Make default resources a better match for en-US requests

When locale fallback landed, resources which specified an 'English'
locale started to be considered a better match for en-US, even though
traditionally, apps tend to ship US English resources under their
default locale.

This fixes that, and makes en-US requests match default locales.

Bug: 26756573
Bug: 26789680
Bug: 26803868
Change-Id: I460c276bfc6ddba0439dcdf87497a0aece0fa05d

4 files changed, 97 insertions, 4 deletions

diff --git a/include/androidfw/LocaleData.h b/include/androidfw/LocaleData.h
index add0ab5eda3f..b14829d07f04 100644
--- a/include/androidfw/LocaleData.h
+++ b/include/androidfw/LocaleData.h
@@ -29,6 +29,8 @@ int localeDataCompareRegions(
 
 void localeDataComputeScript(char out[4], const char* language, const char* region);
 
+bool localeDataIsCloseToUsEnglish(const char* region);
+
 } // namespace android
 
 #endif // _LIBS_UTILS_LOCALE_DATA_H
diff --git a/libs/androidfw/LocaleData.cpp b/libs/androidfw/LocaleData.cpp
index c0c3ab883a42..038ef5839fe2 100644
--- a/libs/androidfw/LocaleData.cpp
+++ b/libs/androidfw/LocaleData.cpp
@@ -70,13 +70,17 @@ uint32_t findParent(uint32_t packed_locale, const char* script) {
 //
 // Returns the number of ancestors written in the output, which is always
 // at least one.
+//
+// (If 'out' is nullptr, we do everything the same way but we simply don't write
+// any results in 'out'.)
 size_t findAncestors(uint32_t* out, ssize_t* stop_list_index,
                      uint32_t packed_locale, const char* script,
                      const uint32_t* stop_list, size_t stop_set_length) {
     uint32_t ancestor = packed_locale;
     size_t count = 0;
     do {
-        out[count++] = ancestor;
+        if (out != nullptr) out[count] = ancestor;
+        count++;
         for (size_t i = 0; i < stop_set_length; i++) {
             if (stop_list[i] == ancestor) {
                 *stop_list_index = (ssize_t) i;
@@ -93,10 +97,9 @@ size_t findDistance(uint32_t supported,
                     const char* script,
                     const uint32_t* request_ancestors,
                     size_t request_ancestors_count) {
-    uint32_t supported_ancestors[MAX_PARENT_DEPTH+1];
     ssize_t request_ancestors_index;
     const size_t supported_ancestor_count = findAncestors(
-            supported_ancestors, &request_ancestors_index,
+            nullptr, &request_ancestors_index,
             supported, script,
             request_ancestors, request_ancestors_count);
     // Since both locales share the same root, there will always be a shared
@@ -198,4 +201,19 @@ void localeDataComputeScript(char out[4], const char* language, const char* regi
     }
 }
 
+const uint32_t ENGLISH_STOP_LIST[2] = {
+    0x656E0000lu, // en
+    0x656E8400lu, // en-001
+};
+const char ENGLISH_CHARS[2] = {'e', 'n'};
+const char LATIN_CHARS[4] = {'L', 'a', 't', 'n'};
+
+bool localeDataIsCloseToUsEnglish(const char* region) {
+    const uint32_t locale = packLocale(ENGLISH_CHARS, region);
+    ssize_t stop_list_index;
+    findAncestors(nullptr, &stop_list_index, locale, LATIN_CHARS, ENGLISH_STOP_LIST, 2);
+    // A locale is like US English if we see "en" before "en-001" in its ancestor list.
+    return stop_list_index == 0; // 'en' is first in ENGLISH_STOP_LIST
+}
+
 } // namespace android
diff --git a/libs/androidfw/ResourceTypes.cpp b/libs/androidfw/ResourceTypes.cpp
index 71e9c92ae726..1d9fe35bac35 100644
--- a/libs/androidfw/ResourceTypes.cpp
+++ b/libs/androidfw/ResourceTypes.cpp
@@ -2196,7 +2196,32 @@ bool ResTable_config::isLocaleBetterThan(const ResTable_config& o,
         // The languages of the two resources are not the same. We can only
         // assume that one of the two resources matched the request because one
         // doesn't have a language and the other has a matching language.
-        return (language[0] != 0);
+        //
+        // We consider the one that has the language specified a better match.
+        //
+        // The exception is that we consider no-language resources a better match
+        // for US English and similar locales than locales that are a descendant
+        // of Internatinal English (en-001), since no-language resources are
+        // where the US English resource have traditionally lived for most apps.
+        if (requested->language[0] == 'e' && requested->language[1] == 'n') {
+            if (requested->country[0] == 'U' && requested->country[1] == 'S') {
+                // For US English itself, we consider a no-locale resource a
+                // better match if the other resource has a country other than
+                // US specified.
+                if (language[0] != '\0') {
+                    return country[0] == '\0' || (country[0] == 'U' && country[1] == 'S');
+                } else {
+                    return !(o.country[0] == '\0' || (o.country[0] == 'U' && o.country[1] == 'S'));
+                }
+            } else if (localeDataIsCloseToUsEnglish(requested->country)) {
+                if (language[0] != '\0') {
+                    return localeDataIsCloseToUsEnglish(country);
+                } else {
+                    return !localeDataIsCloseToUsEnglish(o.country);
+                }
+            }
+        }
+        return (language[0] != '\0');
     }
 
