1#!/bin/bash 2 3# © 2021 and later: Unicode, Inc. and others. 4# License & terms of use: http://www.unicode.org/copyright.html 5 6# This script runs Bazel to create (generate) header files and binary data files 7# with Unicode character/script/collation data that are 8# needed for bootstrapping the ICU4C build to integrate the data. 9 10# Required environment variables: 11# - ICU_SRC - the root directory of ICU source. This directory contains the 12# `icu4c` directory. 13 14ICU4C_COMMON=$ICU_SRC/icu4c/source/common 15ICU4C_DATA_IN=$ICU_SRC/icu4c/source/data/in 16ICU4C_NORM2=$ICU_SRC/icu4c/source/data/unidata/norm2 17 18# Delete the files from the source tree that we need to generate, 19# to make extra sure that we do not depend on their old versions for generating new ones. 20# We cannot just delete *_data.h because ucol_data.h is not a generated header file. 21rm $ICU_SRC/icu4c/source/common/norm2_nfc_data.h 22rm $ICU_SRC/icu4c/source/common/propname_data.h 23rm $ICU_SRC/icu4c/source/common/*_props_data.h 24rm $ICU4C_DATA_IN/*.icu 25rm $ICU4C_DATA_IN/*.nrm 26rm $ICU4C_DATA_IN/coll/*.icu 27# icu4c/source/i18n/collationfcd.cpp is generated by genuca; 28# probably hard to build genuca without depending on the old version. 29 30# Exit this shell script when a command fails. 31set -e 32 33# Generate normalization data files directly into the source tree. 34bazelisk run //icu4c/source/tools/gennorm2 -- -o $ICU4C_COMMON/norm2_nfc_data.h -s $ICU4C_NORM2 nfc.txt --csource 35bazelisk run //icu4c/source/tools/gennorm2 -- -o $ICU4C_DATA_IN/nfc.nrm -s $ICU4C_NORM2 nfc.txt 36bazelisk run //icu4c/source/tools/gennorm2 -- -o $ICU4C_DATA_IN/nfkc.nrm -s $ICU4C_NORM2 nfc.txt nfkc.txt 37bazelisk run //icu4c/source/tools/gennorm2 -- -o $ICU4C_DATA_IN/nfkc_cf.nrm -s $ICU4C_NORM2 nfc.txt nfkc.txt nfkc_cf.txt 38bazelisk run //icu4c/source/tools/gennorm2 -- -o $ICU4C_DATA_IN/uts46.nrm -s $ICU4C_NORM2 nfc.txt uts46.txt 39 40# genprops writes several files directly into the source tree. 41bazelisk run //tools/unicode/c/genprops $ICU_SRC/icu4c 42 43# genuca also writes several files directly into the source tree. 44# We run it twice for different versions of the CLDR root sort order. 45bazelisk run //tools/unicode/c/genuca -- --hanOrder implicit $ICU_SRC/icu4c 46bazelisk run //tools/unicode/c/genuca -- --hanOrder radical-stroke $ICU_SRC/icu4c 47# Also generate the ICU4X versions 48bazelisk run //tools/unicode/c/genuca -- --icu4x --hanOrder implicit $ICU_SRC/icu4c 49bazelisk run //tools/unicode/c/genuca -- --icu4x --hanOrder radical-stroke $ICU_SRC/icu4c 50