src: use simdjson to parse SEA configuration

PR-URL: #59323
Refs: #59288
Reviewed-By: Darshan Sen <[email protected]>
Reviewed-By: Daniel Lemire <[email protected]>
Reviewed-By: Ethan Arrowood <[email protected]>

Author: joyeecheung

src/node_sea.cc

@@ -3,14 +3,14 @@
 #include "blob_serializer_deserializer-inl.h"
 #include "debug_utils-inl.h"
 #include "env-inl.h"
-#include "json_parser.h"
 #include "node_contextify.h"
 #include "node_errors.h"
 #include "node_external_reference.h"
 #include "node_internals.h"
 #include "node_snapshot_builder.h"
 #include "node_union_bytes.h"
 #include "node_v8_platform-inl.h"
+#include "simdjson.h"
 #include "util-inl.h"
 
 // The POSTJECT_SENTINEL_FUSE macro is a string of random characters selected by
@@ -303,79 +303,131 @@ std::optional<SeaConfig> ParseSingleExecutableConfig(
   }
 
   SeaConfig result;
-  JSONParser parser;
-  if (!parser.Parse(config)) {
-    FPrintF(stderr, "Cannot parse JSON from %s\n", config_path);
-    return std::nullopt;
-  }
 
-  result.main_path =
-      parser.GetTopLevelStringField("main").value_or(std::string());
-  if (result.main_path.empty()) {
-    FPrintF(stderr,
-            "\"main\" field of %s is not a non-empty string\n",
-            config_path);
-    return std::nullopt;
-  }
+  simdjson::ondemand::parser parser;
+  simdjson::ondemand::document document;
+  simdjson::ondemand::object main_object;
+  simdjson::error_code error =
+      parser.iterate(simdjson::pad(config)).get(document);
 
-  result.output_path =
-      parser.GetTopLevelStringField("output").value_or(std::string());
-  if (result.output_path.empty()) {
+  if (!error) {
+    error = document.get_object().get(main_object);
+  }
+  if (error) {
     FPrintF(stderr,
-            "\"output\" field of %s is not a non-empty string\n",
-            config_path);
+            "Cannot parse JSON from %s: %s\n",
+            config_path,
+            simdjson::error_message(error));
     return std::nullopt;
   }
 
