ICU-12584 Fix memory allocation for getFormats.
Also:
Add a test case for inner message, other small cleanup.
Style fixes, using better test methods
diff --git a/icu4c/source/i18n/msgfmt.cpp b/icu4c/source/i18n/msgfmt.cpp
index 1b1f670..e39b26b 100644
--- a/icu4c/source/i18n/msgfmt.cpp
+++ b/icu4c/source/i18n/msgfmt.cpp
@@ -810,26 +810,31 @@
// method on this object. We construct and resize an array
// on demand that contains aliases to the subformats[i].format
// pointers.
+
+ // Get total required capacity first (it's refreshed on each call).
+ int32_t totalCapacity = 0;
+ for (int32_t partIndex = 0; (partIndex = nextTopLevelArgStart(partIndex)) >= 0; ++totalCapacity) {};
+
MessageFormat* t = const_cast<MessageFormat*> (this);
cnt = 0;
- if (formatAliases == NULL) {
- t->formatAliasesCapacity = (argTypeCount<10) ? 10 : argTypeCount;
+ if (formatAliases == nullptr) {
+ t->formatAliasesCapacity = totalCapacity;
Format** a = (Format**)
uprv_malloc(sizeof(Format*) * formatAliasesCapacity);
- if (a == NULL) {
+ if (a == nullptr) {
t->formatAliasesCapacity = 0;
- return NULL;
+ return nullptr;
}
t->formatAliases = a;
- } else if (argTypeCount > formatAliasesCapacity) {
+ } else if (totalCapacity > formatAliasesCapacity) {
Format** a = (Format**)
- uprv_realloc(formatAliases, sizeof(Format*) * argTypeCount);
- if (a == NULL) {
+ uprv_realloc(formatAliases, sizeof(Format*) * totalCapacity);
+ if (a == nullptr) {
t->formatAliasesCapacity = 0;
- return NULL;
+ return nullptr;
}
t->formatAliases = a;
- t->formatAliasesCapacity = argTypeCount;
+ t->formatAliasesCapacity = totalCapacity;
}
for (int32_t partIndex = 0; (partIndex = nextTopLevelArgStart(partIndex)) >= 0;) {
diff --git a/icu4c/source/test/intltest/msfmrgts.cpp b/icu4c/source/test/intltest/msfmrgts.cpp
index 78ec942..c35e85d 100644
--- a/icu4c/source/test/intltest/msfmrgts.cpp
+++ b/icu4c/source/test/intltest/msfmrgts.cpp
@@ -1,11 +1,11 @@
// © 2016 and later: Unicode, Inc. and others.
// License & terms of use: http://www.unicode.org/copyright.html
/***********************************************************************
- * COPYRIGHT:
+ * COPYRIGHT:
* Copyright (c) 1997-2016, International Business Machines Corporation
* and others. All Rights Reserved.
***********************************************************************/
-
+
#include "unicode/utypes.h"
#if !UCONFIG_NO_FORMATTING
@@ -53,11 +53,12 @@
TESTCASE_AUTO(Test4142938)
TESTCASE_AUTO(TestChoicePatternQuote)
TESTCASE_AUTO(Test4112104)
+ TESTCASE_AUTO(TestICU12584)
TESTCASE_AUTO(TestAPI)
TESTCASE_AUTO_END;
}
-UBool
+UBool
MessageFormatRegressionTest::failure(UErrorCode status, const char* msg, UBool possibleDataError)
{
if(U_FAILURE(status)) {
@@ -98,7 +99,7 @@
failure(status, "messageFormat->applyPattern");
//Object[] params = {new UnicodeString("BUG"), new Date()};
Formattable params [] = {
- Formattable(UnicodeString("BUG")),
+ Formattable(UnicodeString("BUG")),
Formattable(0, Formattable::kIsDate)
};
UnicodeString tempBuffer;
@@ -111,14 +112,14 @@
//Apply pattern without param and print the result
messageFormatter->applyPattern(pattern[0], status);
failure(status, "messageFormatter->applyPattern");
-
+
// {sfb} how much does this apply in C++?
// do we want to verify that the Formattable* array is not NULL,
// or is that the user's responsibility?
// additionally, what should be the item count?