     // If we are here, both the resources have the same non-empty language as
diff --git a/libs/androidfw/tests/ConfigLocale_test.cpp b/libs/androidfw/tests/ConfigLocale_test.cpp
index 1941563519c8..7b386404fe8c 100644
--- a/libs/androidfw/tests/ConfigLocale_test.cpp
+++ b/libs/androidfw/tests/ConfigLocale_test.cpp
@@ -592,4 +592,52 @@ TEST(ConfigLocaleTest, isLocaleBetterThan_regionComparison) {
     EXPECT_FALSE(config2.isLocaleBetterThan(config1, &request));
 }
 
+// Default resources are considered better matches for US English
+// and US-like English locales than International English locales
+TEST(ConfigLocaleTest, isLocaleBetterThan_UsEnglishIsSpecial) {
+    ResTable_config config1, config2, request;
+
+    fillIn("en", "US", NULL, NULL, &request);
+    fillIn(NULL, NULL, NULL, NULL, &config1);
+    fillIn("en", "001", NULL, NULL, &config2);
+    // default is better than International English
+    EXPECT_TRUE(config1.isLocaleBetterThan(config2, &request));
+    EXPECT_FALSE(config2.isLocaleBetterThan(config1, &request));
+
+    fillIn("en", "US", NULL, NULL, &request);
+    fillIn(NULL, NULL, NULL, NULL, &config1);
+    fillIn("en", "GB", NULL, NULL, &config2);
+    // default is better than British English
+    EXPECT_TRUE(config1.isLocaleBetterThan(config2, &request));
+    EXPECT_FALSE(config2.isLocaleBetterThan(config1, &request));
+
+    fillIn("en", "PR", NULL, NULL, &request);
+    fillIn(NULL, NULL, NULL, NULL, &config1);
+    fillIn("en", "001", NULL, NULL, &config2);
+    // Even for Puerto Rico, default is better than International English
+    EXPECT_TRUE(config1.isLocaleBetterThan(config2, &request));
+    EXPECT_FALSE(config2.isLocaleBetterThan(config1, &request));
+
+    fillIn("en", "US", NULL, NULL, &request);
+    fillIn("en", NULL, NULL, NULL, &config1);
+    fillIn(NULL, NULL, NULL, NULL, &config2);
+    // "English" is better than default, since it's a parent of US English
+    EXPECT_TRUE(config1.isLocaleBetterThan(config2, &request));
+    EXPECT_FALSE(config2.isLocaleBetterThan(config1, &request));
+
+    fillIn("en", "PR", NULL, NULL, &request);
+    fillIn("en", NULL, NULL, NULL, &config1);
+    fillIn(NULL, NULL, NULL, NULL, &config2);
+    // "English" is better than default, since it's a parent of Puerto Rico English
+    EXPECT_TRUE(config1.isLocaleBetterThan(config2, &request));
+    EXPECT_FALSE(config2.isLocaleBetterThan(config1, &request));
+
+    fillIn("en", "US", NULL, NULL, &request);
+    fillIn(NULL, NULL, NULL, NULL, &config1);
+    fillIn("en", "PR", NULL, NULL, &config2);
+    // For US English itself, we prefer default to its siblings in the parent tree
+    EXPECT_TRUE(config1.isLocaleBetterThan(config2, &request));
+    EXPECT_FALSE(config2.isLocaleBetterThan(config1, &request));
+}
+
 }  // namespace android