154 private:
155 MaybeStackArray<char, 20> buf;
156 };
157
UnicodeStringPiece(const UnicodeString & s)158 UnicodeStringPiece::UnicodeStringPiece(const UnicodeString &s) {
159 int32_t len = s.length();
160 int32_t capacity = buf.getCapacity();
161 int32_t requiredCapacity = s.extract(0, len, buf.getAlias(), capacity) + 1;
162 if (capacity < requiredCapacity) {
163 buf.resize(requiredCapacity);
164 capacity = requiredCapacity;
165 s.extract(0, len, buf.getAlias(), capacity);
166 }
167 this->set(buf.getAlias(), requiredCapacity - 1);
168 }
169
170
171
172 //---------------------------------------------------------------------------
173 //
174 // DataDrivenTests
175 // The test cases are in a separate data file,
176 //
177 //---------------------------------------------------------------------------
178
179 // Translate a Formattable::type enum value to a string, for error message formatting.
formattableType(Formattable::Type typ)180 static const char *formattableType(Formattable::Type typ) {
181 static const char *types[] = {"kDate",
182 "kDouble",
183 "kLong",
184 "kString",
185 "kArray",
186 "kInt64",
187 "kObject"
188 };
189 if (typ<0 || typ>Formattable::kObject) {
190 return "Unknown";
191 }
192 return types[typ];
193 }
194
195 const char *
getPath(char * buffer,const char * filename)196 DecimalFormatTest::getPath(char *buffer, const char *filename) {
197 UErrorCode status=U_ZERO_ERROR;
198 const char *testDataDirectory = IntlTest::getSourceTestData(status);
199 DF_CHECK_STATUS;
200
201 strcpy(buffer, testDataDirectory);
202 strcat(buffer, filename);
203 return buffer;
204 }
205
DataDrivenTests()206 void DecimalFormatTest::DataDrivenTests() {
207 char tdd[2048];
208 const char *srcPath;
209 UErrorCode status = U_ZERO_ERROR;
210 int32_t lineNum = 0;
211
212 //
213 // Open and read the test data file.
214 //
215 srcPath=getPath(tdd, "dcfmtest.txt");
216 if(srcPath==NULL) {
217 return; /* something went wrong, error already output */
218 }
219
220 int32_t len;
221 UChar *testData = ReadAndConvertFile(srcPath, len, status);
222 if (U_FAILURE(status)) {
223 return; /* something went wrong, error already output */
224 }
225
226 //
227 // Put the test data into a UnicodeString
228 //
229 UnicodeString testString(FALSE, testData, len);
230
231 RegexMatcher parseLineMat(UnicodeString(
232 "(?i)\\s*parse\\s+"
233 "\"([^\"]*)\"\\s+" // Capture group 1: input text
234 "([ild])\\s+" // Capture group 2: expected parsed type
235 "\"([^\"]*)\"\\s+" // Capture group 3: expected parsed decimal
236 "\\s*(?:#.*)?"), // Trailing comment
237 0, status);
238
239 RegexMatcher formatLineMat(UnicodeString(
240 "(?i)\\s*format\\s+"
241 "(\\S+)\\s+" // Capture group 1: pattern
242 "(ceiling|floor|down|up|halfeven|halfdown|halfup|default|unnecessary)\\s+" // Capture group 2: Rounding Mode
243 "\"([^\"]*)\"\\s+" // Capture group 3: input
244 "\"([^\"]*)\"" // Capture group 4: expected output
245 "\\s*(?:#.*)?"), // Trailing comment
246 0, status);
247
248 RegexMatcher commentMat (UNICODE_STRING_SIMPLE("\\s*(#.*)?$"), 0, status);
249 RegexMatcher lineMat(UNICODE_STRING_SIMPLE("(?m)^(.*?)$"), testString, 0, status);
250
251 if (U_FAILURE(status)){
252 dataerrln("Construct RegexMatcher() error.");
253 delete [] testData;
254 return;
255 }
256
257 //
258 // Loop over the test data file, once per line.
