CutDiffAroundLine makes the incorrect assumption that --- and +++ always represent part of the header of a diff.

This PR adds a flag to its parsing to prevent this problem and adds a streaming parsing technique to CutDiffAroundLine using an io.pipe instead of just sending data to an unbounded buffer.

Fix #14711

Signed-off-by: Andrew Thornton <[email protected]>

Author: zeripath

modules/git/diff.go

@@ -125,30 +125,40 @@ var hunkRegex = regexp.MustCompile(`^@@ -(?P<beginOld>[0-9]+)(,(?P<endOld>[0-9]+
 
 const cmdDiffHead = "diff --git "
 
-func isHeader(lof string) bool {
-	return strings.HasPrefix(lof, cmdDiffHead) || strings.HasPrefix(lof, "---") || strings.HasPrefix(lof, "+++")
+func isHeader(lof string, inHunk bool) bool {
+
+	return strings.HasPrefix(lof, cmdDiffHead) || (!inHunk && (strings.HasPrefix(lof, "---") || strings.HasPrefix(lof, "+++")))
 }
 
 // CutDiffAroundLine cuts a diff of a file in way that only the given line + numberOfLine above it will be shown
 // it also recalculates hunks and adds the appropriate headers to the new diff.
 // Warning: Only one-file diffs are allowed.
-func CutDiffAroundLine(originalDiff io.Reader, line int64, old bool, numbersOfLine int) string {
+func CutDiffAroundLine(originalDiff io.Reader, line int64, old bool, numbersOfLine int) (string, error) {
 	if line == 0 || numbersOfLine == 0 {
 		// no line or num of lines => no diff
-		return ""
+		return "", nil
 	}
+
 	scanner := bufio.NewScanner(originalDiff)
 	hunk := make([]string, 0)
+
 	// begin is the start of the hunk containing searched line
 	// end is the end of the hunk ...
 	// currentLine is the line number on the side of the searched line (differentiated by old)
 	// otherLine is the line number on the opposite side of the searched line (differentiated by old)
 	var begin, end, currentLine, otherLine int64
 	var headerLines int
+
+	inHunk := false
+
 	for scanner.Scan() {
 		lof := scanner.Text()
 		// Add header to enable parsing
-		if isHeader(lof) {
+
+		if isHeader(lof, inHunk) {
+			if strings.HasPrefix(lof, cmdDiffHead) {
+				inHunk = false
+			}
 			hunk = append(hunk, lof)
 			headerLines++
 		}
@@ -157,6 +167,7 @@ func CutDiffAroundLine(originalDiff io.Reader, line int64, old bool, numbersOfLi
 		}
 		// Detect "hunk" with contains commented lof
 		if strings.HasPrefix(lof, "@@") {
+			inHunk = true
 			// Already got our hunk. End of hunk detected!
 			if len(hunk) > headerLines {
 				break
@@ -213,15 +224,19 @@ func CutDiffAroundLine(originalDiff io.Reader, line int64, old bool, numbersOfLi
 			}
 		}
 	}
+	err := scanner.Err()
+	if err != nil {
+		return "", err
+	}
 
 	// No hunk found
 	if currentLine == 0 {
-		return ""
+		return "", nil
 	}
 	// headerLines + hunkLine (1) = totalNonCodeLines
 	if len(hunk)-headerLines-1 <= numbersOfLine {
 		// No need to cut the hunk => return existing hunk
-		return strings.Join(hunk, "\n")
+		return strings.Join(hunk, "\n"), nil
 	}
 	var oldBegin, oldNumOfLines, newBegin, newNumOfLines int64
 	if old {
@@ -256,5 +271,5 @@ func CutDiffAroundLine(originalDiff io.Reader, line int64, old bool, numbersOfLi
 	// construct the new hunk header
 	newHunk[headerLines] = fmt.Sprintf("@@ -%d,%d +%d,%d @@",
 		oldBegin, oldNumOfLines, newBegin, newNumOfLines)
-	return strings.Join(newHunk, "\n")
+	return strings.Join(newHunk, "\n"), nil
 }

modules/git/diff_test.go

@@ -23,8 +23,27 @@ const exampleDiff = `diff --git a/README.md b/README.md
 + cut off
 + cut off`
 
