• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1# Copyright (C) 2016 and later: Unicode, Inc. and others.
2# License & terms of use: http://www.unicode.org/copyright.html
3# Copyright (C) 2010-2014, International Business Machines Corporation and others.
4# All Rights Reserved.
5#
6# Commands for regenerating ICU4C locale data (.txt files) from CLDR.
7#
8# The process requires local copies of
9#    - CLDR (the source of most of the data, and some Java tools)
10#    - ICU4J  (used only for checking the converted data)
11#    - ICU4C  (the destination for the new data, and the source for some of it)
12#
13# For an official CLDR data integration into ICU, these should be clean, freshly
14# checked-out. For released CLDR sources, an alternative to checking out sources
15# for a given version is downloading the zipped sources for the common (core.zip)
16# and tools (tools.zip) directory subtrees from the Data column in
17# [http://cldr.unicode.org/index/downloads].
18#
19# The versions of each of these must match. Included with the release notes for
20# ICU is the version number and/or a CLDR svn tag name for the revision of CLDR
21# that was the source of the data for that release of ICU.
22#
23# Note: Some versions of the OpenJDK will not build the CLDR java utilities.
24#   If you see compilation errors complaining about type incompatibilities with
25#   functions on generic classes, try switching to the Sun JDK.
26#
27# Besides a standard JDK, the process also requires ant
28# (http://ant.apache.org/),
29# plus the xml-apis.jar from the Apache xalan package
30# (http://xml.apache.org/xalan-j/downloads.html).
31#
32# Note: Enough things can (and will) fail in this process that it is best to
33#   run the commands separately from an interactive shell. They should all
34#   copy and paste without problems.
35#
36# It is often useful to save logs of the output of many of the steps in this
37# process. The commands below put log files in /tmp; you may want to put them
38# somewhere else.
39#
40#----
41#
42# IP address whitelisting
43#
44# Parts of the build process (notably building the new ICU data files in step 4)
45# require http: access to files in the CLDR repository; for example, processing
46# the files in icu4c/source/data/xml/ may require access to
47# http://www.unicode.org/repos/cldr/trunk/common/dtd/ldml.dtd
48#
49# Unless you cache the dtds locally by e.g. setting -DCLDR_DTD_CACHE=/tmp, the
50# builds will repeatedly make such requests, which will likely result in the
51# Unicode server blocking access and consequent timeout failures. You can
52# prevent such blockage by having the external IP address of your build system
53# whitelisted with Unicode; contact Rick McGowan or Steven Loomis.
54#
55#----
56#
57# There are several environment variables that need to be defined.
58#
59# a) Java- and ant-related variables
60#
61# JAVA_HOME:     Path to JDK (a directory, containing e.g. bin/java, bin/javac,
62#                etc.); on many systems this can be set using
63#                `/usr/libexec/java_home`.
64#
65# ANT_OPTS:      You may want to set:
66#
67#                -Xmx3072m, to give Java more memory; otherwise it may run out
68#                 of heap.
69#                -DCLDR_DTD_CACHE=/tmp, to reduce frequent http: access to dtds
70#                 and consequent blockage by Unicode server.
71#
72# b) CLDR-related variables
73#
74# CLDR_DIR:      Path to root of CLDR sources, below which are the common and
75#                tools directories.
76# CLDR_CLASSES:  Defined relative to CLDR_DIR. It only needs to be set if you
77#                are not running ant jar for CLDR and have a non-default output
78#                folder for cldr-tools classes.
79#
80# c) ICU-related variables
81# These variables only need to be set if you're directly reusing the
82# commands below.
83#
84# ICU4C_DIR:     Path to root of ICU4C sources, below which is the source dir.
85#
86# ICU4J_ROOT:    Path to root of ICU4J sources, below which is the main dir.
87#
88#----
89#
90# If you are adding or removing locales, or specific kinds of locale data,
91# there are some xml files in the ICU sources that need to be updated (these xml
92# files are used in addition to the CLDR files as inputs to the CLDR data build
93# process for ICU):
94#
95#    icu4c/source/data/icu-config.xml - Update <locales> to add or remove
96#                CLDR locales for inclusion in ICU. Update <paths> to prefer
97#                alt forms for certain paths, or to exclude certain paths; note
98#                that <paths> items can only have draft or alt attributes.
