1
Fork 0
mirror of https://github.com/RGBCube/serenity synced 2025-05-30 21:48:11 +00:00

LibUnicode: Download and parse the special casing UCD file

This adds a SpecialCasing structure to the generated UnicodeData.h/cpp
files. This structure contains casing rules for code points which have
non-1-to-1 upper-to-lower case code point mappings. Further, these rules
may be limited to specific locales or other context.
This commit is contained in:
Timothy Flynn 2021-07-27 10:39:37 -04:00 committed by Linus Groh
parent 350fb0e3db
commit 32ea461385
2 changed files with 219 additions and 22 deletions

View file

@ -3,11 +3,18 @@ option(ENABLE_UNICODE_DATABASE_DOWNLOAD "Enable download of Unicode UCD files at
set(UNICODE_DATA_URL https://www.unicode.org/Public/13.0.0/ucd/UnicodeData.txt)
set(UNICODE_DATA_PATH ${CMAKE_BINARY_DIR}/UCD/UnicodeData.txt)
set(SPECIAL_CASING_URL https://www.unicode.org/Public/13.0.0/ucd/SpecialCasing.txt)
set(SPECIAL_CASING_PATH ${CMAKE_BINARY_DIR}/UCD/SpecialCasing.txt)
if (ENABLE_UNICODE_DATABASE_DOWNLOAD)
if (NOT EXISTS ${UNICODE_DATA_PATH})
message(STATUS "Downloading UCD UnicodeData.txt from ${UNICODE_DATA_URL}...")
file(DOWNLOAD ${UNICODE_DATA_URL} ${UNICODE_DATA_PATH} INACTIVITY_TIMEOUT 10)
endif()
if (NOT EXISTS ${SPECIAL_CASING_PATH})
message(STATUS "Downloading UCD SpecialCasing.txt from ${SPECIAL_CASING_URL}...")
file(DOWNLOAD ${SPECIAL_CASING_URL} ${SPECIAL_CASING_PATH} INACTIVITY_TIMEOUT 10)
endif()
set(UNICODE_GENERATOR CodeGenerators/GenerateUnicodeData)
set(UNICODE_DATA_HEADER UnicodeData.h)
@ -25,18 +32,18 @@ if (ENABLE_UNICODE_DATABASE_DOWNLOAD)
add_custom_command(
OUTPUT ${UNICODE_DATA_HEADER}
COMMAND ${write_if_different} ${UNICODE_DATA_HEADER} ${UNICODE_GENERATOR} -h -u ${UNICODE_DATA_PATH}
COMMAND ${write_if_different} ${UNICODE_DATA_HEADER} ${UNICODE_GENERATOR} -h -u ${UNICODE_DATA_PATH} -s ${SPECIAL_CASING_PATH}
VERBATIM
DEPENDS GenerateUnicodeData
MAIN_DEPENDENCY ${UNICODE_DATA_PATH}
MAIN_DEPENDENCY ${UNICODE_DATA_PATH} ${SPECIAL_CASING_PATH}
)
add_custom_command(
OUTPUT ${UNICODE_DATA_IMPLEMENTATION}
COMMAND ${write_if_different} ${UNICODE_DATA_IMPLEMENTATION} ${UNICODE_GENERATOR} -c -u ${UNICODE_DATA_PATH}
COMMAND ${write_if_different} ${UNICODE_DATA_IMPLEMENTATION} ${UNICODE_GENERATOR} -c -u ${UNICODE_DATA_PATH} -s ${SPECIAL_CASING_PATH}
VERBATIM
DEPENDS GenerateUnicodeData
MAIN_DEPENDENCY ${UNICODE_DATA_PATH}
MAIN_DEPENDENCY ${UNICODE_DATA_PATH} ${SPECIAL_CASING_PATH}
)
set(UNICODE_DATA_SOURCES ${UNICODE_DATA_HEADER} ${UNICODE_DATA_IMPLEMENTATION})