enhance: truncate but show long lines in diffs (#39279)

Co-authored-by: wxiaoguang <wxiaoguang@gmail.com>

modules/charset/escape.go

@@ -8,6 +8,7 @@ import (
 	"io"
 	"strings"
 
+	"gitea.dev/modules/htmlutil"
 	"gitea.dev/modules/setting"
 	"gitea.dev/modules/translation"
 )
@@ -27,13 +28,19 @@ func EscapeOptionsForView() EscapeOptions {
 	}
 }
 
+func EscapeControlHTMLTo(html template.HTML, locale translation.Locale, w htmlutil.HTMLWriter, opts ...EscapeOptions) *EscapeStatus {
+	if !setting.UI.AmbiguousUnicodeDetection {
+		w.WriteHTML(html)
+		return &EscapeStatus{}
+	}
+	escaped, _ := EscapeControlReader(strings.NewReader(string(html)), w.OriginWriter(), locale, opts...)
+	return escaped
+}
+
 // EscapeControlHTML escapes the Unicode control sequences in a provided html document
 func EscapeControlHTML(html template.HTML, locale translation.Locale, opts ...EscapeOptions) (escaped *EscapeStatus, output template.HTML) {
-	if !setting.UI.AmbiguousUnicodeDetection {
+	sb, w := htmlutil.NewHTMLStringWriter()
-		return &EscapeStatus{}, html
+	escaped = EscapeControlHTMLTo(html, locale, w, opts...)
-	}
-	sb := &strings.Builder{}
-	escaped, _ = EscapeControlReader(strings.NewReader(string(html)), sb, locale, opts...) // err has been handled in EscapeControlReader
 	return escaped, template.HTML(sb.String())
 }
 

modules/git/diff.go

@@ -140,6 +140,14 @@ func isHeader(lof string, inHunk bool) bool {
 	return strings.HasPrefix(lof, cmdDiffHead) || (!inHunk && (strings.HasPrefix(lof, "---") || strings.HasPrefix(lof, "+++")))
 }
 
+func NewGitDiffScanner(r io.Reader) *bufio.Scanner {
+	// TODO: GIT-DIFF-PARSE-LONG-LINE: ideally it shouldn't use bufio.Scanner which has a limit.
+	// It will cause errors if a line is very long.
+	scanner := bufio.NewScanner(r)
+	scanner.Buffer(nil, max(512*1024, int(setting.UI.MaxDisplayFileSize/16)))
+	return scanner
+}
+
 // 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.
@@ -149,7 +157,7 @@ func CutDiffAroundLine(originalDiff io.Reader, line int64, old bool, numbersOfLi
 		return "", nil
 	}
 
-	scanner := bufio.NewScanner(originalDiff)
+	scanner := NewGitDiffScanner(originalDiff)
 	hunk := make([]string, 0)
 
 	// begin is the start of the hunk containing searched line

modules/htmlutil/html.go

@@ -93,7 +93,7 @@ type HTMLWriter interface {
 	OriginWriter() io.Writer
 	WriteString(s string) HTMLWriter
 	WriteHTML(s template.HTML) HTMLWriter
-	WriteFormat(fmt template.HTML, args ...any) HTMLWriter
+	WriteFormatf(fmt template.HTML, args ...any) HTMLWriter
 	Err() error
 }
 
@@ -120,7 +120,7 @@ func (h *htmlWriter) WriteHTML(s template.HTML) HTMLWriter {
 	return h
 }
 
-func (h *htmlWriter) WriteFormat(fmt template.HTML, args ...any) HTMLWriter {
+func (h *htmlWriter) WriteFormatf(fmt template.HTML, args ...any) HTMLWriter {
 	if _, err := HTMLPrintf(h.w, fmt, args...); err != nil {
 		h.errs = append(h.errs, err)
 	}
@@ -135,6 +135,11 @@ func NewHTMLWriter(w io.Writer) HTMLWriter {
 	return &htmlWriter{w: w}
 }
 