99#
100#                Note that if a language-only locale (e.g. "de") is included in
101#                <locales>, then all region sublocales for that language that
102#                are present in CLDR data (e.g. "de_AT", "de_BE", "de_CH", etc.)
103#                should also be included in <locales>, per PMC policy decision
104#                2012-05-02 (see http://bugs.icu-project.org/trac/ticket/9298).
105#
106#    icu4c/source/data/build.xml - If you are adding or removing break
107#                iterators, you need to update  <fileset id="brkitr" ...> under
108#                <target name="clean" ...> to clean the correct set of files.
109#
110#    icu4c/source/data/xml/      - If you are adding a new locale, break
111#                iterator, collation tailoring, or rule-based number formatter,
112#                you may need to add a corresponding xml file in (respectively)
113#                the main/, brkitr/, collation/, or rbnf/ subdirectory here.
114#
115#----
116#
117# For an official CLDR data integration into ICU, there are some additional
118# considerations:
119#
120# a) Don't commit anything in ICU sources (and possibly any changes in CLDR
121#    sources, depending on their nature) until you have finished testing and
122#    resolving build issues and test failures for both ICU4C and ICU4J.
123#
124# b) There are version numbers that may need manual updating in CLDR (other
125#    version numbers get updated automatically, based on these):
126#
127#    common/dtd/ldml.dtd                            - update cldrVersion
128#    common/dtd/ldmlBCP47.dtd                       - update cldrVersion
129#    common/dtd/ldmlSupplemental.dtd                - update cldrVersion
130#    tools/java/org/unicode/cldr/util/CLDRFile.java - update GEN_VERSION
131#
132# c) After everything is committed, you will need to tag the CLDR, ICU4J, and
133#    ICU4C sources that ended up being used for the integration; see step 17
134#    below.
135#
136################################################################################
137
138# 1a. Java and ant variables, adjust for your system
139
140export JAVA_HOME=`/usr/libexec/java_home`
141export ANT_OPTS="-Xmx3072m -DCLDR_DTD_CACHE=/tmp"
142
143# 1b. CLDR variables, adjust for your setup; with cygwin it might be e.g.
144# CLDR_DIR=`cygpath -wp /build/cldr`
145
146export CLDR_DIR=$HOME/cldr/trunk
147#export CLDR_CLASSES=$CLDR_DIR/tools/java/classes
148
149# 1c. ICU variables
150
151export ICU4C_DIR=$HOME/icu/trunk/icu4c
152export ICU4J_ROOT=$HOME/icu/trunk/icu4j
153
154# 2. Build the CLDR Java tools
155
156cd $CLDR_DIR/tools/java
157ant jar
158
159# 3. Configure ICU4C, build and test without new data first, to verify that
160# there are no pre-existing errors (configure shown here for MacOSX, adjust
161# for your platform).
162
163cd $ICU4C_DIR/source
164./runConfigureICU MacOSX
165make all 2>&1 | tee /tmp/icu4c-oldData-makeAll.txt
166make check 2>&1 | tee /tmp/icu4c-oldData-makeCheck.txt
167
168# 4. Build the new ICU4C data files; these include .txt files and .mk files.
169# These new files will replace whatever was already present in the ICU4C sources.
170# This process uses ant with ICU's data/build.xml and data/icu-config.xml to
171# operate (via CLDR's ant/CLDRConverterTool.java and ant/CLDRBuild.java) the
172# necessary CLDR tools including LDML2ICUConverter, ConvertTransforms, etc.
173# This process will take several minutes.
174# Keep a log so you can investigate anything that looks suspicious.
175#
176# If you see timeout errors when building the rbnf data, for example, then the
177# system you are building on likely does not have its IP address whitelisted with
178# Unicode for access to the CLDR repository, see note on "IP address whitelisting"
179# near the top of this file.
180
181cd $ICU4C_DIR/source/data
182ant clean
183ant all 2>&1 | tee /tmp/cldr-newData-buildLog.txt
184
185# 5. Check which data files have modifications, which have been added or removed
186# (if there are no changes, you may not need to proceed further). Make sure the
187# list seems reasonable.
