perf: parse chain-id from big genesis file could be slow (backport #18204) (#18267)

Co-authored-by: yihuang <huang@crypto.com>
Co-authored-by: marbar3778 <marbar3778@yahoo.com>

CHANGELOG.md

@@ -50,6 +50,7 @@ Ref: https://keepachangelog.com/en/1.0.0/
 * (x/staking) [#18035](https://github.com/cosmos/cosmos-sdk/pull/18035) Hoisted out of the redelegation loop, the non-changing validator and delegator addresses parsing.
 * (keyring) [#17913](https://github.com/cosmos/cosmos-sdk/pull/17913) Add `NewAutoCLIKeyring` for creating an AutoCLI keyring from a SDK keyring.
 * (x/consensus) [#18041](https://github.com/cosmos/cosmos-sdk/pull/18041) Let `ToProtoConsensusParams()` return an error.
+* [#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

@@ -475,12 +475,16 @@ 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 = genutiltypes.ParseChainIDFromGenesis(reader)
+		if err != nil {
+			panic(fmt.Errorf("failed to parse chain-id from genesis file: %w", err))
+		}
 	}
 
 	snapshotStore, err := GetSnapshotStore(appOpts)

x/genutil/types/chain_id.go

@@ -0,0 +1,69 @@
+package types
+
+import (
+	"encoding/json"
+	"errors"
+	"fmt"
+	"io"
+	"strings"
+
+	"github.com/cometbft/cometbft/types"
+)
+
+const ChainIDFieldName = "chain_id"
+
+// ParseChainIDFromGenesis parses the `chain_id` from a genesis JSON file, aborting early after finding the `chain_id`.
+// For efficiency, it's recommended to place the `chain_id` field before any large entries in the JSON file.
+// Returns an error if the `chain_id` field is not found.
+func ParseChainIDFromGenesis(r io.Reader) (string, error) {
+	dec := json.NewDecoder(r)
+
+	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 for the key type, 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 strings.TrimSpace(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
+}

x/genutil/types/chain_id_test.go

@@ -0,0 +1,130 @@
+package types_test
+
+import (
+	_ "embed"
+	"strings"
+	"testing"
+
+	"github.com/stretchr/testify/require"
+
+	"github.com/cosmos/cosmos-sdk/x/genutil/types"
+)
+
+//go:embed testdata/parse_chain_id.json
+var BenchmarkGenesis string
+
+func TestParseChainIDFromGenesis(t *testing.T) {
+	testCases := []struct {
+		name       string
+		json       string
+		expChainID string
+		expError   string
+	}{
+		{
+			"success",
+			`{
+				"state": {
+				  "accounts": {
+					"a": {}
+				  }
+				},
+				"chain_id": "test-chain-id"
+			}`,
+			"test-chain-id",
+			"",
+		},
+		{
+			"nested",
+			`{
+				"state": {
+				  "accounts": {
+					"a": {}
+				  },
+				  "chain_id": "test-chain-id"
+				}
+			}`,
+			"",
+			"missing chain-id in genesis file",
+		},
+		{
+			"not exist",
+			`{
+				"state": {
+				  "accounts": {
+					"a": {}
+				  }
+				},
+				"chain-id": "test-chain-id"
+			}`,
+			"",
+			"missing chain-id in genesis file",
+		},
+		{
+			"invalid type",
+			`{
+				"chain-id": 1,
+			}`,
+			"",
+			"invalid character '}' looking for beginning of object key string",
+		},
+		{
+			"invalid json",
+			`[ " ': }`,
+			"",
+			"expected {, got [",
+		},
+		{
+			"empty chain_id",
+			`{"chain_id": ""}`,
+			"",
+			"genesis doc must include non-empty chain_id",
+		},
+		{
+			"whitespace chain_id",
+			`{"chain_id": "   "}`,
+			"",
+			"genesis doc must include non-empty chain_id",
+		},
+		{
+			"chain_id too long",
+			`{"chain_id": "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"}`,
+			"",
+			"chain_id in genesis doc is too long",
+		},
+	}
+
+	for _, tc := range testCases {
+		t.Run(tc.name, func(t *testing.T) {
+			chainID, err := types.ParseChainIDFromGenesis(strings.NewReader(tc.json))
+			if tc.expChainID == "" {
+				require.Error(t, err)
+				require.Contains(t, err.Error(), tc.expError)
+			} else {
+				require.NoError(t, err)
+				require.Equal(t, tc.expChainID, chainID)
+			}
+		})
+	}
+}
+
+func BenchmarkParseChainID(b *testing.B) {
+	expChainID := "cronosmainnet_25-1"
+	b.ReportAllocs()
+	b.Run("new", func(b *testing.B) {
+		b.ResetTimer()
+		for i := 0; i < b.N; i++ {
+			chainID, err := types.ParseChainIDFromGenesis(strings.NewReader(BenchmarkGenesis))
+			require.NoError(b, err)
+			require.Equal(b, expChainID, chainID)
+		}
+	})
+
+	b.Run("old", func(b *testing.B) {
+		b.ResetTimer()
+		for i := 0; i < b.N; i++ {
+			doc, err := types.AppGenesisFromReader(strings.NewReader(BenchmarkGenesis))
+			require.NoError(b, err)
+			require.Equal(b, expChainID, doc.ChainID)
+		}
+	})
+}