+func NewHTMLStringWriter() (*strings.Builder, HTMLWriter) {
+	sb := &strings.Builder{}
+	return sb, &htmlWriter{w: sb}
+}
+
 type HTMLBuilder struct {
 	sb strings.Builder
 }

modules/htmlutil/html_test.go

@@ -34,7 +34,7 @@ func TestHTMLBuilder(t *testing.T) {
 func TestHTMLWriter(t *testing.T) {
 	sb := new(strings.Builder)
 	w := NewHTMLWriter(sb)
-	w.WriteString("<").WriteHTML("<hr>").WriteFormat("<span>%s%s</span>", ">", EscapeString(">"))
+	w.WriteString("<").WriteHTML("<hr>").WriteFormatf("<span>%s%s</span>", ">", EscapeString(">"))
 	assert.Equal(t, "&lt;<hr><span>&gt;&gt;</span>", sb.String())
 	assert.NoError(t, w.Err())
 }

modules/markup/jupyter/jupyter.go

@@ -39,18 +39,18 @@ type mimeHandler struct {
 }
 
 func renderCellCodeOutputTextPlain(w htmlutil.HTMLWriter, text string) error {
-	w.WriteFormat(`<div class="cell-output-text"><pre>%s</pre></div>`, text)
+	w.WriteFormatf(`<div class="cell-output-text"><pre>%s</pre></div>`, text)
 	return w.Err()
 }
 
 func renderCellCodeOutputUnsupported(w htmlutil.HTMLWriter, message string) error {
-	w.WriteFormat(`<div class="cell-output-unsupported">%s</div>`, message)
+	w.WriteFormatf(`<div class="cell-output-unsupported">%s</div>`, message)
 	return w.Err()
 }
 
 var dataMimeHandlers = sync.OnceValue(func() []mimeHandler {
 	renderImage := func(w htmlutil.HTMLWriter, subtype, payload string) error {
-		w.WriteFormat(`<div class="cell-output-image"><img src="data:image/%s;base64,%s"></div>`, subtype, payload)
+		w.WriteFormatf(`<div class="cell-output-image"><img src="data:image/%s;base64,%s"></div>`, subtype, payload)
 		return w.Err()
 	}
 	renderUnsupportedOutput := func(message string) func(htmlutil.HTMLWriter, string) error {
@@ -75,11 +75,11 @@ var dataMimeHandlers = sync.OnceValue(func() []mimeHandler {
 			// To future developers:  don't allow custom CSS classes or attributes,
 			// because ".link-action" or "data-fetch-xxx" can send POST requests and lead to XSS.
 			// If you'd really like to support more, do remember to correctly sanitize the values.
-			w.WriteFormat(`<div class="cell-output-html">%s</div>`, markup.Sanitize(d))
+			w.WriteFormatf(`<div class="cell-output-html">%s</div>`, markup.Sanitize(d))
 			return w.Err()
 		}},
 		{"text/latex", func(w htmlutil.HTMLWriter, d string) error {
-			w.WriteFormat(`<div class="cell-output-latex"><pre><code class="language-math display">%s</code></pre></div>`, trimMathDelimiters(d))
+			w.WriteFormatf(`<div class="cell-output-latex"><pre><code class="language-math display">%s</code></pre></div>`, trimMathDelimiters(d))
 			return w.Err()
 		}},
 		{"text/plain", renderCellCodeOutputTextPlain},
@@ -142,14 +142,14 @@ func (renderer) Render(ctx *markup.RenderContext, input io.Reader, outputWriter
 	// the size is (should be) checked and/or limited by the caller to avoid OOM
 	var notebook Notebook
 	if err := json.NewDecoder(input).Decode(&notebook); err != nil {
-		htmlWriter.WriteFormat(`<div class="ui error message">Failed to parse notebook JSON: %v</div>`, err)
+		htmlWriter.WriteFormatf(`<div class="ui error message">Failed to parse notebook JSON: %v</div>`, err)
 		return htmlWriter.Err()
 	}
 
 	// Check nbformat version
 	if notebook.Nbformat < 4 {
 		msg := htmlutil.HTMLFormat("This notebook uses an older format (nbformat %d). Only nbformat 4+ is supported for rendering. Please upgrade the notebook in Jupyter or view the raw JSON.", notebook.Nbformat)
-		htmlWriter.WriteFormat(`<div class="file-not-rendered-prompt">%s</div>`, msg)
+		htmlWriter.WriteFormatf(`<div class="file-not-rendered-prompt">%s</div>`, msg)
 		return htmlWriter.Err()
 	}
 
