git: Add script to check localized texts occurrences in code (#1511)

This script will be executed on every CI run as part of tests, and will
ensure no unlocalised strings are present in the code

Note that texts without the `_lang` suffix will not be checked, e.g.
96fe608d60/plugins/builtin/source/content/views/view_provider_settings.cpp (L10)

.github/workflows/tests.yml

@@ -61,3 +61,17 @@ jobs:
       run: |
         cd build
         ctest --output-on-failure
+
+  langs:
+    name: 🧪 Langs
+    runs-on: ubuntu-22.04
+
+    steps:
+    - name: 🧰 Checkout
+      uses: actions/checkout@v3
+      with:
+        submodules: recursive
+
+    - name: Check langs
+      run:
+        python3 tests/check_langs.py

plugins/builtin/romfs/lang/en_US.json

@@ -148,6 +148,7 @@
         "hex.builtin.menu.file.import.pattern": "Pattern File",
         "hex.builtin.menu.file.import.data_processor": "Data Processor Workspace",
         "hex.builtin.menu.file.import.custom_encoding": "Custom Encoding File",
+        "hex.builtin.menu.file.import.modified_file.popup.invalid_size": "File selected do not have the same size as the current file. Both sizes must match",
         "hex.builtin.menu.file.open_file": "Open File...",
         "hex.builtin.menu.file.open_other": "Open Other...",
         "hex.builtin.menu.file.project": "Project",

plugins/builtin/source/content/out_of_box_experience.cpp

@@ -205,7 +205,7 @@ namespace hex::plugin::builtin {
 
                             // Draw table containing everything that's being reported
                             if (ImGui::CollapsingHeader("hex.builtin.oobe.server_contact.data_collected_title"_lang)) {
-                                if (ImGui::BeginTable("hex.builtin.oobe.server_contact.data_collected_table"_lang, 2,
+                                if (ImGui::BeginTable("hex.builtin.oobe.server_contact.data_collected_table", 2,
                                                      ImGuiTableFlags_Borders | ImGuiTableFlags_RowBg | ImGuiTableFlags_ScrollY | ImGuiTableFlags_NoHostExtendY,
                                                      ImVec2(ImGui::GetContentRegionAvail().x, 150_scaled))) {
                                     ImGui::TableSetupColumn("hex.builtin.oobe.server_contact.data_collected_table.key"_lang);

tests/check_langs.py

@@ -0,0 +1,72 @@
+#!/usr/bin/env python3
+import json
+import re
+import os
+import sys
+
+SHOW_UNUSED_LANGS = "--unused" in sys.argv
+
+# use a regex on code to get all "hex.lang.id"_lang occurences
+def get_lang_occurences_in_code(path):
+    for dir, _, files in os.walk(path):
+        for file in files:
+
+            if not os.path.splitext(file)[1] in (".cpp", ".c", ".hpp", ".h"):
+                continue
+            filepath = os.path.join(dir, file)
+            with open(filepath) as file:
+                for line_num, line in enumerate(file):
+                    for m in re.finditer('"([^"]*?)"_lang', line):
+                        yield (filepath, line_num+1, m.group(1))
+
+# Get langs in a specific json file
+def get_langs(filepath) -> list[str]:
+    if filepath == None:
+        return []
+    elif not os.path.exists(filepath):
+        print(f"Warning: no langs file found at {filepath}")
+        return []
+
+    with open(filepath, "r") as file:
+        data = json.loads(file.read())
+        existing_langs = []
+
+        for key, _ in data["translations"].items():
+            existing_langs.append(key)
+
+        return existing_langs
+
+def check_langs(code_path, bonus_langs, specific_langs_path):
+    print(f"\n--- Checking langs at {code_path}")
+ 
+    specific_langs = get_langs(specific_langs_path)
+    unused_langs = specific_langs.copy()
+    ret = True
+
+    for filepath, line, match in get_lang_occurences_in_code(code_path):
+        try:
+            unused_langs.remove(match)
+        except ValueError:
+            pass
+
+        if not match in bonus_langs + specific_langs:
+            ret = False
+            print(f"Problem: Lang '{match}' at {filepath}:{line} not found")
+    
+    if SHOW_UNUSED_LANGS and len(unused_langs) > 0:
+        print(f"Unused langs in {specific_langs_path}:")
+        for unused_lang in unused_langs:
+            print(unused_lang)
+    return ret
+
+
+ui_langs = get_langs("./plugins/ui/romfs/lang/en_US.json")
+
+exit_ok = True
+exit_ok &= check_langs("./main", ui_langs, None)
+
+for plugin in os.listdir("./plugins"):
+    if plugin == "ui": continue
+    exit_ok &= check_langs(f"./plugins/{plugin}", ui_langs, f"./plugins/{plugin}/romfs/lang/en_US.json")
+
+sys.exit(0 if exit_ok else 1)