259 //
260 while (lineMat.find()) {
261 lineNum++;
262 if (U_FAILURE(status)) {
263 dataerrln("File dcfmtest.txt, line %d: ICU Error \"%s\"", lineNum, u_errorName(status));
264 }
265
266 status = U_ZERO_ERROR;
267 UnicodeString testLine = lineMat.group(1, status);
268 // printf("%s\n", UnicodeStringPiece(testLine).data());
269 if (testLine.length() == 0) {
270 continue;
271 }
272
273 //
274 // Parse the test line. Skip blank and comment only lines.
275 // Separate out the three main fields - pattern, flags, target.
276 //
277
278 commentMat.reset(testLine);
279 if (commentMat.lookingAt(status)) {
280 // This line is a comment, or blank.
281 continue;
282 }
283
284
285 //
286 // Handle "parse" test case line from file
287 //
288 parseLineMat.reset(testLine);
289 if (parseLineMat.lookingAt(status)) {
290 execParseTest(lineNum,
291 parseLineMat.group(1, status), // input
292 parseLineMat.group(2, status), // Expected Type
293 parseLineMat.group(3, status), // Expected Decimal String
294 status
295 );
296 continue;
297 }
298
299 //
300 // Handle "format" test case line
301 //
302 formatLineMat.reset(testLine);
303 if (formatLineMat.lookingAt(status)) {
304 execFormatTest(lineNum,
305 formatLineMat.group(1, status), // Pattern
306 formatLineMat.group(2, status), // rounding mode
307 formatLineMat.group(3, status), // input decimal number
308 formatLineMat.group(4, status), // expected formatted result
309 kFormattable,
310 status);
311
312 execFormatTest(lineNum,
313 formatLineMat.group(1, status), // Pattern
314 formatLineMat.group(2, status), // rounding mode
315 formatLineMat.group(3, status), // input decimal number
316 formatLineMat.group(4, status), // expected formatted result
317 kStringPiece,
318 status);
319 continue;
320 }
321
322 //
323 // Line is not a recognizable test case.
324 //
325 errln("Badly formed test case at line %d.\n%s\n",
326 lineNum, UnicodeStringPiece(testLine).data());
327
328 }
329
330 delete [] testData;
331 }
332
333
334
execParseTest(int32_t lineNum,const UnicodeString & inputText,const UnicodeString & expectedType,const UnicodeString & expectedDecimal,UErrorCode & status)335 void DecimalFormatTest::execParseTest(int32_t lineNum,
336 const UnicodeString &inputText,
337 const UnicodeString &expectedType,
338 const UnicodeString &expectedDecimal,
339 UErrorCode &status) {
340
341 if (U_FAILURE(status)) {
342 return;
343 }
344
345 DecimalFormatSymbols symbols(Locale::getUS(), status);
346 UnicodeString pattern = UNICODE_STRING_SIMPLE("####");
347 DecimalFormat format(pattern, symbols, status);
348 Formattable result;
349 if (U_FAILURE(status)) {
350 dataerrln("file dcfmtest.txt, line %d: %s error creating the formatter.",
351 lineNum, u_errorName(status));
352 return;
353 }
354
355 ParsePosition pos;
356 int32_t expectedParseEndPosition = inputText.length();
357
358 format.parse(inputText, result, pos);
359
360 if (expectedParseEndPosition != pos.getIndex()) {
361 errln("file dcfmtest.txt, line %d: Expected parse position afeter parsing: %d. "
362 "Actual parse position: %d", expectedParseEndPosition, pos.getIndex());
363 return;
364 }
365
366 char expectedTypeC[2];
367 expectedType.extract(0, 1, expectedTypeC, 2, US_INV);
368 Formattable::Type expectType = Formattable::kDate;
369 switch (expectedTypeC[0]) {
370 case 'd': expectType = Formattable::kDouble; break;
371 case 'i': expectType = Formattable::kLong; break;
