Benchmark parsing of pretty-printed JSON

Every input in the benchmark corpus is minified or only lightly spaced, so
none of them exercise the lexer's whitespace handling. Real-world JSON is
frequently indented - configuration files, pretty-printed API responses,
anything kept under version control - where the insignificant whitespace can
outweigh the data itself.

Add a ParseIndented family that re-serializes each document with an
indentation and parses that. The content is identical to the matching
ParseString row, so the pair isolates the cost of the whitespace alone.
Measured here, parsing the indented form costs 16-34% more than the minified
form of the same document, which nothing in the suite currently reports.

Signed-off-by: Niels Lohmann <mail@nlohmann.me>

include/nlohmann/detail/output/serializer.hpp

@@ -63,13 +63,14 @@ class serializer
 
   public:
     /*!
-    @param[in] s  output stream to serialize to
+    @param[in] s  output adapter to serialize to; not owned by the serializer,
+                  so it must outlive it (it lives at the call site)
     @param[in] ichar  indentation character to use
     @param[in] error_handler_  how to react on decoding errors
     */
-    serializer(output_adapter_t<char> s, const char ichar,
+    serializer(output_adapter_protocol<char>& s, const char ichar,
                error_handler_t error_handler_ = error_handler_t::strict)
-        : o(std::move(s))
+        : o(&s)
         , loc(std::localeconv())
         , thousands_sep(loc->thousands_sep == nullptr ? '\0' : std::char_traits<char>::to_char_type(* (loc->thousands_sep)))
         , decimal_point(loc->decimal_point == nullptr ? '\0' : std::char_traits<char>::to_char_type(* (loc->decimal_point)))
@@ -1688,8 +1689,8 @@ class serializer
     }
 
   private:
-    /// the output of the serializer
-    output_adapter_t<char> o = nullptr;
+    /// the output of the serializer (non-owning; the adapter lives at the call site)
+    output_adapter_protocol<char>* o = nullptr;
 
     /// a (hopefully) large enough character buffer
     std::array<char, 64> number_buffer{{}};

include/nlohmann/json.hpp

@@ -1341,7 +1341,8 @@ class basic_json // NOLINT(cppcoreguidelines-special-member-functions,hicpp-spec
                   const error_handler_t error_handler = error_handler_t::strict) const
     {
         string_t result;
-        serializer s(detail::output_adapter<char, string_t>(result), indent_char, error_handler);
+        detail::output_string_adapter<char, string_t> string_adapter(result);
+        serializer s(string_adapter, indent_char, error_handler);
 
         if (indent >= 0)
         {
@@ -4055,7 +4056,8 @@ class basic_json // NOLINT(cppcoreguidelines-special-member-functions,hicpp-spec
         o.width(0);
 
         // do the actual serialization
-        serializer s(detail::output_adapter<char>(o), o.fill());
+        detail::output_stream_adapter<char> stream_adapter(o);
+        serializer s(stream_adapter, o.fill());
         s.dump(j, pretty_print, false, static_cast<unsigned int>(indentation));
         return o;
     }

single_include/nlohmann/json.hpp

@@ -21123,13 +21123,14 @@ class serializer
 
   public:
     /*!
-    @param[in] s  output stream to serialize to
+    @param[in] s  output adapter to serialize to; not owned by the serializer,
+                  so it must outlive it (it lives at the call site)
     @param[in] ichar  indentation character to use
     @param[in] error_handler_  how to react on decoding errors
     */
-    serializer(output_adapter_t<char> s, const char ichar,
+    serializer(output_adapter_protocol<char>& s, const char ichar,
                error_handler_t error_handler_ = error_handler_t::strict)
-        : o(std::move(s))
+        : o(&s)
         , loc(std::localeconv())
         , thousands_sep(loc->thousands_sep == nullptr ? '\0' : std::char_traits<char>::to_char_type(* (loc->thousands_sep)))
         , decimal_point(loc->decimal_point == nullptr ? '\0' : std::char_traits<char>::to_char_type(* (loc->decimal_point)))
@@ -22748,8 +22749,8 @@ class serializer
     }
 
   private:
-    /// the output of the serializer
-    output_adapter_t<char> o = nullptr;
+    /// the output of the serializer (non-owning; the adapter lives at the call site)
+    output_adapter_protocol<char>* o = nullptr;
 
