feat: add db writer, types, logging, util and test infrastructure by farooqazamwasimnl · Pull Request #4 · LF-Decentralized-Trust-labs/fabric-x-block-explorer

farooqazamwasimnl · 2026-02-24T14:07:10Z

Introduces the database writer layer to persist parsed Hyperledger Fabric blocks into PostgreSQL
Adds domain types covering ParsedBlockData, ProcessedBlock, TxRecord, TxNamespaceRecord, ReadOnlyRecord, ReadWriteRecord, BlindWriteRecord, EndorsementRecord, and NamespacePolicyRecord; types are organized hierarchically to eliminate duplicate fields
Adds pgtype nullable helpers for converting between Go pointers and PostgreSQL nullable types
Adds a pgx connection pool factory
Adds BlockWriter that atomically persists a complete block including transactions, namespace entries, reads-only, read-writes, blind-writes, endorsements, and policies within a single DB transaction with rollback on error; all inserts use pgx SendBatch for a single round-trip per table
Adds testcontainers-go based test infrastructure supporting both containerised and local PostgreSQL
Full test coverage for BlockWriter including rollback on error, blind writes, multiple transactions, nil handling, and policy upserts
Adds golangci-lint configuration and lint Makefile target
Updates Makefile with test-no-db, test-requires-db, coverage, and clean targets
Stacks on top of the db schema PR

Signed-off-by: Farooq Azam Wasim <Farooq.Azm.Wasim-CIC.Netherlands@ibm.com>

cendhu

I have done a first pass. Need to do a more in-depth review.

cendhu · 2026-02-25T10:21:34Z

pkg/logging/logging.go

@@ -0,0 +1,165 @@
+/*


directly use flogging from fabric. https://pkg.go.dev/github.com/hyperledger/fabric/common/flogging

this local logging package can be removed.

@cendhu Done

cendhu · 2026-02-25T10:23:47Z

pkg/types/types.go

+	TxID           string
+	ValidationCode int32
+	IsBlindWrite   bool
+	ReadVersion    *uint64


I have a change of mind on this. I feel it would be more straight-forward and easy to understand read-only entries, read-write entries, blind-write entries rather than this optimization? what do you think? It will reflect the block structure as-is.

@cendhu I agree this modular approach and I have updated the changed. Done

cendhu · 2026-02-25T10:24:49Z

pkg/types/types.go

+}
+
+// TxNamespaceRecord represents a namespace within a transaction.
+type TxNamespaceRecord struct {


This just covers one nsID and one nsVersion. The transaction might have touched multiple namespaces. EITHER the method name is wrong or the fields.

@cendhu Done

cendhu · 2026-02-25T10:26:41Z

pkg/types/types.go

+}
+
+// ParsedBlockData contains writes, reads, and namespace records.
+type ParsedBlockData struct {


We always place the caller first and then the callee. Move this method to the top of this file.

@cendhu Done

cendhu · 2026-02-25T10:28:02Z

pkg/types/types.go

+
+// ParsedBlockData contains writes, reads, and namespace records.
+type ParsedBlockData struct {
+	Writes       []WriteRecord


this write-records seem to have a lot of duplicates -- if a tx is writing 10 fields, for each field, we would be duplicating

Namespace string
BlockNum uint64
TxNum uint64
TxID string

@cendhu Done

cendhu · 2026-02-25T10:29:38Z

pkg/constants/constants.go

+
+// Namespace constants
+const (
+	MetaNamespaceID = "_meta"


import this from fabric-x-common

@cendhu Done. I have removed the constant file, the plan is to import such constants from fabric-x-common or fabric-x-committer in future PRs, not define them locally.

cendhu · 2026-02-25T10:29:55Z

pkg/constants/constants.go

+
+// Transaction validation codes
+const (
+	StatusCommitted = 0


import this from fabric-x-common

@cendhu Done

cendhu · 2026-02-25T10:30:06Z

pkg/constants/constants.go

+)
+
+// API default limits
+const (


what are these limits used for?

@cendhu Done. The plan is to import such constants from fabric-x-common or fabric-x-committer in future PRs, not define them locally.

cendhu · 2026-02-25T10:31:52Z

pkg/db/db_writer.go

+		}
+	}()
+
+	q := dbsqlc.New(tx)


can't you use batch in pgx to apply all changes in a single round-trip to the database? or is it taken care by the dbsqlc?

