Remove set_capacity, replace with allocate

jkeiser · jkeiser · commit d93af1161ddd · 2020-03-30T13:49:54.000-07:00
Makes allocation point more predictable
diff --git a/benchmark/bench_dom_api.cpp b/benchmark/bench_dom_api.cpp
@@ -256,7 +256,6 @@ static void print_json(State& state) noexcept {
   // Prints the number of results in twitter.json
   padded_string json = get_corpus(JSON_TEST_PATH);
   dom::parser parser;
-  if (!parser.allocate_capacity(json.length())) { cerr << "allocation failed" << endl; return; }
   if (int error = json_parse(json, parser); error != SUCCESS) { cerr << error_message(error) << endl; return; }
   for (auto _ : state) {
     std::stringstream s;
diff --git a/benchmark/bench_parse_call.cpp b/benchmark/bench_parse_call.cpp
@@ -9,18 +9,18 @@ const padded_string EMPTY_ARRAY("[]", 2);
 SIMDJSON_PUSH_DISABLE_WARNINGS
 SIMDJSON_DISABLE_DEPRECATED_WARNING
 static void json_parse(State& state) {
-  dom::parser parser;
-  if (parser.set_capacity(EMPTY_ARRAY.length())) { return; }
+  ParsedJson pj;
+  if (!pj.allocate_capacity(EMPTY_ARRAY.length())) { return; }
   for (auto _ : state) {
-    auto error = simdjson::json_parse(EMPTY_ARRAY, parser);
+    auto error = json_parse(EMPTY_ARRAY, pj);
     if (error) { return; }
   }
 }
 SIMDJSON_POP_DISABLE_WARNINGS
 BENCHMARK(json_parse);
 static void parser_parse_error_code(State& state) {
   dom::parser parser;
-  if (parser.set_capacity(EMPTY_ARRAY.length())) { return; }
+  if (parser.allocate(EMPTY_ARRAY.length())) { return; }
   for (auto _ : state) {
     auto [doc, error] = parser.parse(EMPTY_ARRAY);
     if (error) { return; }
@@ -29,7 +29,7 @@ static void parser_parse_error_code(State& state) {
 BENCHMARK(parser_parse_error_code);
 static void parser_parse_exception(State& state) {
   dom::parser parser;
-  if (parser.set_capacity(EMPTY_ARRAY.length())) { return; }
+  if (parser.allocate(EMPTY_ARRAY.length())) { return; }
   for (auto _ : state) {
     try {
       UNUSED dom::element doc = parser.parse(EMPTY_ARRAY);
diff --git a/benchmark/benchmarker.h b/benchmark/benchmarker.h
@@ -294,7 +294,10 @@ struct benchmarker {
     // Allocate dom::parser
     collector.start();
     dom::parser parser;
-    error_code alloc_error = parser.set_capacity(json.size());
+    error_code error = parser.allocate(json.size());
+    if (error) {
+      exit_error(string("Unable to allocate_stage ") + to_string(json.size()) + " bytes for the JSON result: " + error_message(error));
+    }
     event_count allocate_count = collector.end();
     allocate_stage << allocate_count;
     // Run it once to get hot buffers
@@ -305,14 +308,11 @@ struct benchmarker {
       }
     }
 
-    if (alloc_error) {
-      exit_error(string("Unable to allocate_stage ") + to_string(json.size()) + " bytes for the JSON result: " + error_message(alloc_error));
-    }
     verbose() << "[verbose] allocated memory for parsed JSON " << endl;
 
     // Stage 1 (find structurals)
     collector.start();
-    error_code error = active_implementation->stage1((const uint8_t *)json.data(), json.size(), parser, false);
+    error = active_implementation->stage1((const uint8_t *)json.data(), json.size(), parser, false);
     event_count stage1_count = collector.end();
     stage1 << stage1_count;
     if (error) {
diff --git a/benchmark/linux/linux-perf-events.h b/benchmark/linux/linux-perf-events.h
@@ -56,25 +56,29 @@ template <int TYPE = PERF_TYPE_HARDWARE> class LinuxEvents {
     temp_result_vec.resize(num_events * 2 + 1);
   }
 
-  ~LinuxEvents() { close(fd); }
+  ~LinuxEvents() { if (fd != -1) { close(fd); } }
 
   inline void start() {
-    if (ioctl(fd, PERF_EVENT_IOC_RESET, PERF_IOC_FLAG_GROUP) == -1) {
-      report_error("ioctl(PERF_EVENT_IOC_RESET)");
-    }
+    if (fd != -1) {
+      if (ioctl(fd, PERF_EVENT_IOC_RESET, PERF_IOC_FLAG_GROUP) == -1) {
+        report_error("ioctl(PERF_EVENT_IOC_RESET)");
+      }
 
