ICU update to 58 part 2

scripts/trim_data.sh

 #!/bin/bash
 # Copyright (c) 2014 The Chromium Authors. All rights reserved.
 # Use of this source code is governed by a BSD-style license that can be
 # found in the LICENSE file.
 
 
+# Remove entries currently not used in Chromium/V8.
+function filter_locale_data {
+  echo Removing unncessary categories in ${localedatapath}
+  for langpath in ${localedatapath}/*.txt
+  do
+    echo Overwriting ${langpath} ...
+    sed -r -i \
+      '/^    characterLabel\{$/,/^    \}$/d
+       /^    AuxExemplarCharacters\{.*\}$/d
+       /^    AuxExemplarCharacters\{$/, /^    \}$/d
+       /^    ExemplarCharacters\{.*\}$/d
+       /^    ExemplarCharacters\{$/, /^    \}$/d
+       /^        (mon|tue|wed|thu|fri|sat|sun|quarter)(|-short|-narrow)\{$/, /^        \}$/d' ${langpath}
+  done
+}
+
 # Remove display names for languages that are not listed in the accept-language
 # list of Chromium.
 function filter_display_language_names {
   for lang in $(grep -v '^#' "${scriptdir}/accept_lang.list")
   do
     # Set $OP to '|' only if $ACCEPT_LANG_PATTERN is not empty.
     OP=${ACCEPT_LANG_PATTERN:+|}
     ACCEPT_LANG_PATTERN="${ACCEPT_LANG_PATTERN}${OP}${lang}"
   done
   ACCEPT_LANG_PATTERN="(${ACCEPT_LANG_PATTERN})[^a-z]"
 
-  echo "Filtering out display names for non-A-L languages ${langdatapath}"
-  for lang in $(grep -v '^#' "${scriptdir}/chrome_ui_languages.list")
+  echo "Filtering out display names for non-A-L languages in ${langdatapath}"
+  for langpath in ${langdatapath}/*.txt
   do
-    target=${langdatapath}/${lang}.txt
+    target=${langpath}
     echo Overwriting ${target} ...
     sed -r -i \
     '/^    Keys\{$/,/^    \}$/d
      /^    Languages\{$/, /^    \}$/ {
        /^    Languages\{$/p
        /^        '${ACCEPT_LANG_PATTERN}'/p
        /^    \}$/p
        d
      }
      /^    Types\{$/,/^    \}$/d
+     /^    Types%short\{$/,/^    \}$/d
+     /^    characterLabelPattern\{$/,/^    \}$/d
      /^    Variants\{$/,/^    \}$/d' ${target}
 
     # Delete an empty "Languages" block. Otherwise, getting the display
     # name for all the language in a given locale (e.g. en_GB) would fail
     # when the above filtering sed command results in an empty "Languages"
     # block.
     sed -r -i \
     '/^    Languages\{$/ {
        N
        /^    Languages\{\n    \}/ d
-     }' ${target}
+    }' ${target}
   done
 }
 
 
 # Keep only the minimum locale data for non-UI languages.
 function abridge_locale_data_for_non_ui_languages {
   for lang in $(grep -v '^#' "${scriptdir}/chrome_ui_languages.list")
   do
     # Set $OP to '|' only if $UI_LANGUAGES is not empty.
     OP=${UI_LANGUAGES:+|}
     UI_LANGUAGES="${UI_LANGUAGES}${OP}${lang}"
   done
 
   EXTRA_LANGUAGES=$(egrep -v -e '^#' -e "(${UI_LANGUAGES})" \
                     "${scriptdir}/accept_lang.list")
 
   echo Creating minimum locale data in ${localedatapath}
   for lang in ${EXTRA_LANGUAGES}
   do
     target=${localedatapath}/${lang}.txt
     [  -e ${target} ] || { echo "missing ${lang}"; continue; }
     echo Overwriting ${target} ...
 
     # Do not include '%%Parent' line on purpose.
     sed -n -r -i \
       '1, /^'${lang}'\{$/p
        /^    "%%ALIAS"\{/p
-       /^    AuxExemplarCharacters\{.*\}$/p
-       /^    AuxExemplarCharacters\{$/, /^    \}$/p
-       /^    ExemplarCharacters\{.*\}$/p
-       /^    ExemplarCharacters\{$/, /^    \}$/p
        /^    (LocaleScript|layout)\{$/, /^    \}$/p
        /^    Version\{.*$/p
        /^\}$/p' ${target}
   done
 
   echo Creating minimum locale data in ${langdatapath}
   for lang in ${EXTRA_LANGUAGES}
   do
     target=${langdatapath}/${lang}.txt
     [  -e ${target} ] || { echo "missing ${lang}"; continue; }
@@ skipping 98 matching lines @@
 }
 
 treeroot="$(dirname "$0")/.."
 dataroot="${treeroot}/source/data"
 scriptdir="${treeroot}/scripts"
 localedatapath="${dataroot}/locales"
 langdatapath="${dataroot}/lang"
 
 
 
+filter_locale_data
 filter_display_language_names
 abridge_locale_data_for_non_ui_languages
 filter_currency_data
 filter_region_data
 remove_legacy_chinese_codepoint_collation
 filter_unit_data
 
 # Chromium OS needs exemplar cities for timezones, but not Chromium.
 # It'll save 400kB (uncompressed), but the size difference in
 # 7z compressed installer is <= 100kB.
 # TODO(jshin): Make separate data files for CrOS and Chromium.
 #remove_exemplar_cities