@@ -205,7 +205,7 @@ func renderCellCode(output htmlutil.HTMLWriter, cell Cell, language string) erro
 	output.WriteHTML(`<div class="cell-line">`)
 	{
 		if executionCount != nil {
-			output.WriteFormat(`<div class="cell-left cell-prompt">In [%d]:</div>`, *executionCount)
+			output.WriteFormatf(`<div class="cell-left cell-prompt">In [%d]:</div>`, *executionCount)
 		} else {
 			output.WriteHTML(`<div class="cell-left cell-prompt">In [ ]:</div>`)
 		}
@@ -213,7 +213,7 @@ func renderCellCode(output htmlutil.HTMLWriter, cell Cell, language string) erro
 		// Highlight code
 		preAttrs, codeAttrs := highlight.CodeBlockAttributes(language)
 		lexer := highlight.DetectChromaLexerByFileName("", language)
-		output.WriteFormat(`<div class="cell-right cell-input"><pre %s><code %s>`, preAttrs, codeAttrs)
+		output.WriteFormatf(`<div class="cell-right cell-input"><pre %s><code %s>`, preAttrs, codeAttrs)
 		output.WriteHTML(highlight.RenderCodeByLexer(lexer, source))
 		output.WriteHTML("</code></pre></div>")
 	}
@@ -232,7 +232,7 @@ func renderCellCode(output htmlutil.HTMLWriter, cell Cell, language string) erro
 		output.WriteHTML(`<div class="cell-line">`)
 		{
 			if hasExecutionResult && executionCount != nil {
-				output.WriteFormat(`<div class="cell-left cell-prompt">Out [%d]:</div>`, *executionCount)
+				output.WriteFormatf(`<div class="cell-left cell-prompt">Out [%d]:</div>`, *executionCount)
 			} else {
 				output.WriteHTML(`<div class="cell-left cell-prompt"></div>`)
 			}
@@ -250,7 +250,7 @@ func renderCellCode(output htmlutil.HTMLWriter, cell Cell, language string) erro
 }
 
 func renderCellPrompt(output htmlutil.HTMLWriter, left, right template.HTML) {
-	output.WriteFormat(`
+	output.WriteFormatf(`
 <div class="notebook-cell">
 	<div class="cell-line">
 		<div class="cell-left cell-prompt">%s</div>
@@ -335,7 +335,7 @@ func renderCellCodeOutput(output htmlutil.HTMLWriter, out Output) {
 	// Stream output
 	if out.OutputType == "stream" && out.Text != nil {
 		streamName := util.Iif(out.Name == "stderr", "stderr", "stdout")
-		output.WriteFormat(`<pre class="cell-output-stream stream-%s">%s</pre>`, streamName, joinSource(out.Text))
+		output.WriteFormatf(`<pre class="cell-output-stream stream-%s">%s</pre>`, streamName, joinSource(out.Text))
 		return
 	}
 
@@ -352,7 +352,7 @@ func renderCellCodeOutput(output htmlutil.HTMLWriter, out Output) {
 		if traceback == "" && out.Ename != "" {
 			traceback = fmt.Sprintf("%s: %s", out.Ename, out.Evalue)
 		}
-		output.WriteFormat(`<pre class="cell-output-error">%s</pre>`, traceback)
+		output.WriteFormatf(`<pre class="cell-output-error">%s</pre>`, traceback)
 		return
 	}
 

options/locale/locale_en-US.json

@@ -2608,8 +2608,8 @@
   "repo.diff.file_image_width": "Width",
   "repo.diff.file_image_height": "Height",
   "repo.diff.file_byte_size": "Size",
+  "repo.diff.line_truncated": "Line truncated",
   "repo.diff.file_suppressed": "File diff suppressed because it is too large.",
-  "repo.diff.file_suppressed_line_too_long": "File diff suppressed because one or more lines are too long.",
   "repo.diff.too_many_files": "Loaded %[1]d of %[2]d files, more files were not shown because too many files have changed in this diff.",
   "repo.diff.show_more": "Show more",
   "repo.diff.load": "Load diff",

services/gitdiff/gitdiff.go

@@ -9,15 +9,16 @@ import (
 	"bytes"
 	"context"
 	"fmt"
-	"html"
 	"html/template"
 	"io"
 	"net/url"
 	"path"
+	"slices"
 	"sort"
 	"strconv"
 	"strings"
 	"time"
+	"unicode/utf8"
 
 	"gitea.dev/models/db"
 	git_model "gitea.dev/models/git"
@@ -79,8 +80,9 @@ type DiffLine struct {
 	Content     string
 	Comments    issues_model.CommentList // related PR code comments
 	SectionInfo *DiffLineSectionInfo
+	IsTruncated bool
 
-	cachedDiffInline *DiffInline
+	cachedDiffInline *DiffInlineComputed
 }
 
 // DiffLineSectionInfo represents diff line section metadata
@@ -327,28 +329,32 @@ func defaultDiffMatchPatch() *diffmatchpatch.DiffMatchPatch {
 	return dmp
 }
 
