From a9156992f4d268c9e7e9c63c2f6d9d9d3be34860 Mon Sep 17 00:00:00 2001
From: LittleSheep <littlesheep.code@hotmail.com>
Date: Fri, 18 Oct 2024 23:36:54 +0800
Subject: [PATCH] :bug: Fix bug crash caused by hash small files

---
 .idea/workspace.xml               |  6 ++--
 pkg/internal/services/analyzer.go | 55 ++++++++++++++++++-------------
 2 files changed, 35 insertions(+), 26 deletions(-)

diff --git a/.idea/workspace.xml b/.idea/workspace.xml
index ca03800..c2c7670 100644
--- a/.idea/workspace.xml
+++ b/.idea/workspace.xml
@@ -4,7 +4,7 @@
     <option name="autoReloadType" value="ALL" />
   </component>
   <component name="ChangeListManager">
-    <list default="true" id="18dd0d68-b4b8-40db-9734-9119b5c848bd" name="更改" comment=":zap: Use exif whitelist to prevent produce garbage data">
+    <list default="true" id="18dd0d68-b4b8-40db-9734-9119b5c848bd" name="更改" comment=":bug: Trying to prevent exiftool causing analyze failed">
       <change beforePath="$PROJECT_DIR$/.idea/workspace.xml" beforeDir="false" afterPath="$PROJECT_DIR$/.idea/workspace.xml" afterDir="false" />
       <change beforePath="$PROJECT_DIR$/pkg/internal/services/analyzer.go" beforeDir="false" afterPath="$PROJECT_DIR$/pkg/internal/services/analyzer.go" afterDir="false" />
     </list>
@@ -117,7 +117,6 @@
     </option>
   </component>
   <component name="VcsManagerConfiguration">
-    <MESSAGE value=":sparkles: Pool clean by lifecycle config" />
     <MESSAGE value=":recycle: Split mark and delete file" />
     <MESSAGE value=":bug: Fix migration issue on pools" />
     <MESSAGE value=":bug: Fix schedule deletion will delete referenced file" />
@@ -142,7 +141,8 @@
     <MESSAGE value=":lock: Fix Attachment will contains GPS information" />
     <MESSAGE value=":sparkles: Save EXIF into file metadata" />
     <MESSAGE value=":zap: Use exif whitelist to prevent produce garbage data" />
-    <option name="LAST_COMMIT_MESSAGE" value=":zap: Use exif whitelist to prevent produce garbage data" />
+    <MESSAGE value=":bug: Trying to prevent exiftool causing analyze failed" />
+    <option name="LAST_COMMIT_MESSAGE" value=":bug: Trying to prevent exiftool causing analyze failed" />
   </component>
   <component name="VgoProject">
     <settings-migrated>true</settings-migrated>
diff --git a/pkg/internal/services/analyzer.go b/pkg/internal/services/analyzer.go
index 6a8de2f..8603ad0 100644
--- a/pkg/internal/services/analyzer.go
+++ b/pkg/internal/services/analyzer.go
@@ -277,32 +277,41 @@ func HashAttachment(file models.Attachment) (hash string, err error) {
 
 	hasher := sha256.New()
 
-	// Hash the first 32KB
-	buf := make([]byte, chunkSize)
-	if _, err := inFile.Read(buf); err != nil && err != io.EOF {
-		return "", fmt.Errorf("error reading file: %v", err)
-	}
-	hasher.Write(buf)
+	if chunkSize*3 <= fileInfo.Size() {
+		// If the total size is smaller than three chunks, then hash the whole file
+		buf := make([]byte, fileInfo.Size())
+		if _, err := inFile.Read(buf); err != nil && err != io.EOF {
+			return "", fmt.Errorf("error reading whole file: %v", err)
+		}
+		hasher.Write(buf)
+	} else {
+		// Hash the first 32KB
+		buf := make([]byte, chunkSize)
+		if _, err := inFile.Read(buf); err != nil && err != io.EOF {
+			return "", fmt.Errorf("error reading file: %v", err)
+		}
+		hasher.Write(buf)
 
-	// Hash the middle 32KB
-	middleOffset := fileInfo.Size() / 2
-	if _, err := inFile.Seek(middleOffset, io.SeekStart); err != nil {
-		return "", fmt.Errorf("error seeking to middle: %v", err)
-	}
-	if _, err := inFile.Read(buf); err != nil && err != io.EOF {
-		return "", fmt.Errorf("error reading middle: %v", err)
-	}
-	hasher.Write(buf)
+		// Hash the middle 32KB
+		middleOffset := fileInfo.Size() / 2
+		if _, err := inFile.Seek(middleOffset, io.SeekStart); err != nil {
+			return "", fmt.Errorf("error seeking to middle: %v", err)
+		}
+		if _, err := inFile.Read(buf); err != nil && err != io.EOF {
+			return "", fmt.Errorf("error reading middle: %v", err)
+		}
+		hasher.Write(buf)
 
-	// Hash the last 32KB
-	endOffset := fileInfo.Size() - chunkSize
-	if _, err := inFile.Seek(endOffset, io.SeekStart); err != nil {
-		return "", fmt.Errorf("error seeking to end: %v", err)
+		// Hash the last 32KB
+		endOffset := fileInfo.Size() - chunkSize
+		if _, err := inFile.Seek(endOffset, io.SeekStart); err != nil {
+			return "", fmt.Errorf("error seeking to end: %v", err)
+		}
+		if _, err := inFile.Read(buf); err != nil && err != io.EOF {
+			return "", fmt.Errorf("error reading end: %v", err)
+		}
+		hasher.Write(buf)
 	}
-	if _, err := inFile.Read(buf); err != nil && err != io.EOF {
-		return "", fmt.Errorf("error reading end: %v", err)
-	}
-	hasher.Write(buf)
 
 	// Hash with the file metadata
 	hasher.Write([]byte(fmt.Sprintf("%d", file.Size)))