188
189git status
190
191# 6. Fix any errors, investigate any warnings. Some warnings are expected,
192# including  warnings for missing versions in locale names which specify some
193# collationvariants, e.g.
194#   [cldr-build] WARNING (ja_JP_TRADITIONAL): No version #??
195#   [cldr-build] WARNING (zh_TW_STROKE): No version #??
196# and warnings for some empty collation bundles, e.g.
197#   [cldr-build] WARNING (en):  warning: No collations found. Bundle will ...
198#   [cldr-build] WARNING (to):  warning: No collations found. Bundle will ...
199#
200# Fixing may entail modifying CLDR source data or tools - for example,
201# updating the validSubLocales for collation data (file a bug if appropriate).
202# Repeat steps 4-5 until there are no build errors and no unexpected
203# warnings.
204
205# 7. Now rebuild ICU4C with the new data and run make check tests.
206# Again, keep a log so you can investigate the errors.
207
208cd $ICU4C_DIR/source
209make check 2>&1 | tee /tmp/icu4c-newData-makeCheck.txt
210
211# 8. Investigate each test case failure. The first run processing new CLDR data
212# from the Survey Tool can result in thousands of failures (in many cases, one
213# CLDR data fix can resolve hundreds of test failures). If the error is caused
214# by bad CLDR data, then file a CLDR bug, fix the data, and regenerate from
215# step 4. If the data is OK but the testcase needs to be updated because the
216# data has legitimately changed, then update the testcase. You will check in
217# the updated testcases along with the new ICU data at the end of this process.
218# Note that if the new data has any differences in structure, you will have to
219# update test/testdata/structLocale.txt or /tsutil/cldrtest/TestLocaleStructure
220# may fail.
221# Repeat steps 4-7 until there are no errors.
222
223# 9. Now run the make check tests in exhaustive mode:
224
225cd $ICU4C_DIR/source
226export INTLTEST_OPTS="-e"
227export CINTLTST_OPTS="-e"
228make check 2>&1 | tee /tmp/icu4c-newData-makeCheckEx.txt
229
230# 10. Again, investigate each failure, fixing CLDR data or ICU test cases as
231# appropriate, and repeating steps 4-7 and 9 until there are no errors.
232
233# 11. Now with ICU4J, build and test without new data first, to verify that
234# there are no pre-existing errors (or at least to have the pre-existing errors
235# as a base for comparison):
236
237cd $ICU4J_ROOT
238ant all 2>&1 | tee /tmp/icu4j-oldData-antAll.txt
239ant check 2>&1 | tee /tmp/icu4j-oldData-antCheck.txt
240
241# 12. Now build the new data and test data for ICU4J
242
243cd $ICU4C_DIR/source/data
244make icu4j-data-install
245cd $ICU4C_DIR/source/test/testdata
246make icu4j-data-install
247
248# 13. Now rebuild ICU4J with the new data and run tests:
249# Keep a log so you can investigate the errors.
250
251cd $ICU4J_ROOT
252ant check 2>&1 | tee /tmp/icu4j-newData-antCheck.txt
253
254# 14. Investigate test case failures; fix test cases and repeat from step 12,
255# or fix CLDR data and repeat from step 4, as appropriate, until; there are no
256# more failures in ICU4C or ICU4J (except failures that were present before you
257# began testing the new CLDR data).
258
259# 15. Check the file changes; then svn add or svn remove as necessary, and
260# commit the changes.
261
262cd $ICU4C_DIR/source
263git status
264# add or remove as necessary
265
266cd $ICU4J_ROOT
267git status
268# add or remove as necessary
269
270cd $HOME/icu/trunk/
271# commit
272
273# 16. For an official CLDR data integration into ICU, now tag the CLDR and
274# ICU sources with an appropriate CLDR milestone (you can check previous
275# tags for format), e.g.:
276
277svn copy svn+ssh://unicode.org/repos/cldr/trunk \
278svn+ssh://unicode.org/repos/cldr/tags/release-NNN \
279--parents -m "cldrbug nnnn: tag cldr sources for NNN"
280
281svn copy svn+ssh://source.icu-project.org/repos/icu/trunk \
282svn+ssh://source.icu-project.org/repos/icu/tags/cldr-NNN \
283--parents -m 'ticket:mmmm: tag the version used for integrating CLDR NNN'
284
285