-// DiffInline is a struct that has a content and escape status
+// DiffInlineComputed is the final computed diff line for template rendering
-type DiffInline struct {
+type DiffInlineComputed struct {
 	EscapeStatus *charset.EscapeStatus
 	Content      template.HTML
+	IsTruncated  bool
 }
 
-// diffInlineWithUnicodeEscape makes a DiffInline with hidden Unicode characters escaped
+// computeDiffInline makes a DiffInline with computed content, e.g.: Unicode escaping, truncation hint, etc
-func diffInlineWithUnicodeEscape(s template.HTML, locale translation.Locale) DiffInline {
+func computeDiffInline(s template.HTML, isTruncated bool, locale translation.Locale) DiffInlineComputed {
-	status, content := charset.EscapeControlHTML(s, locale)
+	sb, w := htmlutil.NewHTMLStringWriter()
-	return DiffInline{EscapeStatus: status, Content: content}
+	status := charset.EscapeControlHTMLTo(s, locale, w)
+	if isTruncated {
+		w.WriteFormatf(`<span class="ui label diff-line-truncated">%s</span>`, locale.Tr("repo.diff.line_truncated"))
+	}
+	return DiffInlineComputed{EscapeStatus: status, IsTruncated: isTruncated, Content: template.HTML(sb.String())}
 }
 
 func (diffSection *DiffSection) getLineContentForRender(lineIdx int, diffLine *DiffLine, fileLanguage string, highlightLines map[int]template.HTML) template.HTML {
-	h, ok := highlightLines[lineIdx-1]
+	if h, ok := highlightLines[lineIdx-1]; ok && !diffLine.IsTruncated {
-	if ok {
 		return h
 	}
 	if diffLine.Content == "" {
 		return ""
 	}
 	if setting.Git.DisableDiffHighlight {
-		return template.HTML(html.EscapeString(diffLine.Content[1:]))
+		return htmlutil.EscapeString(diffLine.Content[1:])
 	}
 	if diffSection.highlightLexer.value == nil {
 		diffSection.highlightLexer.value = highlight.DetectChromaLexerByFileName(diffSection.FileName, fileLanguage)
@@ -356,7 +362,7 @@ func (diffSection *DiffSection) getLineContentForRender(lineIdx int, diffLine *D
 	return highlight.RenderCodeByLexer(diffSection.highlightLexer.value, diffLine.Content[1:])
 }
 
-func (diffSection *DiffSection) getDiffLineForRender(diffLineType DiffLineType, leftLine, rightLine *DiffLine, locale translation.Locale) DiffInline {
+func (diffSection *DiffSection) getDiffLineForRender(diffLineType DiffLineType, leftLine, rightLine *DiffLine, locale translation.Locale) DiffInlineComputed {
 	sideIdx := util.Iif(diffLineType == DiffLineDel, 0, 1) // del=left, add=right
 	lines := [2]*DiffLine{leftLine, rightLine}
 
@@ -376,36 +382,39 @@ func (diffSection *DiffSection) getDiffLineForRender(diffLineType DiffLineType,
 		// left and right are the same, no need to do line-level diff, can just pick any side
 		// caller always uses the "right side" for this type
 		lineHTML := diffSection.getLineContentForRender(rightLine.RightIdx, rightLine, fileLanguage, highlightedRightLines)
-		return diffInlineWithUnicodeEscape(lineHTML, locale)
+		return computeDiffInline(lineHTML, rightLine.IsTruncated, locale)
 	}
 
 	var diffs [2]template.HTML
+	var truncations [2]bool
 	if leftLine != nil {
 		diffs[0] = diffSection.getLineContentForRender(leftLine.LeftIdx, leftLine, fileLanguage, highlightedLeftLines)
+		truncations[0] = leftLine.IsTruncated
 	}
 	if rightLine != nil {
 		diffs[1] = diffSection.getLineContentForRender(rightLine.RightIdx, rightLine, fileLanguage, highlightedRightLines)
+		truncations[1] = rightLine.IsTruncated
 	}
 