// for bug testing purposes, assume that something was set to
// NULL by mistake, and that the length should be non-zero
-
+
//tempBuffer = messageFormatter->format(NULL, 1, tempBuffer, FieldPosition(FieldPosition::DONT_CARE), status);
tempBuffer.remove();
tempBuffer = messageFormatter->format(NULL, 0, tempBuffer, pos, status);
@@ -175,16 +176,16 @@
/* @bug 4031438
* More robust message formats.
*/
-void MessageFormatRegressionTest::Test4031438()
+void MessageFormatRegressionTest::Test4031438()
{
UErrorCode status = U_ZERO_ERROR;
-
+
UnicodeString pattern1("Impossible {1} has occurred -- status code is {0} and message is {2}.");
UnicodeString pattern2("Double '' Quotes {0} test and quoted '{1}' test plus 'other {2} stuff'.");
MessageFormat *messageFormatter = new MessageFormat("", status);
failure(status, "new MessageFormat");
-
+
const UBool possibleDataError = TRUE;
//try {
@@ -209,9 +210,9 @@
NumberFormat *fmt = 0;
UnicodeString temp, temp1;
-
+
for (int i = 0; i < count; i++) {
-
+
// convert to string if not already
Formattable obj = objs[i];
temp.remove();
@@ -245,13 +246,13 @@
//if (objs[i] != NULL && objs[i].getString(temp1) != params[i].getString(temp2)) {
if (temp != temp1) {
errln("Parse failed on object " + objs[i].getString(temp1) + " at index : " + i);
- }
+ }
}
delete fmt;
delete [] objs;
- // {sfb} does this apply? no way to really pass a null Formattable,
+ // {sfb} does this apply? no way to really pass a null Formattable,
// only a null array
/*tempBuffer = messageFormatter->format(null, tempBuffer, FieldPosition(FieldPosition::DONT_CARE), status);
@@ -266,7 +267,7 @@
if (tempBuffer != "Double ' Quotes 7 test and quoted {1} test plus 'other {2} stuff'.")
dataerrln("quote format test (w/ params) failed. - %s", u_errorName(status));
logln("Formatted with params : " + tempBuffer);
-
+
/*tempBuffer = messageFormatter->format(null);
if (!tempBuffer.equals("Double ' Quotes {0} test and quoted {1} test plus other {2} stuff."))
errln("quote format test (w/ null) failed.");
@@ -290,7 +291,7 @@
MessageFormat *fmt = new MessageFormat("There are {0} apples growing on the {1} tree.", status);
failure(status, "new MessageFormat");
UnicodeString str("There is one apple growing on the peach tree.");
-
+
int32_t count = 0;
fmt->parse(str, pos, count);
@@ -300,27 +301,27 @@
pos.setErrorIndex(4);
if (pos.getErrorIndex() != 4)
errln(UnicodeString("setErrorIndex failed, got ") + pos.getErrorIndex() + " instead of 4");
-
+
ChoiceFormat *f = new ChoiceFormat(
"-1#are negative|0#are no or fraction|1#is one|1.0<is 1+|2#are two|2<are more than 2.", status);
failure(status, "new ChoiceFormat");
- pos.setIndex(0);
+ pos.setIndex(0);
pos.setErrorIndex(-1);
Formattable obj;
f->parse("are negative", obj, pos);
if (pos.getErrorIndex() != -1 && obj.getDouble() == -1.0)
errln(UnicodeString("Parse with \"are negative\" failed, at ") + pos.getErrorIndex());
- pos.setIndex(0);
+ pos.setIndex(0);
pos.setErrorIndex(-1);
f->parse("are no or fraction ", obj, pos);
if (pos.getErrorIndex() != -1 && obj.getDouble() == 0.0)
errln(UnicodeString("Parse with \"are no or fraction\" failed, at ") + pos.getErrorIndex());
- pos.setIndex(0);
+ pos.setIndex(0);
pos.setErrorIndex(-1);
f->parse("go postal", obj, pos);
if (pos.getErrorIndex() == -1 && ! uprv_isNaN(obj.getDouble()))
errln(UnicodeString("Parse with \"go postal\" failed, at ") + pos.getErrorIndex());
