Initial static ICU support for wasm

Author: EgorBo

src/libraries/Native/Unix/System.Globalization.Native/pal_icushim.c

@@ -6,6 +6,7 @@
 #include <stdlib.h>
 #include "pal_icushim_internal.h"
 
+#if !defined(STATIC_ICU)
 #if defined(TARGET_UNIX)
 #include <dlfcn.h>
 #elif defined(TARGET_WINDOWS)
@@ -482,3 +483,34 @@ int32_t GlobalizationNative_GetICUVersion()
 
     return (versionInfo[0] << 24) + (versionInfo[1] << 16) + (versionInfo[2] << 8) + versionInfo[3];
 }
+
+#else // !defined(STATIC_ICU)
+
+#include <unicode/putil.h>
+#include <unicode/uversion.h>
+
+int32_t GlobalizationNative_LoadICU(void)
+{
+    const char* icudir = getenv("DOTNET_ICU_DIR");
+    if (!icudir)
+        return 0;
+
+    // path to a directory with icudt___.dat (e.g. icudt67l.dat)
+    // we can also use `udata_setCommonData(const void *data, UErrorCode *err)` API here
+    u_setDataDirectory(icudir);
+    return 1;
+}
+
+void GlobalizationNative_InitICUFunctions(void* icuuc, void* icuin, const char* version, const char* suffix)
+{
+    // no-op for static
+}
+
+int32_t GlobalizationNative_GetICUVersion(void)
+{
+    UVersionInfo versionInfo;
+    u_getVersion(versionInfo);
+
+    return (versionInfo[0] << 24) + (versionInfo[1] << 16) + (versionInfo[2] << 8) + versionInfo[3];
+}
+#endif // defined(STATIC_ICU)

src/libraries/Native/Unix/System.Globalization.Native/pal_icushim_internal.h

@@ -56,6 +56,7 @@
 
 #include "pal_compiler.h"
 
+#if !defined(STATIC_ICU)
 // List of all functions from the ICU libraries that are used in the System.Globalization.Native.so
 #define FOR_ALL_UNCONDITIONAL_ICU_FUNCTIONS \
     PER_FUNCTION_BLOCK(u_charsToUChars, libicuuc) \
@@ -280,3 +281,4 @@ FOR_ALL_ICU_FUNCTIONS
 #define usearch_getMatchedLength(...) usearch_getMatchedLength_ptr(__VA_ARGS__)
 #define usearch_last(...) usearch_last_ptr(__VA_ARGS__)
 #define usearch_openFromCollator(...) usearch_openFromCollator_ptr(__VA_ARGS__)
+#endif // !defined(STATIC_ICU)

src/mono/cmake/config.h.in

@@ -210,6 +210,9 @@
 /* Use static zlib */
 #cmakedefine HAVE_STATIC_ZLIB 1
 
+/* Use static ICU */
+#cmakedefine STATIC_ICU 1
+
 /* Use OS-provided zlib */
 #cmakedefine HAVE_SYS_ZLIB 1
 

src/mono/configure.ac

@@ -1309,6 +1309,12 @@ AC_ARG_WITH(spectre-mitigation,             [  --with-spectre-mitigation=yes,no
 AC_ARG_WITH(spectre-indirect-branch-choice,   [  --with-spectre-indirect-branch-choice=keep,thunk,inline,extern   Convert indirect branches to the specified kind of thunk (defaults to inline)], [], [with_spectre_indirect_branch_choice=inline])
 AC_ARG_WITH(spectre-function-return-choice, [  --with-spectre-function-return-choice=keep,thunk,inline,extern   Convert function return instructions to the specified kind of thunk (defaults to inline)], [], [with_spectre_function_return_choice=inline])
 
+AC_ARG_WITH(static_icu, [  --with-static-icu=yes|no      Integrate ICU statically into the runtime (defaults to no)],[
+	if test x$with_static_icu = xyes ; then
+		AC_DEFINE(STATIC_ICU,1,[Integrate ICU statically into the runtime.])
+	fi
+], [with_static_icu=no])
+
 dnl
 dnl Spectre compiler mitigation flag checks
 dnl
@@ -1696,6 +1702,7 @@ AM_CONDITIONAL(INSTALL_MONOTOUCH, [test "x$with_monotouch" != "xno"])
 AM_CONDITIONAL(INSTALL_MONOTOUCH_WATCH, [test "x$with_monotouch_watch" != "xno"])
 AM_CONDITIONAL(INSTALL_MONOTOUCH_TV, [test "x$with_monotouch_tv" != "xno"])
 AM_CONDITIONAL(BITCODE, test "x$with_bitcode" = "xyes")
+AM_CONDITIONAL(STATIC_ICU, test "x$with_static_icu" = "xyes")
 AM_CONDITIONAL(INSTALL_XAMMAC, [test "x$with_xammac" != "xno"])
 AM_CONDITIONAL(INSTALL_TESTING_AOT_FULL_INTERP, [test "x$with_testing_aot_full_interp" != "xno"])
 AM_CONDITIONAL(INSTALL_TESTING_AOT_HYBRID, [test "x$with_testing_aot_hybrid" != "xno"])
@@ -6836,7 +6843,10 @@ if test x$with_core = xonly; then
 	fi
 	if test x$have_shim_globalization = xyes || test x$cross_compiling = xyes; then
 		ICU_SHIM_PATH=../../../libraries/Native/Unix/System.Globalization.Native
-		if test x$target_osx = xyes; then
+		if test x$target_wasm = xyes && test x$with_static_icu = xyes; then
+			ICU_CFLAGS="-DTARGET_UNIX -DU_DISABLE_RENAMING"
+			have_sys_icu=yes
+		elif test x$target_osx = xyes; then
 			ORIG_CPPFLAGS=$CPPFLAGS
 			# adding icu path to pkg_config_path
 			PKG_CONFIG_PATH=$PKG_CONFIG_PATH:/usr/local/lib/pkgconfig:/usr/local/opt/icu4c/lib/pkgconfig

