Migrate locale data to icu to remove the overlap in luni

Luni

Class

issue

Status

Comments

CurrencyTest

4 failures beacuse of different locale data between ri and icu

Done

ICU4j 3.8 is CLDR 1.5 compatible whereas RI uses the old version

FormatterTest

1 failure becasue icu has different data in Japan locale

Done

Format pattern: %tb Argument: 0 expected:<1> but was:<1 月>, this is also caused by CLDR gap

GregorianCalendarTest

1 failure beacause icu has different data about the MinimalDaysInFirstWeek

Reported to ICU

test_getMinimalDaysInFirstWeek Failure expected:<4> but was:<1>, this is also caused by CLDR gap

LocaleTest

6 failures beacuse of different locale data between ri and icu

Done

regression test for HARMONY-1146 failed because icu does not support the legacy countries. One difference about the variant has been reported as ICU ticket 6096

ScannerTest

2 failures in parsing float value in locale ar_AE

Ongoing

even more testcases fail on RI ...

Text

CollationElementIteratorTest

3 failures

Done

Fails on RI, correct the test

DecimalFormatSymbolsTest

1 failure for serialization compatibility

Ongoing

This test fails on RI too

NumberFormatTest

1 failure for different data in ar_AE

Done

expected:<36-> but was:<٣٦->. I suppose the 36 could be represented as ٣٦ in this locale...google translator proved it. http://translate.google.com/translate_t?langpair=ar|en

RuleBasedCollatorTest

2 failures

Done

Fails on RI, correct the test

SimpleDateFormatTest

9 failures since ICU has different pattern against RI

Done

see below

Differences in DateFormat/SimpleDateFormat
1. Icu supports extra pattern chars[1], refert to http://www.icu-project.org/apiref/icu4j/com/ibm/icu/text/SimpleDateFormat.html; most failures in test_applyLocalizedPatternLjava_lang_String() are caused by that.
2. For pattern "ZZZZ", ICU has different output, for example icu returns GMT-07:00 but RI returns -7000.
3. For pattern "z" and "zzzz", ICU has different output, for example icu returns "GMT+14:00, Line Islands Time" for "z, zzzz" but RI returns "LINT, Line Is. Time" in "Pacific/Kiritimati";
most failures in test_timeZoneFormatting() are caused by above 2.
4. test_formatLjava_util_DateLjava_lang_StringBufferLjava_text_FieldPosition() reported as ICU ticket 6089.
5. icu has different explanation on the pattern 'y'. 


To be continued.

[1]

Symbol   Meaning                 Presentation        Example
 ------   -------                 ------------        -------
 G        era designator          (Text)              AD
 y†       year                    (Number)            1996
 Y*       year (week of year)     (Number)            1997
 u*       extended year           (Number)             4601
 M        month in year           (Text & Number)     July & 07
 d        day in month            (Number)            10
 h        hour in am/pm (1~12)    (Number)            12
 H        hour in day (0~23)      (Number)            0
 m        minute in hour          (Number)            30
 s        second in minute        (Number)            55
 S        fractional second       (Number)            978
 E        day of week             (Text)              Tuesday
 e*       day of week (local 1~7) (Number)            2
 D        day in year             (Number)            189
 F        day of week in month    (Number)            2 (2nd Wed in July)
 w        week in year            (Number)            27
 W        week in month           (Number)            2
 a        am/pm marker            (Text)              PM
 k        hour in day (1~24)      (Number)            24
 K        hour in am/pm (0~11)    (Number)            0
 z        time zone               (Text)              Pacific Standard Time
 Z        time zone (RFC 822)     (Number)            -0800
 v        time zone (generic)     (Text)              Pacific Time
 g*       Julian day              (Number)            2451334
 A*       milliseconds in day     (Number)            69540000
 Q*       quarter in year         (Text & Number)     Q1 & 01
 c*       stand alone day of week (Text & Number)     Tuesday & 2
 L*       stand alone month       (Text & Number)     July & 07
 q*       stand alone quarter     (Text & Number)     Q1 & 01
 '        escape for text         (Delimiter)         'Date='
 ''       single quote            (Literal)           'o''clock'

* These items are not supported by Java's SimpleDateFormat.
† ICU interprets a single 'y' differently than Java.

ID

Description

Status

5085

[classlib][swing] InternationalFormatterRTest.testIncrementDecrement failure

Fixed

5129

[classlib][icu] text formatting degradation after at 592434

Fixed

5122

[classlib][icu] significant performance degradation atfter ICU related commits

Fixed

5061

[classlib][luni][text]migrate locale data to ICU4j's

Fixed

5100

[classlib]Dacapo.jython regression after migrating the locale data

Closed

5406

[classlib]CLDR data in harmony is different from that in RI

Open

5466

[classlib][util] display language is reset to the value of constructor after Locale is set as default

Open

5467

[classlib][util] incompatible country names for certain languages

Open

5448

[classlib][luni] FormatterTest will fail beacuse of different locale data between ri and icu

In Progress

5442

[classlib][util] some testcases of org.apache.harmony.luni.tests.java.util.LocaleTest fails on RI

Closed

5439

[classlib][luni] test_getMinimalDaysInFirstWeek in org.apache.harmony.luni.tests.java.util.GregorianCalendarTest fails on RI

Closed

5407

[classlib][luni]4 testcases in java.util.CurrencyTest fails on RI

Closed

5467

[classlib][util] incompatible country names for certain languages

Open

5459

[classlib][util] default TimeZone ignores raw offset after switch to ICU

Fixed

5461

[classlib][util] wrong daylight savings are reported for "MST" and "EST" timezones after locale data unification

Open

5465

[classlib][util] Locale(String, String, String) capitalize third parameter (variant)

Open

5468

[classlib][util] java.util.Date.toString outputs real time zone name on RI, but prints GMT<+/-offset> on Harmony

In progress

5469

[classlib][util] DateFormat.SHORT has different meaning on Harmony and RI

In progress

ICU trac

6035 6036 6037

6043 6044 6045

6089 6095 6096

6174

In queue to report: Harmony-5436 Harmony-5407

Others

593469 fix the regression that we have different locale data for Europe/London, along with some fixes for other difference against RI 593024 Fix a bug that SimpleDateFormat.parse does not reload the latest timezone 592434 Apply patch Harmony-5061 which removes the duplicate locale data

reverted at 596044

patch recommitted at r612718

JIRA to record non-bug differences:

MigrateToICU (last edited 2009-09-20 21:55:20 by localhost)