     /// a (hopefully) large enough character buffer
     std::array<char, 64> number_buffer{{}};
@@ -24451,7 +24452,8 @@ class basic_json // NOLINT(cppcoreguidelines-special-member-functions,hicpp-spec
                   const error_handler_t error_handler = error_handler_t::strict) const
     {
         string_t result;
-        serializer s(detail::output_adapter<char, string_t>(result), indent_char, error_handler);
+        detail::output_string_adapter<char, string_t> string_adapter(result);
+        serializer s(string_adapter, indent_char, error_handler);
 
         if (indent >= 0)
         {
@@ -27165,7 +27167,8 @@ class basic_json // NOLINT(cppcoreguidelines-special-member-functions,hicpp-spec
         o.width(0);
 
         // do the actual serialization
-        serializer s(detail::output_adapter<char>(o), o.fill());
+        detail::output_stream_adapter<char> stream_adapter(o);
+        serializer s(stream_adapter, o.fill());
         s.dump(j, pretty_print, false, static_cast<unsigned int>(indentation));
         return o;
     }

tests/benchmarks/src/benchmarks.cpp

@@ -81,6 +81,44 @@ BENCHMARK_CAPTURE(ParseString, signed_ints,       TEST_DATA_DIRECTORY "/regressi
 BENCHMARK_CAPTURE(ParseString, unsigned_ints,     TEST_DATA_DIRECTORY "/regression/unsigned_ints.json");
 BENCHMARK_CAPTURE(ParseString, small_signed_ints, TEST_DATA_DIRECTORY "/regression/small_signed_ints.json");
 
+//////////////////////////////////////////////////////////////////////////////
+// parse pretty-printed JSON from string
+//
+// Every file in the corpus above is minified or only lightly spaced, so none of
+// them exercise the lexer's whitespace handling. Real-world JSON is frequently
+// indented - configuration files, pretty-printed API responses, anything kept
+// under version control - where insignificant whitespace can outweigh the data.
+// Re-serializing a document with an indentation and parsing that keeps the
+// content identical to the ParseString row above, so the pair isolates the cost
+// of the whitespace alone.
+//////////////////////////////////////////////////////////////////////////////
+
+static void ParseIndented(benchmark::State& state, const char* filename, int indent)
+{
+    std::ifstream f(filename);
+    std::string str((std::istreambuf_iterator<char>(f)), std::istreambuf_iterator<char>());
+    const std::string indented = json::parse(str).dump(indent);
+
+    while (state.KeepRunning())
+    {
+        state.PauseTiming();
+        auto* j = new json();
+        state.ResumeTiming();
+
+        *j = json::parse(indented);
+
+        state.PauseTiming();
+        delete j;
+        state.ResumeTiming();
+    }
+
+    state.SetBytesProcessed(state.iterations() * indented.size());
+}
+BENCHMARK_CAPTURE(ParseIndented, jeopardy / 4,     TEST_DATA_DIRECTORY "/jeopardy/jeopardy.json",                 4);
+BENCHMARK_CAPTURE(ParseIndented, canada / 4,       TEST_DATA_DIRECTORY "/nativejson-benchmark/canada.json",       4);
+BENCHMARK_CAPTURE(ParseIndented, citm_catalog / 4, TEST_DATA_DIRECTORY "/nativejson-benchmark/citm_catalog.json", 4);
+BENCHMARK_CAPTURE(ParseIndented, twitter / 4,      TEST_DATA_DIRECTORY "/nativejson-benchmark/twitter.json",      4);
+
 //////////////////////////////////////////////////////////////////////////////
 // serialize JSON
 //////////////////////////////////////////////////////////////////////////////

tests/src/unit-convenience.cpp

@@ -98,7 +98,8 @@ void check_escaped(const char* original, const char* escaped = "", bool ensure_a
 void check_escaped(const char* original, const char* escaped, const bool ensure_ascii)
 {
     std::stringstream ss;
-    json::serializer s(nlohmann::detail::output_adapter<char>(ss), ' ');
+    nlohmann::detail::output_stream_adapter<char> adapter(ss);
+    json::serializer s(adapter, ' ');
     s.dump_escaped(original, ensure_ascii);
     s.flush(); // dump_escaped writes into the serializer's internal buffer
     CHECK(ss.str() == escaped);