-
+
delete fmt;
delete f;
}
@@ -334,7 +335,7 @@
{
double limits [] = {1, 20};
UnicodeString formats [] = {
- UnicodeString("xyz"),
+ UnicodeString("xyz"),
UnicodeString("abc")
};
int32_t formats_length = UPRV_LENGTHOF(formats);
@@ -374,7 +375,7 @@
// logln("ChoiceFormat constructor should check for the array lengths");
// cf = null;
//}
- //if (cf != null)
+ //if (cf != null)
// errln(cf->format(5));
//
delete cf;
@@ -390,8 +391,8 @@
{
double limits [] = {3, 1, 2};
UnicodeString formats [] = {
- UnicodeString("Three"),
- UnicodeString("One"),
+ UnicodeString("Three"),
+ UnicodeString("One"),
UnicodeString("Two")
};
ChoiceFormat *cf = new ChoiceFormat(limits, formats, 3);
@@ -475,12 +476,12 @@
void MessageFormatRegressionTest::Test4116444()
{
UnicodeString patterns [] = {
- (UnicodeString)"",
- (UnicodeString)"one",
+ (UnicodeString)"",
+ (UnicodeString)"one",
(UnicodeString) "{0,date,short}"
};
-
- UErrorCode status = U_ZERO_ERROR;
+
+ UErrorCode status = U_ZERO_ERROR;
MessageFormat *mf = new MessageFormat("", status);
failure(status, "new MessageFormat");
@@ -490,7 +491,7 @@
failure(status, "mf->applyPattern", TRUE);
//try {
- int32_t count = 0;
+ int32_t count = 0;
ParsePosition pp(0);
Formattable *array = mf->parse(UnicodeString(""), pp, count);
logln("pattern: \"" + pattern + "\"");
@@ -503,7 +504,7 @@
dataerrln("\"" + array[j].getString(dummy) + "\"");
//else
// log("null");
- if (j < count- 1)
+ if (j < count- 1)
log(",");
}
log("}") ;
@@ -528,7 +529,7 @@
void MessageFormatRegressionTest::Test4114739()
{
- UErrorCode status = U_ZERO_ERROR;
+ UErrorCode status = U_ZERO_ERROR;
MessageFormat *mf = new MessageFormat("<{0}>", status);
failure(status, "new MessageFormat");
@@ -716,7 +717,7 @@
UnicodeString pat;
logln(UnicodeString("") + i + ". pattern :\"" + mf->toPattern(pat) + "\"");
log(" \"" + formatted + "\" parsed as ");
- if (objs == NULL)
+ if (objs == NULL)
logln(" null");
else {
UnicodeString temp;
@@ -805,10 +806,10 @@
form2->setFormat(0, *fileform);
//Object[] testArgs = {new Long(12373), "MyDisk"};
Formattable testArgs [] = {
- Formattable((int32_t)12373),
+ Formattable((int32_t)12373),
Formattable((UnicodeString)"MyDisk")
};
-
+
FieldPosition bogus(FieldPosition::DONT_CARE);
UnicodeString result;
@@ -831,8 +832,8 @@
MessageFormat *mf = new MessageFormat("pattern", status);
failure(status, "new MessageFormat");
UnicodeString texts[] = {
- (UnicodeString)"pattern",
- (UnicodeString)"pat",
+ (UnicodeString)"pattern",
+ (UnicodeString)"pat",
(UnicodeString)"1234"
};
UnicodeString pat;
@@ -861,7 +862,7 @@
* This is actually a problem in ChoiceFormat; it doesn't
* understand single quotes.
*/
-void MessageFormatRegressionTest::Test4142938()
+void MessageFormatRegressionTest::Test4142938()
{
UnicodeString pat = CharsToUnicodeString("''Vous'' {0,choice,0#n''|1#}avez s\\u00E9lectionn\\u00E9 "
"{0,choice,0#aucun|1#{0}} client{0,choice,0#s|1#|2#s} "
@@ -874,7 +875,7 @@
CharsToUnicodeString("'Vous' n'avez s\\u00E9lectionn\\u00E9 aucun clients personnels."),
CharsToUnicodeString("'Vous' avez s\\u00E9lectionn\\u00E9 "),
CharsToUnicodeString("'Vous' avez s\\u00E9lectionn\\u00E9 ")
- };
+ };
UnicodeString SUFFIX [] = {
UnicodeString(),
UNICODE_STRING(" client personnel.", 18),
@@ -914,7 +915,7 @@
* pattern characters '|', '#', '<', and '\u2264'. Two quotes in a row
* is a quote literal.
