Roll ICU to 239289

Roll ICU to 239289

ICU @r=239289 has the following ( https://codereview.chromium.org/102283003/
 ) : 

Update icudt46l.dat and related files.

We're about to enable using icudt46l.dat instead of static-linking it on
Linux, Android and Chrome OS.

While testing that CL, it's found that source/data/in/icudt46l.dat was not
updated last time when {linux,mac,android}/icudt*S and windows/icudt.dll
were updated.

The following two items are also updated:

1) The timezone data (latest version : 2013h)
2) Added trnslit/{tr,el,az}_{Upper,Lower,Title}.txt and
translit/css3transform.txt so that we can support CSS3 locale-dependent
transform.

With the above changes, all the ICU data files (*.S, *dll, *dat) on
all platforms are regenerated.

The section on the ICU data file generation in README.chromium was made
clearer as to where to find / how to generate *S and *dat files if one
wishes to customize ICU data differently than Chromium does.

BUG=72633, 292214, 234797, 304911
TEST=Once ICU is rolled in, 1) build goes fine 2) Israeli DST transition is up to date (see bug 292214 comment 9 as to how to test).
TBR=mark@chromium.org

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=239613

[Mark Mentovai, 2013-12-09 13:54:07]

LGTM

[commit-bot: I haz the power, 2013-12-09 17:53:23]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/jshin@chromium.org/101593005/20001

[commit-bot: I haz the power, 2013-12-10 00:36:44]

Change committed as 239613

[adamk, 2013-12-10 02:14:51]

Looks like this broke a Blink layout test on Mac:
http://test-results.appspot.com/dashboards/flakiness_dashboard.html#tests=fas...

Is this expected? Do we just need a rebaseline? It's hard for me to evaluate the
change in behavior...

[jungshik at Google, 2013-12-10 06:01:08]

On 2013/12/10 02:14:51, adamk wrote:
> Looks like this broke a Blink layout test on Mac:
>
http://test-results.appspot.com/dashboards/flakiness_dashboard.html#tests=fas...
> 
> Is this expected? Do we just need a rebaseline? It's hard for me to evaluate
the
> change in behavior...

Thank you for the alert. 

In the past, there was a mysterious discrepancy in CJK segmentation between Mac
and other platforms. If this test now passes on Mac, that's expected because I
changed the way Mac ICU data is built. 

https://code.google.com/p/chromium/codesearch#chromium/src/third_party/WebKit...
 : All pass


https://code.google.com/p/chromium/codesearch#chromium/src/third_party/WebKit...
  : Partly fail 

Then, we can just delete
platform/mac/fast/text/international/cjk-segmentation-expected.txt.

[adamk, 2013-12-10 18:36:06]

On 2013/12/10 06:01:08, Jungshik Shin wrote:
> On 2013/12/10 02:14:51, adamk wrote:
> > Looks like this broke a Blink layout test on Mac:
> >
>
http://test-results.appspot.com/dashboards/flakiness_dashboard.html#tests=fas...
> > 
> > Is this expected? Do we just need a rebaseline? It's hard for me to evaluate
> the
> > change in behavior...
> 
> Thank you for the alert. 
> 
> In the past, there was a mysterious discrepancy in CJK segmentation between
Mac
> and other platforms. If this test now passes on Mac, that's expected because I
> changed the way Mac ICU data is built. 
> 
>
https://code.google.com/p/chromium/codesearch#chromium/src/third_party/WebKit...
>  : All pass
> 
> 
>
https://code.google.com/p/chromium/codesearch#chromium/src/third_party/WebKit...
>   : Partly fail 
> 
> Then, we can just delete
> platform/mac/fast/text/international/cjk-segmentation-expected.txt.

Ooops, you're quite right, I just misread the test output. Will take care of
this today.

[jungshik at Google, 2013-12-11 08:51:07]

On 2013/12/10 18:36:06, adamk wrote:
> On 2013/12/10 06:01:08, Jungshik Shin wrote:
> > On 2013/12/10 02:14:51, adamk wrote:
> > > Looks like this broke a Blink layout test on Mac:
> > >
> >
>
http://test-results.appspot.com/dashboards/flakiness_dashboard.html#tests=fas...
> > > 
> > > Is this expected? Do we just need a rebaseline? It's hard for me to
evaluate
> > the
> > > change in behavior...
> > 
> > Thank you for the alert. 
> > 
> > In the past, there was a mysterious discrepancy in CJK segmentation between
> Mac
> > and other platforms. If this test now passes on Mac, that's expected because
I
> > changed the way Mac ICU data is built. 
> > 
> >
>
https://code.google.com/p/chromium/codesearch#chromium/src/third_party/WebKit...
> >  : All pass
> > 
> > 
> >
>
https://code.google.com/p/chromium/codesearch#chromium/src/third_party/WebKit...
> >   : Partly fail 
> > 
> > Then, we can just delete
> > platform/mac/fast/text/international/cjk-segmentation-expected.txt.
> 
> Ooops, you're quite right, I just misread the test output. Will take care of
> this today.

Thanks ! I resolved http://crbug.com/70686 as fixed (the myterious bug I
mentioned earlier).