@cendhu Done.. changed to batch processing

cendhu · 2026-02-25T10:32:54Z

.stacked-pr-plan.md

@@ -0,0 +1,107 @@
+# Stacked PR Plan — fabric-x-block-explorer


remove this

@cendhu Done

Signed-off-by: Farooq Azam Wasim <Farooq.Azm.Wasim-CIC.Netherlands@ibm.com>

…h pgx Signed-off-by: Farooq Azam Wasim <Farooq.Azm.Wasim-CIC.Netherlands@ibm.com>

…, tx_blind_writes Signed-off-by: Farooq Azam Wasim <Farooq.Azm.Wasim-CIC.Netherlands@ibm.com>

Signed-off-by: Farooq Azam Wasim <Farooq.Azm.Wasim-CIC.Netherlands@ibm.com>

BlockInfo.Number already carries the block number. The top-level Number field on ProcessedBlock was never read and duplicated the same value set by the parser. Signed-off-by: Farooq Azam Wasim <Farooq.Azm.Wasim-CIC.Netherlands@ibm.com>

cendhu

LGTM.

cendhu · 2026-03-03T05:10:08Z

pkg/db/dbtest/testcontainer.go

+func PrepareTestEnv(t *testing.T) *TestContainer {
+	t.Helper()
+
+	ctx := context.Background()


for test, we should always use t.Context()

cendhu · 2026-03-03T05:10:38Z

pkg/db/dbtest/testcontainer.go

+}
+
+// cleanDatabase drops all tables for a clean test state.
+func cleanDatabase(ctx context.Context, t *testing.T, pool *pgxpool.Pool) {


It is better to drop the whole database instead of tables.

cendhu · 2026-03-03T05:11:32Z

pkg/db/db_test.go

+func TestDatabaseTestEnv(t *testing.T) {
+	env := NewDatabaseTestEnv(t)
+
+	ctx := context.Background()


t.Context()

feat: add db writer, types, logging, util and test infrastructure

3ab8235

Signed-off-by: Farooq Azam Wasim <Farooq.Azm.Wasim-CIC.Netherlands@ibm.com>

cendhu reviewed Feb 25, 2026

View reviewed changes

farooqazamwasimnl added 8 commits February 25, 2026 11:43

chore: remove stacked PR plan file

4d51ac1

Signed-off-by: Farooq Azam Wasim <Farooq.Azm.Wasim-CIC.Netherlands@ibm.com>

refactor: replace custom logging with flogging from hyperledger/fabric

08ffc4e

Signed-off-by: Farooq Azam Wasim <Farooq.Azm.Wasim-CIC.Netherlands@ibm.com>

refactor: remove constants package, defer to fabric types in later PRs

8f17469

Signed-off-by: Farooq Azam Wasim <Farooq.Azm.Wasim-CIC.Netherlands@ibm.com>

refactor: restructure types hierarchically to eliminate duplicate fields

eaf49fe

Signed-off-by: Farooq Azam Wasim <Farooq.Azm.Wasim-CIC.Netherlands@ibm.com>

refactor: use natural composite PKs throughout, batch all inserts wit…

1dd20bf

…h pgx Signed-off-by: Farooq Azam Wasim <Farooq.Azm.Wasim-CIC.Netherlands@ibm.com>

refactor: split tx_reads/tx_writes into tx_reads_only, tx_read_writes…

d79b53c

…, tx_blind_writes Signed-off-by: Farooq Azam Wasim <Farooq.Azm.Wasim-CIC.Netherlands@ibm.com>

chore: add golangci-lint config and clean up code style

5776dcf

Signed-off-by: Farooq Azam Wasim <Farooq.Azm.Wasim-CIC.Netherlands@ibm.com>

cendhu approved these changes Mar 3, 2026

View reviewed changes

cendhu merged commit 4609424 into LF-Decentralized-Trust-labs:main Mar 3, 2026
2 checks passed

		@@ -0,0 +1,107 @@
		# Stacked PR Plan — fabric-x-block-explorer

Conversation

farooqazamwasimnl commented Feb 24, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

cendhu left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

farooqazamwasimnl Feb 25, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

farooqazamwasimnl Feb 25, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

farooqazamwasimnl Feb 25, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

cendhu left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

farooqazamwasimnl commented Feb 24, 2026 •

edited

Loading

farooqazamwasimnl Feb 25, 2026 •

edited

Loading

farooqazamwasimnl Feb 25, 2026 •

edited

Loading

farooqazamwasimnl Feb 25, 2026 •

edited

Loading