372 case 'l': expectType = Formattable::kInt64; break;
373 default:
374 errln("file dcfmtest.tx, line %d: unrecongized expected type \"%s\"",
375 lineNum, InvariantStringPiece(expectedType).data());
376 return;
377 }
378 if (result.getType() != expectType) {
379 errln("file dcfmtest.txt, line %d: expectedParseType(%s) != actual parseType(%s)",
380 lineNum, formattableType(expectType), formattableType(result.getType()));
381 return;
382 }
383
384 StringPiece decimalResult = result.getDecimalNumber(status);
385 if (U_FAILURE(status)) {
386 errln("File %s, line %d: error %s. Line in file dcfmtest.txt: %d:",
387 __FILE__, __LINE__, u_errorName(status), lineNum);
388 return;
389 }
390
391 InvariantStringPiece expectedResults(expectedDecimal);
392 if (decimalResult != expectedResults) {
393 errln("file dcfmtest.txt, line %d: expected \"%s\", got \"%s\"",
394 lineNum, expectedResults.data(), decimalResult.data());
395 }
396
397 return;
398 }
399
400
execFormatTest(int32_t lineNum,const UnicodeString & pattern,const UnicodeString & round,const UnicodeString & input,const UnicodeString & expected,EFormatInputType inType,UErrorCode & status)401 void DecimalFormatTest::execFormatTest(int32_t lineNum,
402 const UnicodeString &pattern, // Pattern
403 const UnicodeString &round, // rounding mode
404 const UnicodeString &input, // input decimal number
405 const UnicodeString &expected, // expected formatted result
406 EFormatInputType inType, // input number type
407 UErrorCode &status) {
408 if (U_FAILURE(status)) {
409 return;
410 }
411
412 DecimalFormatSymbols symbols(Locale::getUS(), status);
413 // printf("Pattern = %s\n", UnicodeStringPiece(pattern).data());
414 DecimalFormat fmtr(pattern, symbols, status);
415 if (U_FAILURE(status)) {
416 dataerrln("file dcfmtest.txt, line %d: %s error creating the formatter.",
417 lineNum, u_errorName(status));
418 return;
419 }
420 if (round=="ceiling") {
421 fmtr.setRoundingMode(DecimalFormat::kRoundCeiling);
422 } else if (round=="floor") {
423 fmtr.setRoundingMode(DecimalFormat::kRoundFloor);
424 } else if (round=="down") {
425 fmtr.setRoundingMode(DecimalFormat::kRoundDown);
426 } else if (round=="up") {
427 fmtr.setRoundingMode(DecimalFormat::kRoundUp);
428 } else if (round=="halfeven") {
429 fmtr.setRoundingMode(DecimalFormat::kRoundHalfEven);
430 } else if (round=="halfdown") {
431 fmtr.setRoundingMode(DecimalFormat::kRoundHalfDown);
432 } else if (round=="halfup") {
433 fmtr.setRoundingMode(DecimalFormat::kRoundHalfUp);
434 } else if (round=="default") {
435 // don't set any value.
436 } else if (round=="unnecessary") {
437 fmtr.setRoundingMode(DecimalFormat::kRoundUnnecessary);
438 } else {
439 fmtr.setRoundingMode(DecimalFormat::kRoundFloor);
440 errln("file dcfmtest.txt, line %d: Bad rounding mode \"%s\"",
441 lineNum, UnicodeStringPiece(round).data());
442 }
443
444 const char *typeStr = "Unknown";
445 UnicodeString result;
446 UnicodeStringPiece spInput(input);
447
448 switch (inType) {
449 case kFormattable:
450 {
451 typeStr = "Formattable";
452 Formattable fmtbl;
453 fmtbl.setDecimalNumber(spInput, status);
454 fmtr.format(fmtbl, result, NULL, status);
455 }
456 break;
457 case kStringPiece:
458 typeStr = "StringPiece";
459 fmtr.format(spInput, result, NULL, status);
460 break;
461 }
462
463 if ((status == U_FORMAT_INEXACT_ERROR) && (result == "") && (expected == "Inexact")) {
464 // Test succeeded.