-	if leftLine != nil && rightLine != nil {
+	if leftLine != nil && rightLine != nil && !truncations[0] && !truncations[1] {
 		// if only some parts of a line are changed, highlight these changed parts as "deleted/added".
 		// "diff" the left&right sides together, then cache the diff result for another side,
 		// because when viewing the diff page, both "deleted" and "added" lines will to be rendered eventually,
 		// so here only diff them once, then next render can just use the cached result, no need to "diff" again.
 		hcd := newHighlightCodeDiff()
 		lineHTMLDel, lineHTMLAdd := hcd.diffLineWithHighlight(diffs[0], diffs[1])
-		leftLine.cachedDiffInline = new(diffInlineWithUnicodeEscape(lineHTMLDel, locale))
+		leftLine.cachedDiffInline = new(computeDiffInline(lineHTMLDel, truncations[0], locale))
-		rightLine.cachedDiffInline = new(diffInlineWithUnicodeEscape(lineHTMLAdd, locale))
+		rightLine.cachedDiffInline = new(computeDiffInline(lineHTMLAdd, truncations[1], locale))
 		return *lines[sideIdx].cachedDiffInline
 	}
 
-	// if left is empty or right is empty (a line is fully deleted or added), then we do not need to diff anymore.
+	// if left is empty or right is empty (a line is fully deleted or added), or either side is truncated (too long),
-	// the tmpl code already adds background colors for these cases.
+	// then we do not need to diff anymore, the tmpl code already adds background colors for these cases.
-	return diffInlineWithUnicodeEscape(diffs[sideIdx], locale)
+	return computeDiffInline(diffs[sideIdx], truncations[sideIdx], locale)
 }
 
 // GetComputedInlineDiffFor computes inline diff for the given line.
-func (diffSection *DiffSection) GetComputedInlineDiffFor(diffLine *DiffLine, locale translation.Locale) DiffInline {
+func (diffSection *DiffSection) GetComputedInlineDiffFor(diffLine *DiffLine, locale translation.Locale) DiffInlineComputed {
 	defer func() {
 		if err := recover(); err != nil {
 			// the logic is too complex in this function, help to catch any panic because Golang template doesn't print the stack
@@ -416,7 +425,7 @@ func (diffSection *DiffSection) GetComputedInlineDiffFor(diffLine *DiffLine, loc
 	switch diffLine.Type {
 	case DiffLineSection:
 		// section content is a diff hunk header, it isn't code diff, its trailing context might come from the file content, might not
-		return diffInlineWithUnicodeEscape(htmlutil.EscapeString(diffLine.Content), locale)
+		return computeDiffInline(htmlutil.EscapeString(diffLine.Content), diffLine.IsTruncated, locale)
 	case DiffLineAdd:
 		compareDiffLine := diffSection.GetLine(diffLine.Match)
 		return diffSection.getDiffLineForRender(DiffLineAdd, compareDiffLine, diffLine, locale)
@@ -457,8 +466,8 @@ type DiffFile struct {
 	IsSubmodule  bool
 	// basic fields but for render purpose only
 	Sections          []*DiffSection
-	IsIncomplete            bool
+	IsIncomplete      bool // file is too large
-	IsIncompleteLineTooLong bool
+	HasTruncatedLines bool // some lines are too long
 
 	// will be filled by the extra loop in GitDiffForRender
 	IsGenerated       bool
@@ -492,6 +501,10 @@ func (diffFile *DiffFile) GetType() int {
 	return int(diffFile.Type)
 }
 
+func (diffFile *DiffFile) CanShowFileViewToggle() bool {
+	return diffFile.IsBlobTypeImage || (diffFile.IsBlobTypeCsv && !diffFile.IsIncomplete && !diffFile.HasTruncatedLines)
+}
+
 type DiffRenderDetail struct {
 	needTailSection               bool
 	leftLineCount, rightLineCount int
@@ -685,62 +698,64 @@ func (diff *Diff) LoadComments(ctx context.Context, issue *issues_model.Issue, c
 
 const cmdDiffHead = "diff --git "
 
-// ParsePatch builds a Diff object from a io.Reader and some parameters.
+// to correctly parse a diff line, the input buffer size should be large enough to read a full line for git diff headers
-func ParsePatch(ctx context.Context, maxLines, maxLineCharacters, maxFiles int, reader io.Reader, skipToFile string) (*Diff, error) {
+var defaultDiffLineBufferSize = 8 * 1024
-	log.Debug("ParsePatch(%d, %d, %d, ..., %s)", maxLines, maxLineCharacters, maxFiles, skipToFile)
-	var curFile *DiffFile
 