-    if (ioctl(fd, PERF_EVENT_IOC_ENABLE, PERF_IOC_FLAG_GROUP) == -1) {
-      report_error("ioctl(PERF_EVENT_IOC_ENABLE)");
+      if (ioctl(fd, PERF_EVENT_IOC_ENABLE, PERF_IOC_FLAG_GROUP) == -1) {
+        report_error("ioctl(PERF_EVENT_IOC_ENABLE)");
+      }
     }
   }
 
   inline void end(std::vector<unsigned long long> &results) {
-    if (ioctl(fd, PERF_EVENT_IOC_DISABLE, PERF_IOC_FLAG_GROUP) == -1) {
-      report_error("ioctl(PERF_EVENT_IOC_DISABLE)");
-    }
+    if (fd != -1) {
+      if (ioctl(fd, PERF_EVENT_IOC_DISABLE, PERF_IOC_FLAG_GROUP) == -1) {
+        report_error("ioctl(PERF_EVENT_IOC_DISABLE)");
+      }
 
-    if (read(fd, temp_result_vec.data(), temp_result_vec.size() * 8) == -1) {
-      report_error("read");
+      if (read(fd, temp_result_vec.data(), temp_result_vec.size() * 8) == -1) {
+        report_error("read");
+      }
     }
     // our actual results are in slots 1,3,5, ... of this structure
     // we really should be checking our ids obtained earlier to be safe
diff --git a/benchmark/parse_stream.cpp b/benchmark/parse_stream.cpp
@@ -36,7 +36,7 @@ int main (int argc, char *argv[]){
         for (auto i = 0; i < 3; i++) {
             //Actual test
             simdjson::dom::parser parser;
-            simdjson::error_code alloc_error = parser.set_capacity(p.size());
+            simdjson::error_code alloc_error = parser.allocate(p.size());
             if (alloc_error) {
                 std::cerr << alloc_error << std::endl;
                 return EXIT_FAILURE;
diff --git a/benchmark/statisticalmodel.cpp b/benchmark/statisticalmodel.cpp
@@ -161,7 +161,7 @@ int main(int argc, char *argv[]) {
 #ifdef __linux__
   simdjson::dom::parser parser;
   const simdjson::implementation &stage_parser = *simdjson::active_implementation;
-  simdjson::error_code alloc_error = parser.set_capacity(p.size());
+  simdjson::error_code alloc_error = parser.allocate(p.size());
   if (alloc_error) {
     std::cerr << alloc_error << std::endl;
     return EXIT_FAILURE;
diff --git a/doc/performance.md b/doc/performance.md
@@ -77,7 +77,7 @@ without bound:
 