-  std::optional<bool> disable_experimental_sea_warning =
-      parser.GetTopLevelBoolField("disableExperimentalSEAWarning");
-  if (!disable_experimental_sea_warning.has_value()) {
-    FPrintF(stderr,
+  bool use_snapshot_value = false;
+  bool use_code_cache_value = false;
+
+  for (auto field : main_object) {
+    std::string_view key;
+    if (field.unescaped_key().get(key)) {
+      FPrintF(stderr, "Cannot read key from %s\n", config_path);
+      return std::nullopt;
+    }
+    if (key == "main") {
+      if (field.value().get_string().get(result.main_path) ||
+          result.main_path.empty()) {
+        FPrintF(stderr,
+                "\"main\" field of %s is not a non-empty string\n",
+                config_path);
+        return std::nullopt;
+      }
+    } else if (key == "output") {
+      if (field.value().get_string().get(result.output_path) ||
+          result.output_path.empty()) {
+        FPrintF(stderr,
+                "\"output\" field of %s is not a non-empty string\n",
+                config_path);
+        return std::nullopt;
+      }
+    } else if (key == "disableExperimentalSEAWarning") {
+      bool disable_experimental_sea_warning;
+      if (field.value().get_bool().get(disable_experimental_sea_warning)) {
+        FPrintF(
+            stderr,
             "\"disableExperimentalSEAWarning\" field of %s is not a Boolean\n",
             config_path);
-    return std::nullopt;
-  }
-  if (disable_experimental_sea_warning.value()) {
-    result.flags |= SeaFlags::kDisableExperimentalSeaWarning;
+        return std::nullopt;
+      }
+      if (disable_experimental_sea_warning) {
+        result.flags |= SeaFlags::kDisableExperimentalSeaWarning;
+      }
+    } else if (key == "useSnapshot") {
+      if (field.value().get_bool().get(use_snapshot_value)) {
+        FPrintF(stderr,
+                "\"useSnapshot\" field of %s is not a Boolean\n",
+                config_path);
+        return std::nullopt;
+      }
+      if (use_snapshot_value) {
+        result.flags |= SeaFlags::kUseSnapshot;
+      }
+    } else if (key == "useCodeCache") {
+      if (field.value().get_bool().get(use_code_cache_value)) {
+        FPrintF(stderr,
+                "\"useCodeCache\" field of %s is not a Boolean\n",
+                config_path);
+        return std::nullopt;
+      }
+      if (use_code_cache_value) {
+        result.flags |= SeaFlags::kUseCodeCache;
+      }
+    } else if (key == "assets") {
+      simdjson::ondemand::object assets_object;
+      if (field.value().get_object().get(assets_object)) {
+        FPrintF(stderr,
+                "\"assets\" field of %s is not a map of strings\n",
+                config_path);
+        return std::nullopt;
+      }
+      simdjson::ondemand::value asset_value;
+      for (auto asset_field : assets_object) {
+        std::string_view key_str;
+        std::string_view value_str;
+        if (asset_field.unescaped_key().get(key_str) ||
+            asset_field.value().get(asset_value) ||
+            asset_value.get_string().get(value_str)) {
+          FPrintF(stderr,
+                  "\"assets\" field of %s is not a map of strings\n",
+                  config_path);
+          return std::nullopt;
+        }
+
+        result.assets.emplace(key_str, value_str);
+      }
+
+      if (!result.assets.empty()) {
+        result.flags |= SeaFlags::kIncludeAssets;
+      }
+    }
   }
 
-  std::optional<bool> use_snapshot = parser.GetTopLevelBoolField("useSnapshot");
-  if (!use_snapshot.has_value()) {
-    FPrintF(
-        stderr, "\"useSnapshot\" field of %s is not a Boolean\n", config_path);
-    return std::nullopt;
-  }
-  if (use_snapshot.value()) {
-    result.flags |= SeaFlags::kUseSnapshot;
+  if (static_cast<bool>(result.flags & SeaFlags::kUseSnapshot) &&
+      static_cast<bool>(result.flags & SeaFlags::kUseCodeCache)) {
+    // TODO(joyeecheung): code cache in snapshot should be configured by
+    // separate snapshot configurations.
+    FPrintF(stderr,
+            "\"useCodeCache\" is redundant when \"useSnapshot\" is true\n");
   }
 
-  std::optional<bool> use_code_cache =
-      parser.GetTopLevelBoolField("useCodeCache");
-  if (!use_code_cache.has_value()) {
-    FPrintF(
-        stderr, "\"useCodeCache\" field of %s is not a Boolean\n", config_path);
+  if (result.main_path.empty()) {
+    FPrintF(stderr,
+            "\"main\" field of %s is not a non-empty string\n",
+            config_path);
     return std::nullopt;
   }
-  if (use_code_cache.value()) {
-    if (use_snapshot.value()) {
-      // TODO(joyeecheung): code cache in snapshot should be configured by
-      // separate snapshot configurations.
-      FPrintF(stderr,
-              "\"useCodeCache\" is redundant when \"useSnapshot\" is true\n");
-    } else {
-      result.flags |= SeaFlags::kUseCodeCache;
-    }
-  }
 
-  auto assets_opt = parser.GetTopLevelStringDict("assets");
-  if (!assets_opt.has_value()) {
+  if (result.output_path.empty()) {
     FPrintF(stderr,
-            "\"assets\" field of %s is not a map of strings\n",
+            "\"output\" field of %s is not a non-empty string\n",
             config_path);
     return std::nullopt;
-  } else if (!assets_opt.value().empty()) {
-    result.flags |= SeaFlags::kIncludeAssets;
-    result.assets = std::move(assets_opt.value());
   }
 
   return result;