-	skipping := skipToFile != ""
+// ParsePatch builds a Diff object by parsing git diff output
+func ParsePatch(ctx context.Context, maxLines, maxLineCharacters, maxFiles int, reader io.Reader, skipToFile string) (_ *Diff, retErr error) {
+	log.Debug("ParsePatch(%d, %d, %d, ..., %s)", maxLines, maxLineCharacters, maxFiles, skipToFile)
 
 	diff := &Diff{Files: make([]*DiffFile, 0)}
-
+	readerSize := max(maxLineCharacters, defaultDiffLineBufferSize)
-	sb := strings.Builder{}
-
-	// OK let's set a reasonable buffer size.
-	// This should be at least the size of maxLineCharacters or 4096 whichever is larger.
-	readerSize := max(maxLineCharacters, 4096)
 
 	input := bufio.NewReaderSize(reader, readerSize)
 	line, err := input.ReadString('\n')
 	if err != nil {
-		if err == io.EOF {
+		return diff, util.Iif(err == io.EOF, nil, err)
-			return diff, nil
-		}
-		return diff, err
 	}
 
-	prepareValue := func(s, p string) string {
+	skipping := skipToFile != ""
+	for {
+		nextLine, err := diff.parseOneDiffFile(ctx, maxLines, maxLineCharacters, maxFiles, &skipping, input, skipToFile, line)
+		if nextLine == "" || err == io.EOF {
+			break
+		} else if err != nil {
+			return diff, err
+		}
+		line = nextLine
+	}
+
+	diff.postProcessFiles()
+	return diff, nil
+}
+
+func (diff *Diff) parseOneDiffFile(ctx context.Context, maxLines, maxLineCharacters, maxFiles int, skipping *bool, input *bufio.Reader, skipToFile, startLine string) (nextLine string, err error) {
+	line := startLine
+
+	extractGitDiffHead := func(s, p string) string {
 		return strings.TrimSpace(strings.TrimPrefix(s, p))
 	}
 
-parsingLoop:
+	{
-	for {
 		// 1. A patch file always begins with `diff --git ` + `a/path b/path` (possibly quoted)
 		// if it does not we have bad input!
 		if !strings.HasPrefix(line, cmdDiffHead) {
-			return diff, fmt.Errorf("invalid first file line: %s", line)
+			return "", fmt.Errorf("invalid first file line: %s", line)
 		}
 
 		if maxFiles > -1 && len(diff.Files) >= maxFiles {
 			lastFile := createDiffFile(line)
 			diff.End = lastFile.Name
 			diff.IsIncomplete = true
-			break parsingLoop
+			return "", nil
 		}
 
-		curFile = createDiffFile(line)
+		curFile := createDiffFile(line)
-		if skipping {
+		if *skipping {
 			if curFile.Name != skipToFile {
-				line, err = skipToNextDiffHead(input)
+				return skipToNextDiffHead(input)
-				if err != nil {
-					if err == io.EOF {
-						return diff, nil
 			}
-					return diff, err
+			*skipping = false
-				}
-				continue
-			}
-			skipping = false
 		}
 
 		diff.Files = append(diff.Files, curFile)
@@ -783,27 +798,23 @@ parsingLoop:
 		//     Binary files a/<path> and b/<path> differ
 		//
 		// but one of a/<path> and b/<path> could be /dev/null.
-	curFileLoop:
 		for {
 			line, err = input.ReadString('\n')
 			if err != nil {
-				if err != io.EOF {
+				return line, err
-					return diff, err
-				}
-				break parsingLoop
 			}
 
 			switch {
 			case strings.HasPrefix(line, cmdDiffHead):
-				break curFileLoop
+				return line, nil
 			case strings.HasPrefix(line, "old mode ") ||
 				strings.HasPrefix(line, "new mode "):
 
