[DRAFT] An approach to linting the translations

Makefile

@@ -146,7 +146,7 @@ hyper$(OBJ_EXTENSION): *.cpp language-data.cpp autohdr.h
 hyper.res: hyper.rc hr-icon.ico
 	windres hyper.rc -O coff -o hyper.res
 
-langen$(EXE_EXTENSION): langen.cpp language-??.cpp language-ptbr.cpp
+langen$(EXE_EXTENSION): langen.cpp language-??.cpp language-ptbr.cpp all-string-literals.ipp
 	$(CXX) -O0 $(CXXFLAGS) $(langen_CXXFLAGS) langen.cpp $(LDFLAGS) -o $@
 
 makeh$(EXE_EXTENSION): makeh.cpp
@@ -155,6 +155,9 @@ makeh$(EXE_EXTENSION): makeh.cpp
 autohdr.h: makeh$(EXE_EXTENSION) language-data.cpp *.cpp
 	./makeh classes.cpp locations.cpp colors.cpp hyperpoint.cpp geometry.cpp goldberg.cpp init.cpp floorshapes.cpp cell.cpp multi.cpp shmup.cpp pattern2.cpp mapeditor.cpp graph.cpp textures.cpp hprint.cpp language.cpp util.cpp complex.cpp *.cpp > autohdr.h
 
+all-string-literals.ipp: lanlint.py *.cpp
+	python lanlint.py > all-string-literals.ipp
+
 language-data.cpp: langen$(EXE_EXTENSION)
 	./langen > language-data.cpp
 
@@ -179,3 +182,4 @@ clean:
 	rm -rf mymake$(EXE_EXTENSION) mymake_files/
 	rm -f hyperrogue$(EXE_EXTENSION) hyper$(OBJ_EXTENSION) $(hyper_RES) savepng$(OBJ_EXTENSION)
 	rm -f hyper.html hyper.js hyper.wasm
+	rm -f all-string-literals.ipp

langen.cpp

@@ -21,6 +21,7 @@
 #endif
 
 template<class T> int isize(const T& x) { return x.size(); }
+template<class T, int N> int isize(const T (&x)[N]) { return N; }
 
 #define NUMLAN 7
 
@@ -30,9 +31,75 @@ std::string current_language;
 
 const char *escape(std::string s, const std::string& dft);
 
+#include "all-string-literals.ipp"
+
+int edit_distance(const std::string &s1, const std::string &s2, int bail)
+{
+  const int m = s1.size();
+  const int n = s2.size();
+  if( m==0 ) return n;
+  if( n==0 ) return m;
+  std::vector<int> costs(n+1);
+  for (int k=0; k<=n; k++) costs[k] = k;
+  int i = 0;
+  for (auto it1 = s1.begin(); it1 != s1.end(); ++it1, ++i ) {
+    costs[0] = i+1;
+    int corner = i;
+    int j = 0;
+    for (auto it2 = s2.begin(); it2 != s2.end(); ++it2, ++j ) {
+      int upper = costs[j+1];
+      if (toupper(*it1) == toupper(*it2)) {
+        costs[j+1] = corner;
+      } else {
+        costs[j+1] = std::min(costs[j], std::min(upper, corner)) + 1;
+      }
+      corner = upper;
+    }
+  }
+  return costs[n];
+}
+
+static auto program_string_set = std::set<std::string>(
+    program_strings, program_strings + isize(program_strings)
+);
+static std::map<std::string, std::vector<std::string>> useless_translations;
+
+static void check_program_strings_for(const std::string& s)
+{
+    if (program_string_set.find(s) != program_string_set.end()) {
+        return;
+    }
+    useless_translations[s].push_back(current_language);
+}
+
+static void print_useless_translations()
+{
+    for (auto&& kv : useless_translations) {
+        const std::string& s = kv.first;
+
+        int min_d = INT_MAX;
+        std::string min_p;
+        for (const std::string& p : program_string_set) {
+            int d = edit_distance(p, s, min_d);
+            if (d < min_d) {
+                min_d = d;
+                min_p = p;
+            }
+        }
+        std::string langs = kv.second[0];
+        for (size_t i=1; i < kv.second.size(); ++i) {
+            langs += ',';
+            langs += kv.second[i];
+        }
+        fprintf(stderr, "Unused translation in %s: %s\n", langs.c_str(), escape(s, s));
+        fprintf(stderr, "Closest match in program: %s\n", escape(min_p, min_p));
+    }
+}
+
 template<class T> struct dictionary {
   std::map<std::string, T> m;
   void add(const std::string& s, T val) {
+    check_program_strings_for(s);
     auto it = m.find(s);
     if (it == m.end()) {
       m.emplace(s, std::move(val));
@@ -443,4 +510,6 @@ int main() {
 
   printf("  };\n");
 
+  print_useless_translations();
+
   }

lanlint.py

@@ -0,0 +1,56 @@
+#!/usr/bin/env python
+
+import glob
+
+def get_file_strings(fname):
+    result = []
+    with open(fname, "r") as f:
+        text = f.read()
+    i = 0
+    start = None
+    current_string = None
+    while i < len(text):
+        if (start is None) and (text[i:i+2] == '//'):
+            while text[i] != '\n':
+                i += 1
+        elif (start is None) and (text[i] == '"'):
+            if (text[i-1] == 'R'):
+                i += 1
+                while text[i] != '"': i += 1
+            else:
+                start = i
+        elif (start is None) and (text[i] in ' \t\n'):
+            pass
+        elif (start is None) and (text[i:i+3] == "'\"'"):
+            i += 2  # skip over character literals that might otherwise confuse us
+        elif (start is None):
+            if current_string:
+                result += [current_string]
+                # print(current_string)
+                # print("---------------------")
+            current_string = ''
+        elif (start is not None) and (text[i] == '\\'):
+            i += 1
+        elif (start is not None) and (text[i] == '"'):
+            current_string += text[start:i+1]
+            start = None
+        else:
+            pass  # just a plain old string character
+        i += 1
+    if current_string:
+        result += [current_string]
+    return result
+
+def get_program_strings():
+    result = []
+    for fname in glob.glob("*.cpp"):
+        if not fname.startswith("language-"):
+            result += get_file_strings(fname)
+    return result
+
+if __name__ == '__main__':
+    program_strings = set(get_program_strings())
+    print('const char *program_strings[] = {')
+    for s in program_strings:
+        print('    %s,' % s)
+    print('};')