Fix range-adaptor handling of an incomplete sequence.

paulhuggett · Jan 13, 2024 · a3ee81c · a3ee81c
1 parent 18585f9
commit a3ee81c
Show file tree

Hide file tree

Showing 2 changed files with 25 additions and 2 deletions.
diff --git a/include/icubaby/icubaby.hpp b/include/icubaby/icubaby.hpp
@@ -1045,7 +1045,9 @@ class transcode_view<FromEncoding, ToEncoding, View>::iterator {
   constexpr std::ranges::iterator_t<View> const& base () const& noexcept { return current_; }
   constexpr std::ranges::iterator_t<View> base () && { return std::move (current_); }
 
-  constexpr value_type const& operator* () const { return state_.front (); }
+  constexpr value_type const& operator* () const {
+    return state_.empty () ? replacement : state_.front ();
+  }
   constexpr std::ranges::iterator_t<View> operator->() const { return state_.front (); }
 
   constexpr iterator& operator++ () {
@@ -1085,6 +1087,7 @@ class transcode_view<FromEncoding, ToEncoding, View>::iterator {
 private:
   std::ranges::iterator_t<View> current_{};
   transcode_view const* parent_ = nullptr;
+  static value_type const replacement = icubaby::replacement_char;
 
   class state {
   public:
@@ -1117,6 +1120,11 @@ class transcode_view<FromEncoding, ToEncoding, View>::iterator {
   mutable state state_{};
 };
 
+template <unicode_char_type FromEncoding, unicode_char_type ToEncoding, std::ranges::input_range View>
+  requires std::ranges::view<View>
+transcode_view<FromEncoding, ToEncoding, View>::iterator::value_type const
+transcode_view<FromEncoding, ToEncoding, View>::iterator::replacement;
+
 template <unicode_char_type FromEncoding, unicode_char_type ToEncoding, std::ranges::input_range View>
   requires std::ranges::view<View>
 constexpr std::ranges::iterator_t<View> transcode_view<FromEncoding, ToEncoding, View>::iterator::state::fill (

diff --git a/unittests/test_u8_32.cpp b/unittests/test_u8_32.cpp
@@ -277,8 +277,23 @@ TEST (Utf8To32, RangesCopy) {
   };
   // clang-format on
   std::vector<char32_t> out32;
-  std::ranges::copy (in | icubaby::ranges::transcode<char8_t, char32_t>, std::back_inserter (out32));
+  auto const r = in | icubaby::ranges::transcode<char8_t, char32_t>;
+  std::ranges::copy (r, std::back_inserter (out32));
   EXPECT_THAT (out32, testing::ElementsAre (char32_t{0x3053}, char32_t{0x3093}, char32_t{0x306B}, char32_t{0x3061},
                                             char32_t{0x306F}, char32_t{0x4E16}, char32_t{0x754C}, char32_t{0x000A}));
+  EXPECT_TRUE (r.well_formed());
 }
+TEST (Utf8To32, RangesBadInput) {
+  // clang-format off
+  std::vector const in{
+    char8_t{0xF3}, char8_t{0x81},
+  };
+  // clang-format on
+  std::vector<char32_t> out32;
+  auto const r = in | icubaby::ranges::transcode<char8_t, char32_t>;
+  std::ranges::copy (r, std::back_inserter (out32));
+  EXPECT_THAT (out32, testing::ElementsAre (char32_t{icubaby::replacement_char}));
+  EXPECT_FALSE (r.well_formed());
+}
+
 #endif  // __cpp_lib_ranges