 				if strings.HasPrefix(line, "old mode ") {
-					curFile.OldEntryMode = prepareValue(line, "old mode ")
+					curFile.OldEntryMode = extractGitDiffHead(line, "old mode ")
 				}
 				if strings.HasPrefix(line, "new mode ") {
-					curFile.EntryMode = prepareValue(line, "new mode ")
+					curFile.EntryMode = extractGitDiffHead(line, "new mode ")
 				}
 				if strings.HasSuffix(line, " 160000\n") {
 					curFile.IsSubmodule, curFile.SubmoduleDiffInfo = true, &SubmoduleDiffInfo{}
@@ -812,33 +823,33 @@ parsingLoop:
 				curFile.IsRenamed = true
 				curFile.Type = DiffFileRename
 				if curFile.isAmbiguous {
-					curFile.OldName = prepareValue(line, "rename from ")
+					curFile.OldName = extractGitDiffHead(line, "rename from ")
 				}
 			case strings.HasPrefix(line, "rename to "):
 				curFile.IsRenamed = true
 				curFile.Type = DiffFileRename
 				if curFile.isAmbiguous {
-					curFile.Name = prepareValue(line, "rename to ")
+					curFile.Name = extractGitDiffHead(line, "rename to ")
 					curFile.isAmbiguous = false
 				}
 			case strings.HasPrefix(line, "copy from "):
 				curFile.IsRenamed = true
 				curFile.Type = DiffFileCopy
 				if curFile.isAmbiguous {
-					curFile.OldName = prepareValue(line, "copy from ")
+					curFile.OldName = extractGitDiffHead(line, "copy from ")
 				}
 			case strings.HasPrefix(line, "copy to "):
 				curFile.IsRenamed = true
 				curFile.Type = DiffFileCopy
 				if curFile.isAmbiguous {
-					curFile.Name = prepareValue(line, "copy to ")
+					curFile.Name = extractGitDiffHead(line, "copy to ")
 					curFile.isAmbiguous = false
 				}
 			case strings.HasPrefix(line, "new file"):
 				curFile.Type = DiffFileAdd
 				curFile.IsCreated = true
 				if strings.HasPrefix(line, "new file mode ") {
-					curFile.EntryMode = prepareValue(line, "new file mode ")
+					curFile.EntryMode = extractGitDiffHead(line, "new file mode ")
 				}
 				if strings.HasSuffix(line, " 160000\n") {
 					curFile.IsSubmodule, curFile.SubmoduleDiffInfo = true, &SubmoduleDiffInfo{}
@@ -892,31 +903,16 @@ parsingLoop:
 					curFile.isAmbiguous = false
 				}
 				// Otherwise do nothing with this line, but now switch to parsing hunks
-				lineBytes, isFragment, err := parseHunks(ctx, curFile, maxLines, maxLineCharacters, input)
+				nextLine, err := parseHunks(ctx, curFile, maxLines, maxLineCharacters, input)
-				if err != nil {
+				return string(nextLine), err
-					if err != io.EOF {
+			default:
-						return diff, err
+				// ignore other extended header lines
-					}
-					break parsingLoop
-				}
-				sb.Reset()
-				_, _ = sb.Write(lineBytes)
-				for isFragment {
-					lineBytes, isFragment, err = input.ReadLine()
-					if err != nil {
-						// Now by the definition of ReadLine this cannot be io.EOF
-						return diff, fmt.Errorf("unable to ReadLine: %w", err)
-					}
-					_, _ = sb.Write(lineBytes)
-				}
-				line = sb.String()
-				sb.Reset()
-
-				break curFileLoop
 			}
 		}
 	}
+}
 
+func (diff *Diff) postProcessFiles() {
 	// TODO: There are numerous issues with this:
 	// - we might want to consider detecting encoding while parsing but...
 	// - we're likely to fail to get the correct encoding here anyway as we won't have enough information
@@ -964,38 +960,18 @@ parsingLoop:
 			}
 		}
 	}
-
-	return diff, nil
 }
 
 func skipToNextDiffHead(input *bufio.Reader) (line string, err error) {
-	// need to skip until the next cmdDiffHead
-	var isFragment, wasFragment bool
-	var lineBytes []byte
 	for {
-		lineBytes, isFragment, err = input.ReadLine()
+		lineBytes, _, err := readGitDiffLineWithDiscard(input)
 		if err != nil {
 			return "", err
 		}
-		if wasFragment {
-			wasFragment = isFragment
-			continue
-		}
 		if bytes.HasPrefix(lineBytes, []byte(cmdDiffHead)) {
-			break
+			return string(lineBytes), nil
 		}
-		wasFragment = isFragment
 	}
-	line = string(lineBytes)
-	if isFragment {
-		var tail string
-		tail, err = input.ReadString('\n')
-		if err != nil {
-			return "", err
-		}
-		line += tail
-	}
-	return line, err
 }
 
 func newDiffSectionForDiffFile(curFile *DiffFile) *DiffSection {
@@ -1007,9 +983,59 @@ func newDiffSectionForDiffFile(curFile *DiffFile) *DiffSection {
 	}
 }
 
