[android, tools] remove unused XML strings; add script to find unused XML strings (#2777)

Signed-off-by: lizzie <lizzie@eden-emu.dev>
Co-authored-by: crueter <crueter@eden-emu.dev>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/2777
Reviewed-by: crueter <crueter@eden-emu.dev>
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
Co-authored-by: lizzie <lizzie@eden-emu.dev>
Co-committed-by: lizzie <lizzie@eden-emu.dev>

3 files changed, 89 insertions, 0 deletions

diff --git a/tools/README.md b/tools/README.md
index 35fd7a390c..061e95bc9c 100644
--- a/tools/README.md
+++ b/tools/README.md
@@ -8,6 +8,7 @@ Tools for Eden and other subprojects.
 
 ## Eden
 
+- `find-unused-strings.pl`: Find unused strings (for Android XML files).
 - `shellcheck.sh`: Ensure POSIX compliance (and syntax sanity) for all tools in this directory and subdirectories.
 - `llvmpipe-run.sh`: Sets environment variables needed to run any command (or Eden) with llvmpipe.
 - `optimize-assets.sh`: Optimize PNG assets with OptiPng.
@@ -20,6 +21,12 @@ Tools for Eden and other subprojects.
 - `clang-format.sh`: Runs `clang-format` on the entire codebase.
     * Requires: clang
 
+## Android
+It's recommended to run these scritps after almost any Android change, as they are relatively fast and important both for APK bloat and CI.
+
+- `unused-strings.sh`: Finds unused strings in `strings.xml` files.
+- `stale-translations.sh`: Finds translated strings that aren't present in the source `strings.xml` file.
+
 ## Translations
 
 - [Translation Scripts](./translations)
\ No newline at end of file
diff --git a/tools/stale-translations.sh b/tools/stale-translations.sh
new file mode 100755
index 0000000000..791b857a36
--- /dev/null
+++ b/tools/stale-translations.sh
@@ -0,0 +1,34 @@
+#!/bin/sh -e
+
+# SPDX-FileCopyrightText: Copyright 2025 Eden Emulator Project
+# SPDX-License-Identifier: GPL-3.0-or-later
+
+ANDROID=src/android/app/src/main
+STRINGS=$ANDROID/res/values/strings.xml
+
+TMP_DIR=$(mktemp -d)
+
+SRC="$TMP_DIR"/src
+LOCALES="$TMP_DIR"/locales
+TRANSLATED="$TMP_DIR"/translated
+COMBINED="$TMP_DIR"/combined
+
+# We start out by getting the list of source strings so we can compare all locale strings to this
+grep -e "string name" $STRINGS | grep -v 'translatable="false"' | cut -d'"' -f2 > "$SRC"
+
+# ... then we determine what locales to check
+find $ANDROID/res/values-* -name 'strings.xml' > "$LOCALES"
+
+while IFS= read -r locale; do
+	echo "Locale $(echo "$locale" | cut -d"-" -f2- | cut -d"/" -f1)"
+	grep -e "string name" "$locale" | cut -d'"' -f2 > "$TRANSLATED"
+	cat "$TRANSLATED" "$SRC" | sort | uniq -u > "$COMBINED"
+
+	# This will also include strings that are present in the source but NOT translated, so we filter those out now:
+	while IFS= read -r unused; do
+		{ grep -e "string name=\"$unused\"" "$STRINGS" >/dev/null; } || \
+		{ echo "-- Removing unused translation $unused" && sed "/string name=\"$unused\"/d" "$locale" > "$locale.new" && mv "$locale.new" "$locale"; }
+	done < "$COMBINED"
+
+	echo "-- done"
+done < "$LOCALES"
\ No newline at end of file
diff --git a/tools/unused-strings.sh b/tools/unused-strings.sh
new file mode 100755
index 0000000000..ed8c5d9584
--- /dev/null
+++ b/tools/unused-strings.sh
@@ -0,0 +1,48 @@
+#!/bin/sh -e
+
+# SPDX-FileCopyrightText: Copyright 2025 Eden Emulator Project
+# SPDX-License-Identifier: GPL-3.0-or-later
+
+ANDROID=src/android/app/src/main
+STRINGS=$ANDROID/res/values/strings.xml
+
+TMP_DIR=$(mktemp -d)
+
+USED="$TMP_DIR"/used
+UNUSED="$TMP_DIR"/unused
+FILTERED="$TMP_DIR"/filtered
+ALL_STRINGS="$TMP_DIR"/all
+
+# First we find what files to search and what strings are defined
+find $ANDROID -type f -iname '*.xml' -a -not -iname '*strings.xml' -o -iname '*.kt' | grep -v drawable >"$TMP_DIR"/files
+grep -e "string name" $STRINGS | cut -d'"' -f2 >"$ALL_STRINGS"
+
+# then get a list of all strings that are actually used
+set +e
+while IFS= read -r file; do
+    grep -o 'R\.string\.[a-zA-Z0-9_]\+' "$file" >> "$USED"
+    grep -o '@string/[a-zA-Z0-9_]\+' "$file" >> "$USED"
+done <"$TMP_DIR"/files
+set -e
+
+# filter out "@string/" and "R.string." from the strings to get the raw names
+sed 's/R.string.\|@string\///' "$USED" | sort -u | grep -v app_name_suffixed > "$FILTERED"
+
+# now we run a sort + uniq -u pass - this basically removes all strings that are
+# present in BOTH the used strings list AND strings.xml
+# thus giving us strings that are either ONLY used in strings.xml OR in the files
+
+# NB: This also gives strings that may be found in comments but nowhere else,
+# as well as some android builtins, but sed-ing those out of strings.xml is a no-op
+
+# e.g. appbar_scrolling_view_behavior
+cat "$FILTERED" "$ALL_STRINGS" | sort | uniq -u > "$UNUSED"
+
+# finally, print out all unused strings and remove them from ALL strings.xml definitions
+while IFS= read -r string; do
+	echo "$string"
+    find $ANDROID/res -iname '*strings.xml' | while IFS= read -r file; do
+        sed "/string name=\"$string\"/d" "$file" > "$file.new"
+		mv "$file.new" "$file"
+	done
+done < "$UNUSED"