Merge pull request #91937 from bruvzg/icu_751

Update ICU to 75.1

COPYRIGHT.txt

@@ -285,7 +285,7 @@ License: HarfBuzz
 
 Files: ./thirdparty/icu4c/
 Comment: International Components for Unicode
-Copyright: 1991-2021, Unicode
+Copyright: 2016-2024, Unicode, Inc.
 License: Unicode
 
 Files: ./thirdparty/jpeg-compressor/

modules/text_server_adv/SCsub

@@ -468,7 +468,7 @@ if env["builtin_icu4c"]:
     ]
     thirdparty_sources = [thirdparty_dir + file for file in thirdparty_sources]
 
-    icu_data_name = "icudt74l.dat"
+    icu_data_name = "icudt75l.dat"
 
     if env.editor_build:
         env_icu.Depends("#thirdparty/icu4c/icudata.gen.h", "#thirdparty/icu4c/" + icu_data_name)

modules/text_server_adv/gdextension_build/SConstruct

@@ -703,7 +703,7 @@ thirdparty_icu_sources = [
 ]
 thirdparty_icu_sources = [thirdparty_icu_dir + file for file in thirdparty_icu_sources]
 
-icu_data_name = "icudt74l.dat"
+icu_data_name = "icudt75l.dat"
 
 if env["static_icu_data"]:
     env_icu.Depends("../../../thirdparty/icu4c/icudata.gen.h", "../../../thirdparty/icu4c/" + icu_data_name)

thirdparty/README.md

@@ -391,7 +391,7 @@ Files extracted from upstream source:
 ## icu4c
 
 - Upstream: https://github.com/unicode-org/icu
-- Version: 74.2 (2d029329c82c7792b985024b2bdab5fc7278fbc8, 2023)
+- Version: 75.1 (7750081bda4b3bc1768ae03849ec70f67ea10625, 2024)
 - License: Unicode
 
 Files extracted from upstream source:
@@ -403,7 +403,7 @@ Files extracted from upstream source:
 
 Files generated from upstream source:
 
-- The `icudt74l.dat` built with the provided `godot_data.json` config file (see
+- The `icudt75l.dat` built with the provided `godot_data.json` config file (see
   https://github.com/unicode-org/icu/blob/master/docs/userguide/icu_data/buildtool.md
   for instructions).
 
@@ -413,7 +413,7 @@ Files generated from upstream source:
 3. Reconfigure ICU with custom data config:
    `ICU_DATA_FILTER_FILE={GODOT_SOURCE}/thirdparty/icu4c/godot_data.json ./runConfigureICU {PLATFORM} --with-data-packaging=common`
 4. Delete `data/out` folder and rebuild data: `cd data && rm -rf ./out && make`
-5. Copy `source/data/out/icudt74l.dat` to the `{GODOT_SOURCE}/thirdparty/icu4c/icudt74l.dat`
+5. Copy `source/data/out/icudt75l.dat` to the `{GODOT_SOURCE}/thirdparty/icu4c/icudt75l.dat`
 
 
 ## jpeg-compressor

thirdparty/icu4c/LICENSE

@@ -2,7 +2,7 @@ UNICODE LICENSE V3
 
 COPYRIGHT AND PERMISSION NOTICE
 
-Copyright © 2016-2023 Unicode, Inc.
+Copyright © 2016-2024 Unicode, Inc.
 
 NOTICE TO USER: Carefully read the following legal agreement. BY
 DOWNLOADING, INSTALLING, COPYING OR OTHERWISE USING DATA FILES, AND/OR
@@ -38,6 +38,8 @@ not be used in advertising or otherwise to promote the sale, use or other
 dealings in these Data Files or Software without prior written
 authorization of the copyright holder.
 
