From e3b87dd4dc4bcea85753e5b5266d231b7aa3ad3f Mon Sep 17 00:00:00 2001
From: Somnath <snb895@outlook.com>
Date: Wed, 17 Apr 2024 16:58:28 +0400
Subject: [PATCH] Fix missing prune log index (#9968)

Existing prune had some confusing logic (thanks to me).

In PR, change the logic to:
if receipt, or any individual log in the receipt contains an address in
`noPruneContracts` the entire log must be stored and indexed.

Also fixes a regression that logs were not getting indexed - this was
due to
```
if l.BlockNumber < pruneBlock && cfg.noPruneContracts != nil && !cfg.noPruneContracts[l.Address] {
```
Since the individual log object didn't have the block number - something
that shouldn't matter.
In PR, change the logic to:
Use `blockNum` from outer loop
---
 eth/stagedsync/stage_execute.go   | 17 ++++++++---
 eth/stagedsync/stage_log_index.go | 51 +++++++++++++++++++++----------
 2 files changed, 48 insertions(+), 20 deletions(-)

diff --git a/eth/stagedsync/stage_execute.go b/eth/stagedsync/stage_execute.go
index c363def9b46..b58e605ebae 100644
--- a/eth/stagedsync/stage_execute.go
+++ b/eth/stagedsync/stage_execute.go
@@ -208,12 +208,21 @@ func executeBlock(
 func gatherNoPruneReceipts(receipts *types.Receipts, chainCfg *chain.Config) bool {
 	cr := types.Receipts{}
 	for _, r := range *receipts {
-		for _, l := range r.Logs {
-			if chainCfg.NoPruneContracts[l.Address] {
-				cr = append(cr, r)
-				break
+		toStore := false
+		if chainCfg.NoPruneContracts != nil && chainCfg.NoPruneContracts[r.ContractAddress] {
+			toStore = true
+		} else {
+			for _, l := range r.Logs {
+				if chainCfg.NoPruneContracts != nil && chainCfg.NoPruneContracts[l.Address] {
+					toStore = true
+					break
+				}
 			}
 		}
+
+		if toStore {
+			cr = append(cr, r)
+		}
 	}
 	receipts = &cr
 	if receipts.Len() > 0 {
diff --git a/eth/stagedsync/stage_log_index.go b/eth/stagedsync/stage_log_index.go
index 6919ffbfc0f..d0c19faa1c9 100644
--- a/eth/stagedsync/stage_log_index.go
+++ b/eth/stagedsync/stage_log_index.go
@@ -130,7 +130,7 @@ func promoteLogIndex(logPrefix string, tx kv.RwTx, start uint64, endBlock uint64
 	reader := bytes.NewReader(nil)
 
 	if endBlock != 0 && endBlock-start > 100 {
-		logger.Info(fmt.Sprintf("[%s] processing", logPrefix), "from", start, "to", endBlock)
+		logger.Info(fmt.Sprintf("[%s] processing", logPrefix), "from", start, "to", endBlock, "pruneTo", pruneBlock)
 	}
 
 	for k, v, err := logs.Seek(dbutils.LogKey(start, 0)); k != nil; k, v, err = logs.Next() {
@@ -178,10 +178,26 @@ func promoteLogIndex(logPrefix string, tx kv.RwTx, start uint64, endBlock uint64
 			return fmt.Errorf("receipt unmarshal failed: %w, blocl=%d", err, blockNum)
 		}
 
-		for _, l := range ll {
-			if l.BlockNumber < pruneBlock && cfg.noPruneContracts != nil && !cfg.noPruneContracts[l.Address] {
+		toStore := true
+		// if pruning is enabled, and noPruneContracts isn't configured for the chain, don't index
+		if blockNum < pruneBlock {
+			toStore = false
+			if cfg.noPruneContracts == nil {
 				continue
 			}
+			for _, l := range ll {
+				// if any of the log address is in noPrune, store and index all logs for this txId
+				if cfg.noPruneContracts[l.Address] {
+					toStore = true
+					break
+				}
+			}
+		}
+
+		if !toStore {
+			continue
+		}
+		for _, l := range ll {
 			for _, topic := range l.Topics {
 				topicStr := string(topic.Bytes())
 				m, ok := topics[topicStr]
@@ -477,27 +493,30 @@ func pruneLogIndex(logPrefix string, tx kv.RwTx, tmpDir string, pruneFrom, prune
 				return fmt.Errorf("receipt unmarshal failed: %w, block=%d", err, binary.BigEndian.Uint64(k))
 			}
 
-			notToPrune := false // To identify whether this log key has addr in noPruneContracts
+			toPrune := true
 			for _, l := range logs {
-				// If any of the logs have an address in noPruneContracts, then the whole tx is related to it, and must not be pruned, including addr and topic indexes
-				if noPruneContracts != nil && noPruneContracts[l.Address] || notToPrune {
-					notToPrune = true
-					continue
+				// No logs (or sublogs) for this txId should be pruned
+				// if one of the logs belongs to noPruneContracts lis
+				if noPruneContracts != nil && noPruneContracts[l.Address] {
+					toPrune = false
+					break
 				}
-				for _, topic := range l.Topics {
-					if err := topics.Collect(topic.Bytes(), nil); err != nil {
+			}
+			if toPrune {
+				for _, l := range logs {
+					for _, topic := range l.Topics {
+						if err := topics.Collect(topic.Bytes(), nil); err != nil {
+							return err
+						}
+					}
+					if err := addrs.Collect(l.Address.Bytes(), nil); err != nil {
 						return err
 					}
 				}
-				if err := addrs.Collect(l.Address.Bytes(), nil); err != nil {
+				if err := pruneLogKeyCollector.Collect(k, nil); err != nil {
 					return err
 				}
 			}
-			// No logs (or sublogs) for this txId should be pruned
-			// if one of the logs belongs to noPruneContracts list
-			if !notToPrune {
-				pruneLogKeyCollector.Collect(k, nil)
-			}
 		}
 	}