+const breakingDiff = `diff --git a/aaa.sql b/aaa.sql
+index d8e4c92..19dc8ad 100644
+--- a/aaa.sql
++++ b/aaa.sql
+@@ -1,9 +1,10 @@
+ --some comment
+--- some comment 5
++--some coment 2
++-- some comment 3
+ create or replace procedure test(p1 varchar2)
+ is
+ begin
+---new comment
+ dbms_output.put_line(p1);
++--some other comment
+ end;
+ /
+`
+
 func TestCutDiffAroundLine(t *testing.T) {
-	result := CutDiffAroundLine(strings.NewReader(exampleDiff), 4, false, 3)
+	result, _ := CutDiffAroundLine(strings.NewReader(exampleDiff), 4, false, 3)
 	resultByLine := strings.Split(result, "\n")
 	assert.Len(t, resultByLine, 7)
 	// Check if headers got transferred
@@ -37,18 +56,44 @@ func TestCutDiffAroundLine(t *testing.T) {
 	assert.Equal(t, "+ Build Status", resultByLine[4])
 
 	// Must be same result as before since old line 3 == new line 5
-	newResult := CutDiffAroundLine(strings.NewReader(exampleDiff), 3, true, 3)
+	newResult, _ := CutDiffAroundLine(strings.NewReader(exampleDiff), 3, true, 3)
 	assert.Equal(t, result, newResult, "Must be same result as before since old line 3 == new line 5")
 
-	newResult = CutDiffAroundLine(strings.NewReader(exampleDiff), 6, false, 300)
+	newResult, _ = CutDiffAroundLine(strings.NewReader(exampleDiff), 6, false, 300)
 	assert.Equal(t, exampleDiff, newResult)
 
-	emptyResult := CutDiffAroundLine(strings.NewReader(exampleDiff), 6, false, 0)
+	emptyResult, _ := CutDiffAroundLine(strings.NewReader(exampleDiff), 6, false, 0)
 	assert.Empty(t, emptyResult)
 
 	// Line is out of scope
-	emptyResult = CutDiffAroundLine(strings.NewReader(exampleDiff), 434, false, 0)
+	emptyResult, _ = CutDiffAroundLine(strings.NewReader(exampleDiff), 434, false, 0)
 	assert.Empty(t, emptyResult)
+
+	// Handle minus diffs properly
+	minusDiff, _ := CutDiffAroundLine(strings.NewReader(breakingDiff), 2, false, 4)
+
+	expected := `diff --git a/aaa.sql b/aaa.sql
+--- a/aaa.sql
++++ b/aaa.sql
+@@ -1,9 +1,10 @@
+ --some comment
+--- some comment 5
++--some coment 2`
+	assert.Equal(t, expected, minusDiff)
+
+	// Handle minus diffs properly
+	minusDiff, _ = CutDiffAroundLine(strings.NewReader(breakingDiff), 3, false, 4)
+
+	expected = `diff --git a/aaa.sql b/aaa.sql
+--- a/aaa.sql
++++ b/aaa.sql
+@@ -1,9 +1,10 @@
+ --some comment
+--- some comment 5
++--some coment 2
++-- some comment 3`
+
+	assert.Equal(t, expected, minusDiff)
 }
 
 func BenchmarkCutDiffAroundLine(b *testing.B) {
@@ -69,7 +114,7 @@ func ExampleCutDiffAroundLine() {
  Docker Pulls
 + cut off
 + cut off`
-	result := CutDiffAroundLine(strings.NewReader(diff), 4, false, 3)
+	result, _ := CutDiffAroundLine(strings.NewReader(diff), 4, false, 3)
 	println(result)
 }
 

modules/migrations/gitea_uploader.go

@@ -6,7 +6,6 @@
 package migrations
 
 import (
-	"bytes"
 	"context"
 	"fmt"
 	"io"
@@ -802,13 +801,20 @@ func (g *GiteaLocalUploader) CreateReviews(reviews ...*base.Review) error {
 			}
 