src/mono/mono.proj

@@ -635,6 +635,7 @@
       <_MonoConfigureParams Include="--disable-icall-tables"/>
       <_MonoConfigureParams Include="--disable-crash-reporting"/>
       <_MonoConfigureParams Include="--with-bitcode=yes"/>
+      <_MonoConfigureParams Condition="'$(ICU_REPO)' != ''" Include="--with-static-icu=yes"/>
       <_MonoConfigureParams Include="--enable-minimal=ssa,com,jit,reflection_emit_save,portability,assembly_remapping,attach,verifier,full_messages,appdomains,shadowcopy,security,sgen_marksweep_conc,sgen_split_nursery,sgen_gc_bridge,logging,remoting,shared_perfcounters,sgen_debug_helpers,soft_debug,interpreter,assert_messages,cleanup,mdb,gac,threads,$(_MonoEnableMinimal)"/>
       <_MonoCFLAGS Include="-fexceptions" />
       <_MonoCXXFLAGS Include="-fexceptions -s DISABLE_EXCEPTION_CATCHING=0" />

src/mono/mono/metadata/Makefile.am

@@ -158,6 +158,14 @@ nodist_libmonoruntime_shimglobalization_la_SOURCES = \
 									@ICU_SHIM_PATH@/entrypoints.c
 
 libmonoruntime_shimglobalization_la_CFLAGS = @ICU_CFLAGS@ -I$(top_srcdir)/../libraries/Native/Unix/System.Globalization.Native/ -I$(top_srcdir)/../libraries/Native/Unix/Common/
+
+if STATIC_ICU
+nodist_libmonoruntime_shimglobalization_la_SOURCES += @ICU_SHIM_PATH@/pal_icushim_static.c
+libmonoruntime_shimglobalization_la_CFLAGS += -I$(ICU_REPO)/icu4c/source/wasm-usr/include/ \
+-L$(ICU_REPO)/icu4c/source/wasm-build/lib/ -licui18n -licuuc -licutu -licuio \
+-L$(ICU_REPO)/icu4c/source/wasm-build/stubdata -licudata
+endif
+
 endif
 endif
 

src/mono/wasm/Makefile

@@ -56,6 +56,17 @@ MONO_LIBS = \
 	$(MONO_BIN_DIR)/libmono-icall-table.a \
 	${SYS_NATIVE_DIR}/libSystem.Native.a
 
+ifeq ($(ICU_REPO),)
+	@echo "ICU_REPO is not set."
+else
+	MONO_LIBS := $(MONO_LIBS) \
+	$(ICU_REPO)/icu4c/source/wasm-build/lib/libicui18n.a \
+	$(ICU_REPO)/icu4c/source/wasm-build/lib/libicuio.a \
+	$(ICU_REPO)/icu4c/source/wasm-build/lib/libicuuc.a \
+	$(ICU_REPO)/icu4c/source/wasm-build/lib/libicutu.a \
+	$(ICU_REPO)/icu4c/source/wasm-build/stubdata/libicudata.a
+endif
+
 EMCC_FLAGS=--profiling-funcs -s WASM=1 -s ALLOW_MEMORY_GROWTH=1 -s BINARYEN=1 -s ALIASING_FUNCTION_POINTERS=0 -s NO_EXIT_RUNTIME=1 -s "EXTRA_EXPORTED_RUNTIME_METHODS=['ccall', 'FS_createPath', 'FS_createDataFile', 'cwrap', 'setValue', 'getValue', 'UTF8ToString', 'addFunction']" -s "EXPORTED_FUNCTIONS=['_putchar']" --source-map-base http://example.com  -s WASM_OBJECT_FILES=0 -s FORCE_FILESYSTEM=1 -s USE_ZLIB=1
 EMCC_DEBUG_FLAGS =-g -Os -s ASSERTIONS=1 -DENABLE_NETCORE=1 -DDEBUG=1
 EMCC_RELEASE_FLAGS=-Oz --llvm-opts 2 --llvm-lto 1 -DENABLE_NETCORE=1

src/mono/wasm/runtime/driver.c

@@ -322,7 +322,7 @@ mono_wasm_load_runtime (const char *managed_path, int enable_debugging)
 	monoeg_g_setenv ("COMPlus_DebugWriteToStdErr", "1", 0);
 #endif
 #ifdef ENABLE_NETCORE
-	monoeg_g_setenv ("DOTNET_SYSTEM_GLOBALIZATION_INVARIANT", "1", 0);
+	monoeg_g_setenv ("DOTNET_SYSTEM_GLOBALIZATION_INVARIANT", getenv("DOTNET_ICU_DIR") ? "0" : "1", 0);
 #endif
 
 	mini_parse_debug_option ("top-runtime-invoke-unhandled");

src/mono/wasm/wasm.targets

@@ -9,6 +9,7 @@
 
   <ItemGroup>
     <WasmPInvokeModules Include="libSystem.Native"/>
+    <WasmPInvokeModules Include="QCall"/>
     <WasmPInvokeAssemblies Include="$(MonoArtifactsPath)\System.Private.CoreLib.dll"/>
     <WasmPInvokeAssemblies Include="$(ArtifactsBinDir)\System.Runtime\$(NetCoreAppCurrent)-$(Configuration)\System.Runtime.dll"/>
     <WasmPInvokeAssemblies Include="$(ArtifactsBinDir)\System.Console\$(NetCoreAppCurrent)-Browser-$(Configuration)\System.Console.dll"/>