The previous commit ( https://codereview.chromium.org/1836993003 )
has misbuilt Android data file/assembly. All the currency data
files were truncated.

BUG=598404
TEST=See comment 1 in the bug

Review URL: https://codereview.chromium.org/1856563002 .

Update the timezone source files and the pre-built data file
for all platforms.

This is done directly in chrome/m50 branch of ICU (which is ICU 54.1)
because Chrome ToT uses ICU 56.1.

BUG=598404
TEST=See comment 1 in the bug

Review URL: https://codereview.chromium.org/1836993003 .

fread is used in putil.cpp without examining the return value. Suppress
this warning for ICU code only.

This is done for chromium/m49 branch (that can be used for head, too)
because origin/master is already ICU 56.1 that does not need this
warning to be suppressed.

Original Patch by slan@ : https://codereview.chromium.org/1704763002/
BUG=587965
R=slan@chromium.org

Review URL: https://codereview.chromium.org/1719453003 .

Set fLastSetLiteral to U_SENTINEL in the ctor and check if it's still
U_SENTINEL before adding the range.

https://codereview.chromium.org/1656003004/ is the master CL being
cherry-picked.

BUG=582470
TEST=See comments 13 and 15 in the bug
TBR=mmoroz

Review URL: https://codereview.chromium.org/1654213004 .

This is to cherry-pick a fix for
http://bugs.icu-project.org/trac/ticket/12130

BUG=575206
TEST=See comment 51 in the bug
R=mmoroz@chromium.org

Review URL: https://codereview.chromium.org/1659693002 .

* Big5 : https://www.w3.org/Bugs/Public/show_bug.cgi?id=27878

  Special case the following four more code points in
  addition to U+5341, U+5345 that are already special cased.

      U+2550, U+255E, U+2561, U+256A

  For those 6 code points, the last pointer instead of the first pointer
  in index-big5.txt is used for round-trip. The first pointer is for
  decoding-only.

* KOI8-U ( https://www.w3.org/Bugs/Public/show_bug.cgi?id=17053 )

  - 0xAE and 0xBE are mapped to U+04[50]E instead of U+255[DC].
  - Add an alias KOI8-RU

BUG=544228
TEST=1. http://goo.gl/reGQPU :  encoding(form) test
 2. Layout test: fast/encoding/*

R=jsbell@chromium.org

Review URL: https://codereview.chromium.org/1514253003 .

The new GN rule will enable removal of code which copies icudtl.dat in Android
GN code (for instance "copy_content_shell_assets").

BUG=547162
R=dpranke@chromium.org

Review URL: https://codereview.chromium.org/1419583008 .

Previously GN would implicitly make a group's deps public. I'm making this behave like other target types, so ICU's group's deps need to be public to forward the build information to dependents.

TBR=dpranke@chromium.org

Review URL: https://codereview.chromium.org/1374173003 .

In line 1105 of rbt_pars.cpp, variablesVector.elementAt(j) is cast to
|UnicodeSet *|, but sometimes it's |Quantifier *|. Casting to a common
parent class |UnicodeFunctor *| seems to work (this patch), but it may
hide a deeper issue.

Further investigation will be done in the upstream bug:
  http://bugs.icu-project.org/trac/ticket/11937

This CL was independently made by krasin@ and jshin@.
( https://codereview.chromium.org/1366563006/ )

BUG=528799
TEST=wtf_unittests --gtest_filter="*ToUpp*" and --gtest_filter="*ToLow*" with both regular and cfi build.
TBR=krasin@chromium.org

Review URL: https://codereview.chromium.org/1369653003 .

The warning complains that StringThreadTest is declared in the global namespace,
but the friend line says `friend class StringThreadTest;` instead of
`friend class ::StringThreadTest`.  MSVC doesn't require this ::, but gcc does
-- without it, the friend line declares a new class in the namespace containing
UnicodeString.

Since StringThreadTest was in test/threadtest which isn't built as part of
Chromium's build and which was removed in upstream ICU in ICU 37392, just
delete the friend line, it has no effect.

BUG=505296
R=jshin@chromium.org

Review URL: https://codereview.chromium.org/1317603003 .

Also add a few comments to the .gn file.

BUG=505316
R=jshin@chromium.org

Review URL: https://codereview.chromium.org/1301493002 .