 			var patch string
-			patchBuf := new(bytes.Buffer)
-			if err := git.GetRepoRawDiffForFile(g.gitRepo, pr.MergeBase, headCommitID, git.RawDiffNormal, comment.TreePath, patchBuf); err != nil {
-				// We should ignore the error since the commit maybe removed when force push to the pull request
-				log.Warn("GetRepoRawDiffForFile failed when migrating [%s, %s, %s, %s]: %v", g.gitRepo.Path, pr.MergeBase, headCommitID, comment.TreePath, err)
-			} else {
-				patch = git.CutDiffAroundLine(patchBuf, int64((&models.Comment{Line: int64(line + comment.Position - 1)}).UnsignedLine()), line < 0, setting.UI.CodeCommentLines)
-			}
+			reader, writer := io.Pipe()
+			defer func() {
+				_ = reader.Close()
+				_ = writer.Close()
+			}()
+			go func() {
+				if err := git.GetRepoRawDiffForFile(g.gitRepo, pr.MergeBase, headCommitID, git.RawDiffNormal, comment.TreePath, writer); err != nil {
+					// We should ignore the error since the commit maybe removed when force push to the pull request
+					log.Warn("GetRepoRawDiffForFile failed when migrating [%s, %s, %s, %s]: %v", g.gitRepo.Path, pr.MergeBase, headCommitID, comment.TreePath, err)
+				}
+				_ = writer.Close()
+			}()
+
+			patch, _ = git.CutDiffAroundLine(reader, int64((&models.Comment{Line: int64(line + comment.Position - 1)}).UnsignedLine()), line < 0, setting.UI.CodeCommentLines)
 
 			var c = models.Comment{
 				Type:        models.CommentTypeCode,

services/pull/review.go

@@ -6,13 +6,14 @@
 package pull
 
 import (
-	"bytes"
 	"fmt"
+	"io"
 	"regexp"
 	"strings"
 
 	"code.gitea.io/gitea/models"
 	"code.gitea.io/gitea/modules/git"
+	"code.gitea.io/gitea/modules/log"
 	"code.gitea.io/gitea/modules/notification"
 	"code.gitea.io/gitea/modules/setting"
 )
@@ -179,11 +180,24 @@ func createCodeComment(doer *models.User, repo *models.Repository, issue *models
 		if len(commitID) == 0 {
 			commitID = headCommitID
 		}
-		patchBuf := new(bytes.Buffer)
-		if err := git.GetRepoRawDiffForFile(gitRepo, pr.MergeBase, headCommitID, git.RawDiffNormal, treePath, patchBuf); err != nil {
-			return nil, fmt.Errorf("GetRawDiffForLine[%s, %s, %s, %s]: %v", gitRepo.Path, pr.MergeBase, headCommitID, treePath, err)
+		reader, writer := io.Pipe()
+		defer func() {
+			_ = reader.Close()
+			_ = writer.Close()
+		}()
+		go func() {
+			if err := git.GetRepoRawDiffForFile(gitRepo, pr.MergeBase, headCommitID, git.RawDiffNormal, treePath, writer); err != nil {
+				_ = writer.CloseWithError(fmt.Errorf("GetRawDiffForLine[%s, %s, %s, %s]: %v", gitRepo.Path, pr.MergeBase, headCommitID, treePath, err))
+				return
+			}
+			_ = writer.Close()
+		}()
+
+		patch, err = git.CutDiffAroundLine(reader, int64((&models.Comment{Line: line}).UnsignedLine()), line < 0, setting.UI.CodeCommentLines)
+		if err != nil {
+			log.Error("Error whilst generating patch: %v", err)
+			return nil, err
 		}
-		patch = git.CutDiffAroundLine(patchBuf, int64((&models.Comment{Line: line}).UnsignedLine()), line < 0, setting.UI.CodeCommentLines)
 	}
 	return models.CreateComment(&models.CreateCommentOptions{
 		Type:        models.CommentTypeCode,