+SPDX-License-Identifier: Unicode-3.0
+
 ----------------------------------------------------------------------
 
 Third-Party Software Licenses

thirdparty/icu4c/common/brkeng.cpp

@@ -114,13 +114,11 @@ UnhandledEngine::handleCharacter(UChar32 c) {
  */
 
 ICULanguageBreakFactory::ICULanguageBreakFactory(UErrorCode &/*status*/) {
-    fEngines = 0;
+    fEngines = nullptr;
 }
 
 ICULanguageBreakFactory::~ICULanguageBreakFactory() {
-    if (fEngines != 0) {
-        delete fEngines;
-    }
+    delete fEngines;
 }
 
 void ICULanguageBreakFactory::ensureEngines(UErrorCode& status) {

thirdparty/icu4c/common/brkiter.cpp

@@ -438,17 +438,14 @@ BreakIterator::makeInstance(const Locale& loc, int32_t kind, UErrorCode& status)
             UTRACE_ENTRY(UTRACE_UBRK_CREATE_LINE);
             uprv_strcpy(lb_lw, "line");
             UErrorCode kvStatus = U_ZERO_ERROR;
-            CharString value;
-            CharStringByteSink valueSink(&value);
-            loc.getKeywordValue("lb", valueSink, kvStatus);
+            auto value = loc.getKeywordValue<CharString>("lb", kvStatus);
             if (U_SUCCESS(kvStatus) && (value == "strict" || value == "normal" || value == "loose")) {
                 uprv_strcat(lb_lw, "_");
                 uprv_strcat(lb_lw, value.data());
             }
             // lw=phrase is only supported in Japanese and Korean
             if (uprv_strcmp(loc.getLanguage(), "ja") == 0 || uprv_strcmp(loc.getLanguage(), "ko") == 0) {
-                value.clear();
-                loc.getKeywordValue("lw", valueSink, kvStatus);
+                value = loc.getKeywordValue<CharString>("lw", kvStatus);
                 if (U_SUCCESS(kvStatus) && value == "phrase") {
                     uprv_strcat(lb_lw, "_");
                     uprv_strcat(lb_lw, value.data());
@@ -500,7 +497,7 @@ BreakIterator::makeInstance(const Locale& loc, int32_t kind, UErrorCode& status)
 Locale
 BreakIterator::getLocale(ULocDataLocaleType type, UErrorCode& status) const {
     if (type == ULOC_REQUESTED_LOCALE) {
-        return Locale(requestLocale);
+        return {requestLocale};
     }
     U_LOCALE_BASED(locBased, *this);
     return locBased.getLocale(type, status);

thirdparty/icu4c/common/bytesinkutil.h

@@ -7,18 +7,52 @@
 #ifndef BYTESINKUTIL_H
 #define BYTESINKUTIL_H
 
+#include <type_traits>
+
 #include "unicode/utypes.h"
 #include "unicode/bytestream.h"
 #include "unicode/edits.h"
+#include "charstr.h"
 #include "cmemory.h"
 #include "uassert.h"
+#include "ustr_imp.h"
 
 U_NAMESPACE_BEGIN
 
 class ByteSink;
-class CharString;
 class Edits;
 
+class U_COMMON_API CharStringByteSink : public ByteSink {
+public:
+    CharStringByteSink(CharString* dest);
+    ~CharStringByteSink() override;
+
+    CharStringByteSink() = delete;
+    CharStringByteSink(const CharStringByteSink&) = delete;
+    CharStringByteSink& operator=(const CharStringByteSink&) = delete;
+
+    void Append(const char* bytes, int32_t n) override;
+
+    char* GetAppendBuffer(int32_t min_capacity,
+                          int32_t desired_capacity_hint,
+                          char* scratch,
+                          int32_t scratch_capacity,
+                          int32_t* result_capacity) override;
+
+private:
+    CharString& dest_;
+};
+
+// CharString doesn't provide the public API that StringByteSink requires a
+// string class to have so this template specialization replaces the default
+// implementation of StringByteSink<CharString> with CharStringByteSink.
+template<>
+class StringByteSink<CharString> : public CharStringByteSink {
+ public:
+  StringByteSink(CharString* dest) : CharStringByteSink(dest) { }
+  StringByteSink(CharString* dest, int32_t /*initialAppendCapacity*/) : CharStringByteSink(dest) { }
+};
+
 class U_COMMON_API ByteSinkUtil {
 public:
     ByteSinkUtil() = delete;  // all static
@@ -57,30 +91,64 @@ class U_COMMON_API ByteSinkUtil {
                                  ByteSink &sink, uint32_t options, Edits *edits,
                                  UErrorCode &errorCode);
 