465 status = U_ZERO_ERROR;
466 return;
467 }
468
469 if (U_FAILURE(status)) {
470 errln("[%s] file dcfmtest.txt, line %d: format() returned %s.",
471 typeStr, lineNum, u_errorName(status));
472 status = U_ZERO_ERROR;
473 return;
474 }
475
476 if (result != expected) {
477 errln("[%s] file dcfmtest.txt, line %d: expected \"%s\", got \"%s\", %s",
478 typeStr, lineNum, UnicodeStringPiece(expected).data(), UnicodeStringPiece(result).data(),
479 u_errorName(status));
480 }
481 }
482
483
484 //-------------------------------------------------------------------------------
485 //
486 // Read a text data file, convert it from UTF-8 to UChars, and return the data
487 // in one big UChar * buffer, which the caller must delete.
488 //
489 // (Lightly modified version of a similar function in regextst.cpp)
490 //
491 //--------------------------------------------------------------------------------
ReadAndConvertFile(const char * fileName,int32_t & ulen,UErrorCode & status)492 UChar *DecimalFormatTest::ReadAndConvertFile(const char *fileName, int32_t &ulen,
493 UErrorCode &status) {
494 UChar *retPtr = NULL;
495 char *fileBuf = NULL;
496 const char *fileBufNoBOM = NULL;
497 FILE *f = NULL;
498
499 ulen = 0;
500 if (U_FAILURE(status)) {
501 return retPtr;
502 }
503
504 //
505 // Open the file.
506 //
507 f = fopen(fileName, "rb");
508 if (f == 0) {
509 dataerrln("Error opening test data file %s\n", fileName);
510 status = U_FILE_ACCESS_ERROR;
511 return NULL;
512 }
513 //
514 // Read it in
515 //
516 int32_t fileSize;
517 int32_t amtRead;
518 int32_t amtReadNoBOM;
519
520 fseek( f, 0, SEEK_END);
521 fileSize = ftell(f);
522 fileBuf = new char[fileSize];
523 fseek(f, 0, SEEK_SET);
524 amtRead = static_cast<int32_t>(fread(fileBuf, 1, fileSize, f));
525 if (amtRead != fileSize || fileSize <= 0) {
526 errln("Error reading test data file.");
527 goto cleanUpAndReturn;
528 }
529
530 //
531 // Look for a UTF-8 BOM on the data just read.
532 // The test data file is UTF-8.
533 // The BOM needs to be there in the source file to keep the Windows &
534 // EBCDIC machines happy, so force an error if it goes missing.
535 // Many Linux editors will silently strip it.
536 //
537 fileBufNoBOM = fileBuf + 3;
538 amtReadNoBOM = amtRead - 3;
539 if (fileSize<3 || uprv_strncmp(fileBuf, "\xEF\xBB\xBF", 3) != 0) {
540 // TODO: restore this check.
541 errln("Test data file %s is missing its BOM", fileName);
542 fileBufNoBOM = fileBuf;
543 amtReadNoBOM = amtRead;
544 }
545
546 //
547 // Find the length of the input in UTF-16 UChars
548 // (by preflighting the conversion)
549 //
550 u_strFromUTF8(NULL, 0, &ulen, fileBufNoBOM, amtReadNoBOM, &status);
551
552 //
553 // Convert file contents from UTF-8 to UTF-16
554 //
555 if (status == U_BUFFER_OVERFLOW_ERROR) {
556 // Buffer Overflow is expected from the preflight operation.
557 status = U_ZERO_ERROR;
558 retPtr = new UChar[ulen+1];
559 u_strFromUTF8(retPtr, ulen+1, NULL, fileBufNoBOM, amtReadNoBOM, &status);
560 }
561
562 cleanUpAndReturn:
563 fclose(f);
564 delete[] fileBuf;
565 if (U_FAILURE(status)) {
566 errln("ICU Error \"%s\"\n", u_errorName(status));
567 delete retPtr;
568 retPtr = NULL;
569 }
570 return retPtr;
571 }
572
573 #endif /* !UCONFIG_NO_REGULAR_EXPRESSIONS */
574
575