🐛 skipping BOM for iterators #602

I totally forgot about byte order marks in this scenario.
nlohmann · Jun 2, 2017 · d19c5ce · d19c5ce
1 parent c043ba6
commit d19c5ce
Show file tree

Hide file tree

Showing 3 changed files with 27 additions and 6 deletions.
diff --git a/src/json.hpp b/src/json.hpp
@@ -8814,7 +8814,7 @@ class basic_json
             // store number of bytes in the buffer
             fill_size = static_cast<size_t>(is.gcount());
 
-            // skip byte-order mark
+            // skip byte order mark
             if (fill_size >= 3 and buffer[0] == '\xEF' and buffer[1] == '\xBB' and buffer[2] == '\xBF')
             {
                 buffer_pos += 3;
@@ -8911,7 +8911,13 @@ class basic_json
       public:
         input_buffer_adapter(const char* b, size_t l)
             : input_adapter(), cursor(b), limit(b + l), start(b)
-        {}
+        {
+            // skip byte order mark
+            if (l >= 3 and b[0] == '\xEF' and b[1] == '\xBB' and b[2] == '\xBF')
+            {
+                cursor += 3;
+            }
+        }
 
         // delete because of pointer members
         input_buffer_adapter(const input_buffer_adapter&) = delete;

diff --git a/test/src/unit-regression.cpp b/test/src/unit-regression.cpp
@@ -1169,4 +1169,10 @@ TEST_CASE("regression tests")
         std::vector<uint8_t> vec = {'"', '\\', '"', 'X', '"', '"'};
         CHECK_THROWS_AS(json::parse(vec), json::parse_error);
     }
+
+    SECTION("issue #602 - BOM not skipped when using json:parse(iterator)")
+    {
+        std::string i = "\xef\xbb\xbf{\n   \"foo\": true\n}";
+        CHECK_NOTHROW(json::parse(i.begin(), i.end()));
+    }
 }
diff --git a/test/src/unit-unicode.cpp b/test/src/unit-unicode.cpp
@@ -1012,10 +1012,19 @@ TEST_CASE("Unicode", "[hide]")
 
     SECTION("ignore byte-order-mark")
     {
-        // read a file with a UTF-8 BOM
-        std::ifstream f("test/data/json_nlohmann_tests/bom.json");
-        json j;
-        CHECK_NOTHROW(f >> j);
+        SECTION("in a stream")
+        {
+            // read a file with a UTF-8 BOM
+            std::ifstream f("test/data/json_nlohmann_tests/bom.json");
+            json j;
+            CHECK_NOTHROW(f >> j);
+        }
+
+        SECTION("with an iterator")
+        {
+            std::string i = "\xef\xbb\xbf{\n   \"foo\": true\n}";
+            CHECK_NOTHROW(json::parse(i.begin(), i.end()));
+        }
     }
 
     SECTION("error for incomplete/wrong BOM")