From fef37cdd228de7054c228a6800798b49f2034369 Mon Sep 17 00:00:00 2001
From: Richard Gomez <rmgomez368@gmail.com>
Date: Mon, 14 Oct 2024 13:09:43 -0400
Subject: [PATCH] feat: propagate file info in log context

---
 pkg/sources/filesystem/filesystem.go |  6 +++---
 pkg/sources/git/git.go               | 31 ++++++++--------------------
 pkg/sources/github/github.go         |  6 ++++--
 3 files changed, 16 insertions(+), 27 deletions(-)

diff --git a/pkg/sources/filesystem/filesystem.go b/pkg/sources/filesystem/filesystem.go
index 74ae9bcae133..98805882d932 100644
--- a/pkg/sources/filesystem/filesystem.go
+++ b/pkg/sources/filesystem/filesystem.go
@@ -153,7 +153,7 @@ func (s *Source) scanDir(ctx context.Context, path string, chunksChan chan *sour
 var skipSymlinkErr = errors.New("skipping symlink")
 
 func (s *Source) scanFile(ctx context.Context, path string, chunksChan chan *sources.Chunk) error {
-	logger := ctx.Logger().WithValues("path", path)
+	fileCtx := context.WithValues(ctx, "path", path)
 	fileStat, err := os.Lstat(path)
 	if err != nil {
 		return fmt.Errorf("unable to stat file: %w", err)
@@ -168,7 +168,7 @@ func (s *Source) scanFile(ctx context.Context, path string, chunksChan chan *sou
 	}
 	defer inputFile.Close()
 
-	logger.V(3).Info("scanning file")
+	fileCtx.Logger().V(3).Info("scanning file")
 
 	chunkSkel := &sources.Chunk{
 		SourceType: s.Type(),
@@ -185,7 +185,7 @@ func (s *Source) scanFile(ctx context.Context, path string, chunksChan chan *sou
 		Verify: s.verify,
 	}
 
-	return handlers.HandleFile(ctx, inputFile, chunkSkel, sources.ChanReporter{Ch: chunksChan})
+	return handlers.HandleFile(fileCtx, inputFile, chunkSkel, sources.ChanReporter{Ch: chunksChan})
 }
 
 // Enumerate implements SourceUnitEnumerator interface. This implementation simply
diff --git a/pkg/sources/git/git.go b/pkg/sources/git/git.go
index ca6f82f967b6..c6d35c436e45 100644
--- a/pkg/sources/git/git.go
+++ b/pkg/sources/git/git.go
@@ -657,9 +657,8 @@ func (s *Git) ScanCommits(ctx context.Context, repo *git.Repository, path string
 				logger.Error(
 					err,
 					"error handling binary file",
-					"filename", fileName,
 					"commit", commitHash,
-					"file", diff.PathB,
+					"path", fileName,
 				)
 			}
 			continue
@@ -677,9 +676,8 @@ func (s *Git) ScanCommits(ctx context.Context, repo *git.Repository, path string
 			if err != nil {
 				ctx.Logger().Error(
 					err, "error creating reader for commits",
-					"filename", fileName,
 					"commit", fullHash,
-					"file", diff.PathB,
+					"path", fileName,
 				)
 				return nil
 			}
@@ -687,11 +685,10 @@ func (s *Git) ScanCommits(ctx context.Context, repo *git.Repository, path string
 
 			data := make([]byte, d.Len())
 			if _, err := io.ReadFull(reader, data); err != nil {
-				ctx.Logger().Error(
+				logger.Error(
 					err, "error reading diff content for commit",
-					"filename", fileName,
 					"commit", fullHash,
-					"file", diff.PathB,
+					"path", fileName,
 				)
 				return nil
 			}
@@ -829,7 +826,7 @@ func (s *Git) ScanStaged(ctx context.Context, repo *git.Repository, path string,
 	)
 	for diff := range diffChan {
 		fullHash := diff.Commit.Hash
-		logger := ctx.Logger().WithValues("filename", diff.PathB, "commit", fullHash, "file", diff.PathB)
+		logger := ctx.Logger().WithValues("commit", fullHash, "path", diff.PathB)
 		logger.V(2).Info("scanning staged changes from git")
 
 		if scanOptions.MaxDepth > 0 && depth >= scanOptions.MaxDepth {
@@ -877,7 +874,7 @@ func (s *Git) ScanStaged(ctx context.Context, repo *git.Repository, path string,
 				Verify:         s.verify,
 			}
 			if err := s.handleBinary(ctx, gitDir, reporter, chunkSkel, commitHash, fileName); err != nil {
-				logger.Error(err, "error handling binary file", "filename", fileName)
+				logger.Error(err, "error handling binary file")
 			}
 			continue
 		}
@@ -887,24 +884,14 @@ func (s *Git) ScanStaged(ctx context.Context, repo *git.Repository, path string,
 
 			reader, err := d.ReadCloser()
 			if err != nil {
-				ctx.Logger().Error(
-					err, "error creating reader for staged",
-					"filename", fileName,
-					"commit", fullHash,
-					"file", diff.PathB,
-				)
+				logger.Error(err, "error creating reader for staged")
 				return nil
 			}
 			defer reader.Close()
 
 			data := make([]byte, d.Len())
 			if _, err := reader.Read(data); err != nil {
-				ctx.Logger().Error(
-					err, "error reading diff content for staged",
-					"filename", fileName,
-					"commit", fullHash,
-					"file", diff.PathB,
-				)
+				logger.Error(err, "error reading diff content for staged")
 				return nil
 			}
 			chunk := sources.Chunk{
@@ -1255,7 +1242,7 @@ func (s *Git) handleBinary(ctx context.Context, gitDir string, reporter sources.
 
 	defer func() { _ = cmd.Wait() }()
 
-	return handlers.HandleFile(ctx, stdout, chunkSkel, reporter, handlers.WithSkipArchives(s.skipArchives))
+	return handlers.HandleFile(fileCtx, stdout, chunkSkel, reporter, handlers.WithSkipArchives(s.skipArchives))
 }
 
 func (s *Source) Enumerate(ctx context.Context, reporter sources.UnitReporter) error {
diff --git a/pkg/sources/github/github.go b/pkg/sources/github/github.go
index b5d3ef90e487..8c945834731d 100644
--- a/pkg/sources/github/github.go
+++ b/pkg/sources/github/github.go
@@ -1492,8 +1492,10 @@ func (s *Source) scanTarget(ctx context.Context, target sources.ChunkingTarget,
 		SourceMetadata: &source_metadatapb.MetaData{
 			Data: &source_metadatapb.MetaData_Github{Github: meta},
 		},
-		Verify: s.verify}
-	return handlers.HandleFile(ctx, readCloser, &chunkSkel, reporter)
+		Verify: s.verify,
+	}
+	fileCtx := context.WithValues(ctx, "path", meta.GetFile())
+	return handlers.HandleFile(fileCtx, readCloser, &chunkSkel, reporter)
 }
 
 func (s *Source) ChunkUnit(ctx context.Context, unit sources.SourceUnit, reporter sources.ChunkReporter) error {