1.
In Win32NumberFormat, rename the private format() method to variadicFormat().
It's just an implementation detail of the class. Fixes:
..\..\third_party\icu\source\i18n/winnmfmt.h(146,20) :  error(clang):
        'icu_54::Win32NumberFormat::format' hides overloaded virtual functions
        [-Werror,-Woverloaded-virtual]
    UnicodeString &format(int32_t numDigits, UnicodeString &appendTo, const wchar_t *format, ...) const;
                   ^
..\..\third_party\icu\source\i18n\unicode/numfmt.h(255,28) :  note(clang):
        hidden overloaded virtual function 'icu_54::NumberFormat::format'
        declared here: different number of parameters (4 vs 3)
    virtual UnicodeString& format(const Formattable& obj,
                           ^
...

2.
In Win32DateFormat, remove the two-argument format() function that just
forwards to the base class.  (If this happens to produce compile errors,
the function should be replaced by
`using Win32DateFormat::format(UDate, UnicodeString&) const;`) Fixes:
..\..\third_party\icu\source\i18n/windtfmt.h(55,20) :  error(clang):
        'icu_54::Win32DateFormat::format' hides overloaded virtual functions
        [-Werror,-Woverloaded-virtual]
    UnicodeString& format(UDate date, UnicodeString& appendTo) const;
                   ^
..\..\third_party\icu\source\i18n\unicode/datefmt.h(240,28) :  note(clang):
        hidden overloaded virtual function 'icu_54::DateFormat::format'
        declared here: different number of parameters (4 vs 2)
    virtual UnicodeString& format(const Formattable& obj,
                           ^
...

BUG=505301
R=jshin@chromium.org

Review URL: https://codereview.chromium.org/1296803002 .

The Chromium build currently defines both is_component_build (boolean)
and component_mode (a string). We're trying to standardize on the boolean.

TBR=jshin

Review URL: https://codereview.chromium.org/1273843002 .

Upstream CL : http://bugs.icu-project.org/trac/changeset/37670

BUG=508444
TEST=See bug
R=inferno@chromium.org

Review URL: https://codereview.chromium.org/1233663005 .

This is going back to PS #4 in
https://codereview.chromium.org/1148463002/

BUG=466959, 508049, 508440
TEST=See bugs
TBR=thakis@chromium.org

Review URL: https://codereview.chromium.org/1237873002 .

Takes an upstream patch (to be included in 56.x release) and patch
our ICU 54.

Clang-specific atomic ops will be used when clang is used to compile but C+11 atomic is disabled/unavailable.

patch: http://bugs.icu-project.org/trac/changeset/37326
bug: http://bugs.icu-project.org/trac/ticket/11599

BUG=466959,508049,508440
TEST=See bugs
R=glider@chromium.org, thakis@chromium.org

Review URL: https://codereview.chromium.org/1148463002 .

No intended behavior change (except fewer duplicated compiler flags).

BUG=505319
TBR=mgiuca@chromium.org

Review URL: https://codereview.chromium.org/1220363002.

This warning is currently globally suppressed in Chromium, but this
suppression will be switched off, making a local suppression necessary.

BUG=505319
R=thakis@chromium.org

Review URL: https://codereview.chromium.org/1224463003.

BUG=505304
R=thakis@chromium.org

Review URL: https://codereview.chromium.org/1215103002.

Some Windows-specific ICU code uses this pattern, which triggers a clang warning.

BUG=505302
R=thakis@chromium.org

Review URL: https://codereview.chromium.org/1209413007.

R=brettw@chromium.org, dpranke@chromium.org

Review URL: https://codereview.chromium.org/1183493003.

Else clang's -Winfinite-recursion will warn about filteredbrk.cpp

BUG=none
R=scottmg@chromium.org

Review URL: https://codereview.chromium.org/1191653011.

Previously, |referenceCounter != (uint32_t)~0| was used to
check whether refcounting is used or not.

Cherry-pick an upstream change:

   http://bugs.icu-project.org/trac/changeset/37476

BUG=465175
TEST=see the bug
R=jyasskin@chromium.org

Review URL: https://codereview.chromium.org/1171203002.

This allows targets that use GN to correctly note that
they need the .dat file to be included in .isolates.

R=brettw@chromium.org, jshin@chromium.org
BUG=

Review URL: https://codereview.chromium.org/1167613003.

I hang my head in shame.

BUG=498033
TBR=hans@chromium.org

Review URL: https://codereview.chromium.org/1177533002.

This ports https://codereview.chromium.org/515263002/ and
https://codereview.chromium.org/508343002/

I don't see these warnings locally, but they do show up on the bots
for some reason. Probably good to have gyp and gn match anyways.

BUG=498033
R=hans@chromium.org

Review URL: https://codereview.chromium.org/1171153002

1. Add a one-way (encoding-only/fromUnicode) mapping for U+2212 to
   Shift_JIS, EUC-JP and ISO-2022-JP. The last just uses Shift_JIS.

