Make it so that we can disable ICU.

Author: lemire

CMakeLists.txt

@@ -21,6 +21,7 @@ add_subdirectory(src)
 set(CMAKE_MODULE_PATH ${CMAKE_CURRENT_SOURCE_DIR}/scripts/cmake)
 
 option(ADA_BENCHMARKS "Build benchmarks" OFF)
+option(ADA_USE_ICU "Use ICU if available" ON)
 
 if(BUILD_TESTING OR ADA_BENCHMARKS)
   include(${PROJECT_SOURCE_DIR}/cmake/import.cmake)
@@ -37,6 +38,7 @@ else()
   endif()
 endif(BUILD_TESTING)
 
+if(ADA_USE_ICU)
 
 if (${CMAKE_SYSTEM_NAME} MATCHES "Darwin")
   message(STATUS "Apple system detected.")
@@ -83,7 +85,12 @@ else(ICU_FOUND)
     return()
   endif(MSVC)
 endif(ICU_FOUND)
+
+else(ADA_USE_ICU)
+target_compile_definitions(ada PUBLIC ADA_HAS_ICU=0)
+
 ### Going forward, we have ICU for sure, except under Windows.
+endif(ADA_USE_ICU)
 
 If(ADA_BENCHMARKS)
   message(STATUS "Ada benchmarks enabled.")

README.md

@@ -11,7 +11,7 @@ Specification for URL parser can be found from the
 ## Requirements
 
 - A recent C++ compiler supporting C++17. We test GCC 9 or better, LLVM 10 or better and Microsoft Visual Studio 2022.
-- We use [ICU](https://icu.unicode.org) under non-Windows systems (macOS, Linux). Under Windows, [we rely on Microsoft's builtin IdnToAscii function](https://learn.microsoft.com/en-us/windows/win32/api/winnls/nf-winnls-idntoascii).
+- We use [ICU](https://icu.unicode.org) when it is available.
 
 ## Usage
 

include/ada/common_defs.h

@@ -218,5 +218,6 @@ namespace ada {
 #include <unicode/utf8.h>
 #endif // ADA_HAS_ICU
 
+#define ADA_WINDOWS_TO_ASCII_FALLBACK 0 // we never use anything but ICU. No fallback.
 
 #endif // ADA_COMMON_DEFS_H

singleheader/CMakeLists.txt

@@ -50,7 +50,11 @@ if (Python3_Interpreter_FOUND)
   if(MSVC) 
     target_link_libraries(ada-singleheader-include-source INTERFACE Normaliz)
   else() 
-    target_link_libraries(ada-singleheader-include-source INTERFACE ICU::uc ICU::i18n)
+    if(ADA_USE_ICU)
+      target_link_libraries(ada-singleheader-include-source INTERFACE ICU::uc ICU::i18n)
+    else(ADA_USE_ICU)
+      target_compile_definitions(ada-singleheader-include-source INTERFACE ADA_HAS_ICU=0)
+    endif(ADA_USE_ICU)
   endif()
   if (BUILD_TESTING)
     add_executable(demo $<BUILD_INTERFACE:${CMAKE_CURRENT_BINARY_DIR}/demo.cpp>)

src/unicode.cpp

@@ -411,8 +411,11 @@ constexpr static uint8_t is_forbidden_domain_code_point_table[] = {
     }
     out.value().resize(length); // we possibly want to call :shrink_to_fit otherwise we use 255 bytes.
     out.value().shrink_to_fit();
-
-#elif defined(_WIN32)
+    if(std::any_of(out.value().begin(), out.value().end(), ada::unicode::is_forbidden_domain_code_point)) {
+      out = std::nullopt;
+      return false;
+    }
+#elif defined(_WIN32) && ADA_WINDOWS_TO_ASCII_FALLBACK
     (void)be_strict; // unused.
     // Fallback on the system if ICU is not available.
     // Windows function assumes UTF-16.
@@ -522,14 +525,14 @@ constexpr static uint8_t is_forbidden_domain_code_point_table[] = {
     }
     out = std::string(required_buffer_size, 0);
     for(size_t i = 0; i < required_buffer_size; i++) { (*out)[i] = char(ascii_buffer.get()[i]); }
-#else
-    (void)be_strict; // unused.
-    out = input; // We cannot do much more for now.
-#endif
     if(std::any_of(out.value().begin(), out.value().end(), ada::unicode::is_forbidden_domain_code_point)) {
       out = std::nullopt;
       return false;
     }
+#else
+    (void)be_strict; // unused.
+    out = input; // We cannot do much more for now.
+#endif
     return true;
   }
 

tests/wpt_tests.cpp

@@ -473,6 +473,7 @@ int main(int argc, char** argv) {
     results[name] = true; // we pretend. The setters fail under Windows due to IDN issues.
 #endif // _WIN32
   }
+#if ADA_HAS_ICU
   name = "urltestdata_encoding("+std::string(ADA_URLTESTDATA_JSON)+")";
   if(all_tests || name.find(filter) != std::string::npos) {
     results[name] = urltestdata_encoding(ADA_URLTESTDATA_JSON);
@@ -481,6 +482,7 @@ int main(int argc, char** argv) {
   if(all_tests || name.find(filter) != std::string::npos) {
     results[name] = urltestdata_encoding(URLTESTDATA_JSON);
   }
+#endif
   name = "verifydnslength_tests("+std::string(VERIFYDNSLENGTH_TESTS_JSON)+")";
   if(all_tests || name.find(filter) != std::string::npos) {
     results[name] = verifydnslength_tests(VERIFYDNSLENGTH_TESTS_JSON);