perf: parse chain-id from big genesis file could be slow

CHANGELOG.md

@@ -67,6 +67,7 @@ Ref: https://keepachangelog.com/en/1.0.0/
 * (rpc) [#17470](https://github.com/cosmos/cosmos-sdk/pull/17470) Avoid open 0.0.0.0 to public by default and add `listen-ip-address` argument for `testnet init-files` cmd.
 * (types) [#17670](https://github.com/cosmos/cosmos-sdk/pull/17670) Use `ctx.CometInfo` in place of `ctx.VoteInfos`
 * [#17733](https://github.com/cosmos/cosmos-sdk/pull/17733) Ensure `buf export` exports all proto dependencies
+* [#18204](https://github.com/cosmos/cosmos-sdk/pull/18204) Use streaming json parser to parse chain-id from genesis file.
 
 ### Bug Fixes
 

server/util.go

@@ -37,7 +37,6 @@ import (
 	sdk "github.com/cosmos/cosmos-sdk/types"
 	"github.com/cosmos/cosmos-sdk/types/mempool"
 	"github.com/cosmos/cosmos-sdk/version"
-	genutiltypes "github.com/cosmos/cosmos-sdk/x/genutil/types"
 )
 
 // ServerContextKey defines the context key used to retrieve a server.Context from
@@ -485,12 +484,16 @@ func DefaultBaseappOptions(appOpts types.AppOptions) []func(*baseapp.BaseApp) {
 	chainID := cast.ToString(appOpts.Get(flags.FlagChainID))
 	if chainID == "" {
 		// fallback to genesis chain-id
-		appGenesis, err := genutiltypes.AppGenesisFromFile(filepath.Join(homeDir, "config", "genesis.json"))
+		reader, err := os.Open(filepath.Join(homeDir, "config", "genesis.json"))
 		if err != nil {
 			panic(err)
 		}
+		defer reader.Close()
 
-		chainID = appGenesis.ChainID
+		chainID, err = sdk.ParseChainIDFromGenesis(reader)
+		if err != nil {
+			panic(fmt.Errorf("failed to parse chain-id from genesis file: %w", err))
+		}
 	}
 
 	snapshotStore, err := GetSnapshotStore(appOpts)

types/chain_id.go

@@ -0,0 +1,70 @@
+package types
+
+import (
+	"encoding/json"
+	"errors"
+	"fmt"
+	"io"
+
+	"github.com/cometbft/cometbft/types"
+)
+
+const ChainIDFieldName = "chain_id"
+
+// ParseChainIDFromGenesis parses the `chain_id` from the genesis json file and abort early,
+// it still parses the values before the `chain_id` field, particularly if the `app_state` field is
+// before the `chain_id` field, it will parse the `app_state` value, user must make sure the `chain_id`
+// is put before `app_state` or other big entries to enjoy the efficiency.
+func ParseChainIDFromGenesis(r io.Reader) (string, error) {
+	dec := json.NewDecoder(r)
+
+	var t json.Token
+	t, err := dec.Token()
+	if err != nil {
+		return "", err
+	}
+	if t != json.Delim('{') {
+		return "", fmt.Errorf("expected {, got %s", t)
+	}
+
+	for dec.More() {
+		t, err = dec.Token()
+		if err != nil {
+			return "", err
+		}
+		key, ok := t.(string)
+		if !ok {
+			return "", fmt.Errorf("expected string, got %s", t)
+		}
+
+		if key == ChainIDFieldName {
+			var chainId string
+			if err := dec.Decode(&chainId); err != nil {
+				return "", err
+			}
+			if err := validateChainID(chainId); err != nil {
+				return "", err
+			}
+			return chainId, nil
+		}
+
+		// skip the value
+		var value json.RawMessage
+		if err := dec.Decode(&value); err != nil {
+			return "", err
+		}
+	}
+
+	return "", errors.New("missing chain-id in genesis file")
+}
+
+func validateChainID(chainID string) error {
+	if chainID == "" {
+		return errors.New("genesis doc must include non-empty chain_id")
+	}
+	if len(chainID) > types.MaxChainIDLen {
+		return fmt.Errorf("chain_id in genesis doc is too long (max: %d)", types.MaxChainIDLen)
+	}
+
+	return nil
+}

types/chain_id_test.go

@@ -0,0 +1,113 @@
+package types
+
+import (
+	_ "embed"
+	"strings"
+	"testing"
+
+	genutiltypes "github.com/cosmos/cosmos-sdk/x/genutil/types"
+	"github.com/stretchr/testify/require"
+)
+
+//go:embed testdata/parse_chain_id.json
+var BenchmarkGenesis string
+
+func TestParseChainIDFromGenesis(t *testing.T) {
+	testCases := []struct {
+		name       string
+		json       string
+		expChainID string
+	}{
+		{
+			"success",
+			`{
+				"state": {
+				  "accounts": {
+					"a": {}
+				  }
+				},
+				"chain_id": "test-chain-id"
+			}`,
+			"test-chain-id",
+		},
+		{
+			"nested",
+			`{
+				"state": {
+				  "accounts": {
+					"a": {}
+				  },
+				  "chain_id": "test-chain-id"
+				}
+			}`,
+			"",
+		},
+		{
+			"not exist",
+			`{
+				"state": {
+				  "accounts": {
+					"a": {}
+				  }
+				},
+				"chain-id": "test-chain-id"
+			}`,
+			"",
+		},
+		{
+			"invalid type",
+			`{
+				"chain-id":1,
+			}`,
+			"",
+		},
+		{
+			"invalid json",
+			`[ " ': }`,
+			"",
+		},
+		{
+			"empty chain_id",
+			`{"chain_id": ""}`,
+			"",
+		},
+		{
+			"chain_id too long",
+			`{"chain_id": "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"}`,
+			"",
+		},
+	}
+
+	for _, tc := range testCases {
+		t.Run(tc.name, func(t *testing.T) {
+			chain_id, err := ParseChainIDFromGenesis(strings.NewReader(tc.json))
+			if tc.expChainID == "" {
+				require.Error(t, err)
+			} else {
+				require.NoError(t, err)
+				require.Equal(t, tc.expChainID, chain_id)
+			}
+		})
+	}
+}
+
+func BenchmarkParseChainID(b *testing.B) {
+	b.ReportAllocs()
+	b.Run("new", func(b *testing.B) {
+		b.ResetTimer()
+		for i := 0; i < b.N; i++ {
+			chainId, err := ParseChainIDFromGenesis(strings.NewReader(BenchmarkGenesis))
+			require.NoError(b, err)
+			require.Equal(b, "test_777-1", chainId)
+		}
+	})
+
+	b.Run("old", func(b *testing.B) {
+		b.ResetTimer()
+		for i := 0; i < b.N; i++ {
+			doc, err := genutiltypes.AppGenesisFromReader(strings.NewReader(BenchmarkGenesis))
+			require.NoError(b, err)
+			require.Equal(b, "test_777-1", doc.ChainID)
+		}
+	})
+}