   ```c++
   dom::parser parser(0); // This parser will refuse to automatically grow capacity
-  simdjson::error_code allocate_error = parser.set_capacity(1024*1024); // This allocates enough capacity to handle documents <= 1MB
+  simdjson::error_code allocate_error = parser.allocate(1024*1024); // This allocates enough capacity to handle documents <= 1MB
   if (allocate_error) { cerr << allocate_error << endl; exit(1); }
 
   for (web_request request : listen()) {
diff --git a/include/simdjson/document.h b/include/simdjson/document.h
@@ -350,7 +350,7 @@ class document {
   std::unique_ptr<uint8_t[]> string_buf;
 
 private:
-  inline error_code set_capacity(size_t len) noexcept;
+  inline error_code allocate(size_t len) noexcept;
   template<typename T>
   friend class simdjson::minify;
   friend class parser;
@@ -609,13 +609,10 @@ class parser {
   * @param max_capacity The maximum document length the parser can automatically handle. The parser
   *    will allocate more capacity on an as needed basis (when it sees documents too big to handle)
   *    up to this amount. The parser still starts with zero capacity no matter what this number is:
-  *    to allocate an initial capacity, call set_capacity() after constructing the parser. Defaults
-  *    to SIMDJSON_MAXSIZE_BYTES (the largest single document simdjson can process).
-  * @param max_depth The maximum depth--number of nested objects and arrays--this parser can handle.
-  *    This will not be allocated until parse() is called for the first time. Defaults to
-  *    DEFAULT_MAX_DEPTH.
+  *    to allocate an initial capacity, call allocate() after constructing the parser.
+  *    Defaults to SIMDJSON_MAXSIZE_BYTES (the largest single document simdjson can process).
   */
-  really_inline parser(size_t max_capacity = SIMDJSON_MAXSIZE_BYTES, size_t max_depth = DEFAULT_MAX_DEPTH) noexcept;
+  really_inline parser(size_t max_capacity = SIMDJSON_MAXSIZE_BYTES) noexcept;
 
   /**
    * Take another parser's buffers and state.
@@ -835,6 +832,30 @@ class parser {
   /** @private We do not want to allow implicit conversion from C string to std::string. */
   really_inline simdjson_result<element> parse_many(const char *buf, size_t batch_size = DEFAULT_BATCH_SIZE) noexcept = delete;
 
+  /**
+   * Ensure this parser has enough memory to process JSON documents up to `capacity` bytes in length
+   * and `max_depth` depth.
+   *
+   * @param capacity The new capacity.
+   * @param max_depth The new max_depth. Defaults to DEFAULT_MAX_DEPTH.
+   * @return The error, if there is one.
+   */
+  WARN_UNUSED inline error_code allocate(size_t capacity, size_t max_depth = DEFAULT_MAX_DEPTH) noexcept;
+
+  /**
+   * @private deprecated because it returns bool instead of error_code, which is our standard for
+   * failures. Use allocate() instead.
+   *
+   * Ensure this parser has enough memory to process JSON documents up to `capacity` bytes in length
+   * and `max_depth` depth.
+   *
+   * @param capacity The new capacity.
+   * @param max_depth The new max_depth. Defaults to DEFAULT_MAX_DEPTH.
+   * @return true if successful, false if allocation failed.
+   */
+  [[deprecated("Use allocate() instead.")]]
+  WARN_UNUSED inline bool allocate_capacity(size_t capacity, size_t max_depth = DEFAULT_MAX_DEPTH) noexcept;
+
   /**
    * The largest document this parser can support without reallocating.
    *
@@ -858,17 +879,6 @@ class parser {
    */
   really_inline size_t max_depth() const noexcept;
 
-  /**
-   * Set capacity. This is the largest document this parser can support without reallocating.
-   *
-   * This will allocate or deallocate as necessary.
-   *
-   * @param capacity The new capacity, in bytes.
-   *
-   * @return MEMALLOC if unsuccessful, SUCCESS otherwise.
-   */
-  WARN_UNUSED inline error_code set_capacity(size_t capacity) noexcept;
-
   /**
    * Set max_capacity. This is the largest document this parser can automatically support.
    *
@@ -880,32 +890,6 @@ class parser {
    */
   really_inline void set_max_capacity(size_t max_capacity) noexcept;
 
-  /**
-   * Set the maximum level of nested object and arrays supported by this parser.
-   *
-   * This will allocate or deallocate as necessary.
-   *
-   * @param max_depth The new maximum depth, in bytes.
-   *
-   * @return MEMALLOC if unsuccessful, SUCCESS otherwise.
-   */
-  WARN_UNUSED inline error_code set_max_depth(size_t max_depth) noexcept;
-
-  /**
-   * @private @deprecated Use set_capacity() instead.
-   *
-   * Ensure this parser has enough memory to process JSON documents up to `capacity` bytes in length
-   * and `max_depth` depth.
-   *
-   * Equivalent to calling set_capacity() and set_max_depth().
-   *
-   * @param capacity The new capacity.
-   * @param max_depth The new max_depth. Defaults to DEFAULT_MAX_DEPTH.
-   * @return true if successful, false if allocation failed.
-   */
-  [[deprecated("Use set_capacity() instead.")]]
-  WARN_UNUSED inline bool allocate_capacity(size_t capacity, size_t max_depth = DEFAULT_MAX_DEPTH) noexcept;
-
   /** @private Use the new DOM API instead */
   class Iterator;
   /** @private Use simdjson_error instead */
@@ -988,25 +972,25 @@ class parser {
 
 private:
   /**
-   * The maximum document length this parser supports.
+   * The maximum document length this parser will automatically support.
    *
-   * Buffers are large enough to handle any document up to this length.
+   * The parser will not be automatically allocated above this amount.
    */
-  size_t _capacity{0};
+  size_t _max_capacity;
 
   /**
-   * The maximum document length this parser will automatically support.
+   * The maximum document length this parser supports.
    *
-   * The parser will not be automatically allocated above this amount.
+   * Buffers are large enough to handle any document up to this length.
    */
-  size_t _max_capacity;
+  size_t _capacity{0};
 
   /**
    * The maximum depth (number of nested objects and arrays) supported by this parser.
    *
    * Defaults to DEFAULT_MAX_DEPTH.
    */
-  size_t _max_depth;
+  size_t _max_depth{0};
 
   /**
    * The loaded buffer (reused each time load() is called)
diff --git a/include/simdjson/inline/document.h b/include/simdjson/inline/document.h
diff --git a/tests/readme_examples.cpp b/tests/readme_examples.cpp
