[Browser] Add mono_wasm_get_icudt_name API

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

@@ -3,6 +3,7 @@
 //
 
 #include <stdlib.h>
+#include <string.h>
 #include <stdio.h>
 #include "pal_icushim_internal.h"
 #include "pal_icushim.h"
@@ -11,6 +12,13 @@
 #include <unicode/localpointer.h>
 #include <unicode/utrace.h>
 
+#if defined(TARGET_UNIX)
+#include <strings.h>
+#elif defined(TARGET_WINDOWS)
+#define strcasecmp _stricmp
+#define strncasecmp _strnicmp
+#endif
+
 static int32_t isLoaded = 0;
 static int32_t isDataSet = 0;
 
@@ -30,6 +38,37 @@ static void U_CALLCONV icu_trace_data(const void* context, int32_t fnNumber, int
 #ifdef __EMSCRIPTEN__
 #include <emscripten.h>
 
+EMSCRIPTEN_KEEPALIVE const char* mono_wasm_get_icudt_name(const char* culture);
+
+EMSCRIPTEN_KEEPALIVE const char* mono_wasm_get_icudt_name(const char* culture)
+{
+    // Based on https://github.com/dotnet/icu/tree/maint/maint-67/icu-filters
+
+    // Use full one if culture is null or empty
+    if (!culture || strlen(culture) < 2)
+        return "icudt.dat";
+
+    // CJK: starts with "ja", "ko" or "zh"
+    if (!strncasecmp("ja", culture, 2) || 
+        !strncasecmp("ko", culture, 2) || 
+        !strncasecmp("zh", culture, 2))
+        return "icudt_CJK.dat"; // contains "en" as well.
+
+    // EFIGS
+    const char* efigsCultures[15] = {
+        "en-US", "fr-FR", "es-ES", "it-IT", "de-DE",
+        "en_US", "fr_FR", "es_ES", "it_IT", "de_DE",
+        "en",    "fr",    "es",    "it",    "de"
+    };
+
+    for (int i = 0; i < 15; i++)
+        if (!strcasecmp(culture, efigsCultures[i]))
+            return "icudt_EFIGS.dat";
+
+    // full except CJK cultures
+    return "icudt_no_CJK.dat";
+}
+
 EMSCRIPTEN_KEEPALIVE int32_t mono_wasm_load_icu_data(void * pData);
 
 EMSCRIPTEN_KEEPALIVE int32_t mono_wasm_load_icu_data(void * pData)

src/mono/wasm/runtime/library_mono.js

@@ -71,6 +71,7 @@ var MonoSupportLib = {
 			module ["mono_load_runtime_and_bcl_args"] = MONO.mono_load_runtime_and_bcl_args;
 			module ["mono_wasm_load_bytes_into_heap"] = MONO.mono_wasm_load_bytes_into_heap;
 			module ["mono_wasm_load_icu_data"] = MONO.mono_wasm_load_icu_data;
+			module ["mono_wasm_get_icudt_name"] = MONO.mono_wasm_get_icudt_name;
 			module ["mono_wasm_globalization_init"] = MONO.mono_wasm_globalization_init;
 			module ["mono_wasm_get_loaded_files"] = MONO.mono_wasm_get_loaded_files;
 			module ["mono_wasm_new_root_buffer"] = MONO.mono_wasm_new_root_buffer;
@@ -1379,6 +1380,14 @@ var MonoSupportLib = {
 			return ok;
 		},
 
+		// Get icudt.dat exact filename that matches given culture, examples:
+		//   "ja" -> "icudt_CJK.dat"
+		//   "en_US" (or "en-US" or just "en") -> "icudt_EFIGS.dat"
+		// etc, see "mono_wasm_get_icudt_name" implementation in pal_icushim_static.c
+		mono_wasm_get_icudt_name: function (culture) {
+			return Module.ccall ('mono_wasm_get_icudt_name', 'string', ['string'], [culture]);
+		},
+
 		_finalize_startup: function (args, ctx) {
 			var loaded_files_with_debug_info = [];