   See https://www.w3.org/Bugs/Public/show_bug.cgi?id=28661

2. Make GBK aliases list compliant to the encoding spec.

3. Add "xA3xA0 => U+3000" to the GBK (windows-936)  and gb18030. This makes it possible to remove the corresponding override in Blink

4. Modify the following to GBK (windows-936). See [1]
  - Add U+01F9 <=> \xA8\xBF
  - Drop U+E7C8 <=> \xA8\xBF

5. The following change is put on hold (NOT included in the CL)  until the resolution of [1]

  - Add U+1E3F <=> \xA8\xBC
  - Drop U+E7C7 <=> \xA8\xBC

The corresponding Blink CL is https://codereview.chromium.org/1167523003/

[1] https://www.w3.org/Bugs/Public/show_bug.cgi?id=28740#c3

BUG=425417,493824
TEST=Once ICU is rolled to this CL, Blink layout test fast/encoding/*.
R=jsbell@chromium.org

Review URL: https://codereview.chromium.org/1162723008

1. -Wno-return-type-c-linkage
 - became unnecessary with ICU bug 9504 fixed 3 years ago.
 - remove from both BUILD.gn and icu.gyp

2. -Wnoheader-hygine
 - became unnecessary a long time ago.
 - remove from BUILD.gn (icu.gyp does not have it)

3. -Wstring-conversion and -Wheader-hygine
 - add to BUILD.gn by NOT excluding clang:extra_warnings

BUG=114660
TEST=Build icuuc/icui18n targets (both gyp and GN) after regenerating build files.
R=brettw@chromium.org, thakis@chromium.org

Review URL: https://codereview.chromium.org/1158933005

A fix by mbarbella@.

BUG=444573
TEST=See the bug
R=inferno@chromium.org, mbarbella@chromium.org

Review URL: https://codereview.chromium.org/1145963004

  Egypt will not observe DST in 2015.

  See http://mm.icann.org/pipermail/tz-announce/2015-April/000031.html
  and http://en.wikipedia.org/wiki/Egypt_Standard_Time

BUG=473288
TBR=mark@chromium.org,jsbell@chromium.org
TEST=In JS console, run the following to get 1:00 PM for 2015
instead of 2:00 PM. In 2014, it should return 2:00 PM.
  may_25_2015_1100 = new Date("05/25/2015 11:00Z")
  may_25_2015_1100.toLocaleString("en", {timeZone: "Africa/Cairo"})
  may_25_2014_1100 = new Date("05/25/2014 11:00Z")
  may_25_2014_1100.toLocaleString("en", {timeZone: "Africa/Cairo"})

Review URL: https://codereview.chromium.org/1138093004

The GYP and GN build's defines have diverged. THis brings GN's defines in line with GYP's current set.

R=jshin@chromium.org

Review URL: https://codereview.chromium.org/1130383003

prep_toolchain is now a no-op.

BUG=456902
R=jshin@chromium.org

Review URL: https://codereview.chromium.org/1058613005

Due to caching, the 2nd timezone change is not detected properly.

Upstream bug: http://bugs.icu-project.org/trac/ticket/11623

BUG=406382
TEST=manual. Call detectHostTimeZone() every time the OS tz is changed on Linux and check if it returns the correct present timezone.
R=mark@chromium.org

Review URL: https://codereview.chromium.org/1058353002

We need this new API to resolve bug 406382 across platforms.

See
http://bugs.icu-project.org/trac/ticket/11358
http://bugs.icu-project.org/trac/changeset/37148

BUG=406382
TEST=None
R=mark@chromium.org
TBR=mark@chromium.org

Review URL: https://codereview.chromium.org/1062503003

Due to a bug in patch_locale.sh, calendars intended for being stripped
away are not removed from the ICU data for Android.

Fixing the bug cuts down the ICU data size by 10kB on Android.

BUG=472142
TEST=android/icudtl.dat is ~10kB smaller than before. (6,267,808 bytes)
R=mark@chromium.org

Review URL: https://codereview.chromium.org/1052333003

1. Update the IANA tz db to 2015b.
  - http://mm.icann.org/pipermail/tz-announce/2015-March/000029.html
  - Mongolia decided to observe DST again in 2015 starting on the last
    Sunday in March.
  - Palestine's DST start date is corrected to be March 28 instead of 27th.

2. Add a script to download the tz database files (update_tz.sh)

3. Check in scripts/make_n_copy_data.sh that I've been using to build ICU
   data/assembly files and update README.chromium.