-private:
-    static void appendNonEmptyUnchanged(const uint8_t *s, int32_t length,
-                                        ByteSink &sink, uint32_t options, Edits *edits);
-};
-
-class U_COMMON_API CharStringByteSink : public ByteSink {
-public:
-    CharStringByteSink(CharString* dest);
-    ~CharStringByteSink() override;
-
-    CharStringByteSink() = delete;
-    CharStringByteSink(const CharStringByteSink&) = delete;
-    CharStringByteSink& operator=(const CharStringByteSink&) = delete;
-
-    void Append(const char* bytes, int32_t n) override;
+    /**
+     * Calls a lambda that writes to a ByteSink with a CheckedArrayByteSink
+     * and then returns through u_terminateChars(), in order to implement
+     * the classic ICU4C C API writing to a fix sized buffer on top of a
+     * contemporary C++ API.
+     *
+     * @param buffer receiving buffer
+     * @param capacity capacity of receiving buffer
+     * @param lambda that gets called with the sink as an argument
+     * @param status set to U_BUFFER_OVERFLOW_ERROR on overflow
+     * @return number of bytes written, or needed (in case of overflow)
+     * @internal
+     */
+    template <typename F,
+              typename = std::enable_if_t<
+                  std::is_invocable_r_v<void, F, ByteSink&, UErrorCode&>>>
+    static int32_t viaByteSinkToTerminatedChars(char* buffer, int32_t capacity,
+                                                F&& lambda,
+                                                UErrorCode& status) {
+        if (U_FAILURE(status)) { return 0; }
+        CheckedArrayByteSink sink(buffer, capacity);
+        lambda(sink, status);
+        if (U_FAILURE(status)) { return 0; }
+
+        int32_t reslen = sink.NumberOfBytesAppended();
+
+        if (sink.Overflowed()) {
+            status = U_BUFFER_OVERFLOW_ERROR;
+            return reslen;
+        }
+
+        return u_terminateChars(buffer, capacity, reslen, &status);
+    }
 
-    char* GetAppendBuffer(int32_t min_capacity,
-                          int32_t desired_capacity_hint,
-                          char* scratch,
-                          int32_t scratch_capacity,
-                          int32_t* result_capacity) override;
+    /**
+     * Calls a lambda that writes to a ByteSink with a CharStringByteSink and
+     * then returns a CharString, in order to implement a contemporary C++ API
+     * on top of a C/C++ compatibility ByteSink API.
+     *
+     * @param lambda that gets called with the sink as an argument
+     * @param status to check and report
+     * @return the resulting string, or an empty string (in case of error)
+     * @internal
+     */
+    template <typename F,
+              typename = std::enable_if_t<
+                  std::is_invocable_r_v<void, F, ByteSink&, UErrorCode&>>>
+    static CharString viaByteSinkToCharString(F&& lambda, UErrorCode& status) {
+        if (U_FAILURE(status)) { return {}; }
+        CharString result;
+        CharStringByteSink sink(&result);
+        lambda(sink, status);
+        return result;
+    }
 
 private:
-    CharString& dest_;
+    static void appendNonEmptyUnchanged(const uint8_t *s, int32_t length,
+                                        ByteSink &sink, uint32_t options, Edits *edits);
 };
 
 U_NAMESPACE_END