Fill the indentation buffer once instead of once per flush

@gregmarr's point on the fill-and-flush loop: flushing does not disturb what
the write buffer holds, so an indentation spanning several buffer-fulls only
has to be written into the buffer once and can then be handed to the adapter
as many times as needed. The loop re-filled it every time, doing work it
already knew was there.

put_indent() now fills the room left in the buffer, and if anything remains,
flushes, fills the buffer once, and re-flushes that same content. It also
returns early for a zero-width indentation, which is what the closing brace of
every outermost value asks for.

Measured over a dump(), counting memset calls and bytes inside put_indent:

    indent       before              after
         4       1 call /     4 B    1 call /     4 B
      2000       2 calls /  2000 B   2 calls /  2046 B
    100000      98 calls / 100000 B  2 calls /  2046 B

The wide case is now constant work rather than proportional to the indentation
width; ordinary widths are unchanged. Tests extended to cover several whole
buffer-fulls and an exact multiple of the buffer size.

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

include/nlohmann/detail/output/serializer.hpp

@@ -718,24 +718,52 @@ class serializer
 
     Writes the indentation straight into the buffer instead of copying it out of
     a pre-grown indentation string, so no auxiliary string has to be sized,
-    resized, or kept in sync with the deepest nesting level reached. An
+    resized, or kept in sync with the deepest nesting level reached.
-    indentation wider than the buffer simply fills and flushes it repeatedly.
+
+    An indentation wider than the buffer is emitted by filling the buffer with
+    the indentation character once and flushing that same content repeatedly:
+    flushing does not disturb what the buffer holds, so re-filling it between
+    flushes would be redundant work.
     */
     void put_indent(unsigned int indent)
     {
-        while (indent > 0)
+        // closing braces at the outermost level ask for no indentation at all
+        if (indent == 0)
         {
-            if (JSON_HEDLEY_UNLIKELY(write_buffer_pos == write_buffer.size()))
+            return;
-            {
-                flush();
-            }
-
-            const std::size_t chunk = (std::min)(static_cast<std::size_t>(indent),
-                                                 write_buffer.size() - write_buffer_pos);
-            std::memset(write_buffer.data() + write_buffer_pos, indent_char, chunk);
-            write_buffer_pos += chunk;
-            indent -= static_cast<unsigned int>(chunk);
         }
+
+        const std::size_t capacity = write_buffer.size();
+
+        // fill whatever room is left in the buffer; this is the whole job
+        // whenever the indentation is narrower than the buffer, which is the
+        // case for every sane indent_step
+        const std::size_t head = (std::min)(static_cast<std::size_t>(indent), capacity - write_buffer_pos);
+        std::memset(write_buffer.data() + write_buffer_pos, indent_char, head);
+        write_buffer_pos += head;
+        indent -= static_cast<unsigned int>(head);
+
+        if (JSON_HEDLEY_LIKELY(indent == 0))
+        {
+            return;
+        }
+
+        // the buffer is full and the remainder spans whole buffer-fulls: flush
+        // what is pending, then fill the buffer with the indentation character
+        // exactly once and hand the same bytes to the adapter as often as needed
+        flush();
+        std::memset(write_buffer.data(), indent_char, capacity);
+
+        while (indent >= capacity)
+        {
+            write_buffer_pos = capacity;
+            flush();
+            indent -= static_cast<unsigned int>(capacity);
+        }
+
+        // the buffer still holds indentation characters throughout, so the tail
+        // only has to be claimed, not written again
+        write_buffer_pos = indent;
     }
 
     /*!

single_include/nlohmann/json.hpp

@@ -21651,24 +21651,52 @@ class serializer
 
     Writes the indentation straight into the buffer instead of copying it out of
     a pre-grown indentation string, so no auxiliary string has to be sized,
-    resized, or kept in sync with the deepest nesting level reached. An
+    resized, or kept in sync with the deepest nesting level reached.
-    indentation wider than the buffer simply fills and flushes it repeatedly.
+
+    An indentation wider than the buffer is emitted by filling the buffer with
+    the indentation character once and flushing that same content repeatedly:
+    flushing does not disturb what the buffer holds, so re-filling it between
+    flushes would be redundant work.
     */
     void put_indent(unsigned int indent)
     {
-        while (indent > 0)
+        // closing braces at the outermost level ask for no indentation at all
+        if (indent == 0)
         {
-            if (JSON_HEDLEY_UNLIKELY(write_buffer_pos == write_buffer.size()))
+            return;
-            {
-                flush();
-            }
-
-            const std::size_t chunk = (std::min)(static_cast<std::size_t>(indent),
-                                                 write_buffer.size() - write_buffer_pos);
-            std::memset(write_buffer.data() + write_buffer_pos, indent_char, chunk);
-            write_buffer_pos += chunk;
-            indent -= static_cast<unsigned int>(chunk);
         }
+
+        const std::size_t capacity = write_buffer.size();
+
+        // fill whatever room is left in the buffer; this is the whole job
+        // whenever the indentation is narrower than the buffer, which is the
+        // case for every sane indent_step
+        const std::size_t head = (std::min)(static_cast<std::size_t>(indent), capacity - write_buffer_pos);
+        std::memset(write_buffer.data() + write_buffer_pos, indent_char, head);
+        write_buffer_pos += head;
+        indent -= static_cast<unsigned int>(head);
+
+        if (JSON_HEDLEY_LIKELY(indent == 0))
+        {
+            return;
+        }
+
+        // the buffer is full and the remainder spans whole buffer-fulls: flush
+        // what is pending, then fill the buffer with the indentation character
+        // exactly once and hand the same bytes to the adapter as often as needed
+        flush();
+        std::memset(write_buffer.data(), indent_char, capacity);
+
+        while (indent >= capacity)
+        {
+            write_buffer_pos = capacity;
+            flush();
+            indent -= static_cast<unsigned int>(capacity);
+        }
+
+        // the buffer still holds indentation characters throughout, so the tail
+        // only has to be claimed, not written again
+        write_buffer_pos = indent;
     }
 
     /*!

tests/src/unit-serialization.cpp

@@ -485,6 +485,12 @@ TEST_CASE("indentation is written straight into the write buffer")
         // 2000 > the 1024-byte write buffer, and > the 512 the indentation
         // string used to start at
         CHECK(j.dump(2000) == "{\n" + std::string(2000, ' ') + "\"a\": 1\n}");
+        // several whole buffer-fulls, so the buffer is refilled once and then
+        // flushed repeatedly
+        CHECK(j.dump(5000) == "{\n" + std::string(5000, ' ') + "\"a\": 1\n}");
+        CHECK(j.dump(5000, '\t') == "{\n" + std::string(5000, '\t') + "\"a\": 1\n}");
+        // an exact multiple of the buffer size
+        CHECK(j.dump(4096) == "{\n" + std::string(4096, ' ') + "\"a\": 1\n}");
     }
 
     SECTION("a non-space indentation character is used throughout")