*/
-void MessageFormatRegressionTest::TestChoicePatternQuote()
+void MessageFormatRegressionTest::TestChoicePatternQuote()
{
// ICU 4.8 ChoiceFormat (like PluralFormat & SelectFormat)
// returns the chosen string unmodified, so that it is usable in a MessageFormat.
@@ -958,7 +959,7 @@
catch (IllegalArgumentException e) {
errln("Fail: Pattern \"" + DATA[i] + "\" -> " + e);
}*/
-
+
delete cf;
delete cf2;
}
@@ -969,7 +970,7 @@
* MessageFormat.equals(null) throws a NullPointerException. The JLS states
* that it should return false.
*/
-void MessageFormatRegressionTest::Test4112104()
+void MessageFormatRegressionTest::Test4112104()
{
UErrorCode status = U_ZERO_ERROR;
MessageFormat *format = new MessageFormat("", status);
@@ -987,11 +988,37 @@
delete format;
}
+void MessageFormatRegressionTest::TestICU12584() {
+ // We were able to handle only 10 due to the bug ICU-12584.
+ UnicodeString pattern(
+ u"{1}{2}{3}{1}{2}{3}{1}{2}{3}{1}{2}{3}{1}{2}{3}{1}{2}{3}{1}{2}{3}{1}"
+ u"{2}{3}{1}{2}{3}{1}{2}{3}{1}{2}{3}{1}{2}{3}{1}{2}{3}{1}{2}{3}");
+
+ UErrorCode status = U_ZERO_ERROR;
+
+ MessageFormat msg(pattern, status);
+ failure(status, "Flat message");
+
+ int32_t count;
+ msg.getFormats(count);
+ assertEquals("Plain placeholder match", 42, count);
+
+ // Make sure we iterate only over top level arguments (so we don't over allocate).
+ UnicodeString inner_pattern("{1}{1,select,fem {1} masc {2} other {3}}{2}");
+ status = U_ZERO_ERROR;
+ MessageFormat inner_msg(inner_pattern, status);
+ failure(status, "Inner message");
+
+ count = 0;
+ inner_msg.getFormats(count);
+ assertEquals("Inner placeholder match", 3, count);
+}
+
void MessageFormatRegressionTest::TestAPI() {
UErrorCode status = U_ZERO_ERROR;
MessageFormat *format = new MessageFormat("", status);
failure(status, "new MessageFormat");
-
+
// Test adoptFormat
MessageFormat *fmt = new MessageFormat("",status);
format->adoptFormat("some_name",fmt,status); // Must at least pass a valid identifier.
diff --git a/icu4c/source/test/intltest/msfmrgts.h b/icu4c/source/test/intltest/msfmrgts.h
index b346c87..3330184 100644
--- a/icu4c/source/test/intltest/msfmrgts.h
+++ b/icu4c/source/test/intltest/msfmrgts.h
@@ -1,25 +1,25 @@
// © 2016 and later: Unicode, Inc. and others.
// License & terms of use: http://www.unicode.org/copyright.html
/********************************************************************
- * COPYRIGHT:
+ * COPYRIGHT:
* Copyright (c) 1997-2009, International Business Machines Corporation and
* others. All Rights Reserved.
********************************************************************/
#ifndef _MESSAGEFORMATREGRESSIONTEST_
#define _MESSAGEFORMATREGRESSIONTEST_
-
+
#include "unicode/utypes.h"
#if !UCONFIG_NO_FORMATTING
#include "intltest.h"
-/**
+/**
* Performs regression test for MessageFormat
**/
-class MessageFormatRegressionTest: public IntlTest {
-
+class MessageFormatRegressionTest: public IntlTest {
+
// IntlTest override
void runIndexedTest( int32_t index, UBool exec, const char* &name, char* par );
public:
@@ -45,6 +45,7 @@
void Test4142938(void);
void TestChoicePatternQuote(void);
void Test4112104(void);
+ void TestICU12584(void);
void TestAPI(void);
protected:
@@ -53,6 +54,6 @@
};
#endif /* #if !UCONFIG_NO_FORMATTING */
-
+
#endif // _MESSAGEFORMATREGRESSIONTEST_
//eof