4. Update android/patch_locale.sh to apply android/brkitr.patch as well.

BUG=473288
TEST=1. In JavaScript console, run the following.
  mar27_2015_1200 = new Date("03/27/2015 12:00Z")
  mar28_2015_1200 = new Date("03/28/2015 12:00Z")
  mar27_2015_1200.toLocaleString("en", {timeZone: "Asia/Gaza"}
  mar28_2015_1200.toLocaleString("en", {timeZone: "Asia/Gaza"}
  apr15_2014_1200 = new Date("04/15/2014 12:00Z")
  apr15_2015_1200 = new Date("04/15/2015 12:00Z")
  apr15_2014_1200.toLocaleString("en", {timeZone: "Asia/Ulan_Bator"}
  apr15_2015_1200.toLocaleString("en", {timeZone: "Asia/Ulan_Bator"}

In Asia/Gaza, Mar 27 12:00Z should be 2PM and mar28 12:00Z should be 3PM.
In Asia/Ulan_Bator, April 15 12:00Z should be 8PM in 2014 and should be 9PM
in 2015. Ulan_Bator does not work due to http://crbug.com/364374.

R=mark@chromium.org

Review URL: https://codereview.chromium.org/1051193002

BUG=470291
TEST=None
R=tsepez@chromium.org

Review URL: https://codereview.chromium.org/1032783002

1. Cherry-pick security patches from the upstream.
  a. BiDi: integer overflow
    http://bugs.icu-project.org/trac/ticket/11451
    http://crbug.com/451799
    bidi.patch was updated to include patches for this issue.
  b. data race in cmemory; remove an unnecessary check with a global variable.
    http://bugs.icu-project.org/trac/ticket/11538
    http://crbug.com/223352
    cmemory.patch was added
  c. Locale::getBaseName() thread-safety
    http://bugs.icu-project.org/trac/ticket/11547
    http://crbug.com/467836
    locid.patch was updated to include patches for this issue.

2. Add UCONFIG_NO_NON_HTML5_CONVERSION=1 to BUILD.gn

3. Clean up
  a. Update README.chromium to flag patches that are already in the upstream.
  b. Split pkg_gen.patch from data.build.patch. pkg_gen.patch was already
     in the upstream while the rest of data.build.patch is Chromium-specific.
  c. Delete an unused converters.patch.

BUG=223352,451799,467836
TEST=See bugs 223352, 467836, 468716, and 466838
R=jyasskin@chromium.org, mark@chromium.org

Review URL: https://codereview.chromium.org/1020303002

I forgot to include this change in
  https://codereview.chromium.org/984233002

TBR=jochen@chromium.org
BUG=NONE

Review URL: https://codereview.chromium.org/1021673003

1. Update ucmlocal.mk and convertrs.txt to refer to euc-kr-html.ucm
instead of windows-949.ucm

2. Tighten up the valid code range for the following converters:

   EUC-KR, Shift_JIS, Big5

This is to add back an ASCII range byte to the stream per
the encoding spec when they're either illegal as a 'trail byte' or
there's no assigned code point for a "lead + trail" sequence.
For instance, with this change, '0xF3 0x41' in EUC-KR is converted to
'U+FFFD U+0041' instead of 'U+FFFD'.

This change requires adding 2 ~ 8 new states to the conversion
table of each converter mentioned above leading to 6.5kB net increase
in the final data size.

3. Tighten the trail byte range for 2-byte sequences starting with 0x8E
from [A1,E2] to [A1,DF] in EUC-JP and update the corresponding generating
script.

4. Change the substitution characters for EUC-JP and Shift_JIS to
match other converters. i.e. make them produce U+FFFD when encountering
an invalid input. Before this chaange, they emitted U+001A.

5. Enable 'U_CHARSET_IS_UTF8' configuration flag.
Chromium/Blink does not rely on ICU for the code conversion between
the 'system native encoding' (if it's one of legacy encodings)
and Unicode. With this configuration, we can cut down the code size
a bit.

6. Update the icudtl.dat (all platforms) and assembly files (mac,linux)
   and the icudata dll (windows)

See https://codereview.chromium.org/1026453002 for a new blink test
added ( fast/encoding/char-decoding-invalid-trail.html )

BUG=450312,430823
TEST=Blink: fast/encoding/char-decoding-{truncated,invalid-trail}.html
TEST=base_unittests --gtest_filter=*Conv*, browser_tests --gtest_filter=*ncoding*
R=jsbell@chromium.org, mark@chromium.org

Review URL: https://codereview.chromium.org/984233002