-func parseHunks(ctx context.Context, curFile *DiffFile, maxLines, maxLineCharacters int, input *bufio.Reader) (lineBytes []byte, isFragment bool, err error) {
+func readGitDiffLineWithDiscard(r *bufio.Reader) (_ []byte, truncated bool, _ error) {
-	sb := strings.Builder{}
+	// HINT: GIT-DIFF-PARSE-LONG-LINE: it can't use Scanner which has a default limit and will cause errors if a line is very long
+	line, isPrefix, err := r.ReadLine()
+	if !isPrefix {
+		return line, false, err
+	}
 
+	if bytes.HasPrefix(line, []byte(cmdDiffHead)) {
+		// "diff head" is special, even if it is very long, we still want to fully read it
+		line = slices.Clone(line)
+		lineRemaining, err := r.ReadBytes('\n')
+		lineRemaining = bytes.TrimRight(lineRemaining, "\r\n")
+		return append(line, lineRemaining...), false, err
+	}
+
+	// discard remaining bytes, only return the prefix
+	line = slices.Clone(line)
+	for isPrefix && err == nil {
+		_, isPrefix, err = r.ReadLine()
+	}
+	return line, true, err
+}
+
+// tryFixTruncatedString tries to fix the truncated diff line by removing the last corrupted rune
+func tryFixTruncatedString(s string) string {
+	b := util.UnsafeStringToBytes(s)
+	var idx int
+	for idx = 0; idx < len(s); {
+		r, l := utf8.DecodeRune(b[idx:])
+		if r == utf8.RuneError && l == 1 {
+			break
+		}
+		idx += l
+	}
+	// for valid utf8 diff line, remove the last truncated rune
+	remainingLen := len(s) - idx
+	if idx > 0 && remainingLen < utf8.UTFMax {
+		return s[:idx]
+	}
+
+	// for non-utf8 diff line, try to find an ASCII char at the ending, remove the potentially truncated chars after that
+	for i := 1; i <= utf8.UTFMax; i++ {
+		idx = len(s) - i
+		if 0 < idx && idx+1 < len(s) && s[idx] < 127 {
+			return s[:idx+1]
+		}
+	}
+
+	// otherwise, just return the input as is
+	return s
+}
+
+func parseHunks(ctx context.Context, curFile *DiffFile, maxLines, maxLineCharacters int, input *bufio.Reader) (nextLine []byte, err error) {
 	var curSection *DiffSection
 	curFileLFSPrefix := false
 
@@ -1022,27 +1048,12 @@ func parseHunks(ctx context.Context, curFile *DiffFile, maxLines, maxLineCharact
 	}
 
 	for {
-		for isFragment {
+		lineBytes, truncated, err := readGitDiffLineWithDiscard(input)
-			curFile.IsIncomplete = true
-			curFile.IsIncompleteLineTooLong = true
-			_, isFragment, err = input.ReadLine()
 		if err != nil {
-				// Now by the definition of ReadLine this cannot be io.EOF
+			return lineBytes, err
-				return nil, false, fmt.Errorf("unable to ReadLine: %w", err)
 		}
-		}
+		if bytes.HasPrefix(lineBytes, []byte(cmdDiffHead)) {
-		sb.Reset()
+			return lineBytes, err
-		lineBytes, isFragment, err = input.ReadLine()
-		if err != nil {
-			if err == io.EOF {
-				return lineBytes, isFragment, err
-			}
-			err = fmt.Errorf("unable to ReadLine: %w", err)
-			return nil, false, err
-		}
-		if lineBytes[0] == 'd' {
-			// End of hunks
-			return lineBytes, isFragment, err
 		}
 
 		switch lineBytes[0] {
@@ -1051,19 +1062,7 @@ func parseHunks(ctx context.Context, curFile *DiffFile, maxLines, maxLineCharact
 				curFile.IsIncomplete = true
 				continue
 			}
-
+			line := string(lineBytes)
-			_, _ = sb.Write(lineBytes)
-			for isFragment {
-				// This is very odd indeed - we're in a section header and the line is too long
-				// This really shouldn't happen...
-				lineBytes, isFragment, err = input.ReadLine()
-				if err != nil {
-					// Now by the definition of ReadLine this cannot be io.EOF
-					return nil, false, fmt.Errorf("unable to ReadLine: %w", err)
-				}
-				_, _ = sb.Write(lineBytes)
-			}
-			line := sb.String()
 
 			// Create a new section to represent this hunk
 			curSection = newDiffSectionForDiffFile(curFile)
@@ -1086,7 +1085,7 @@ func parseHunks(ctx context.Context, curFile *DiffFile, maxLines, maxLineCharact
 		case '\\':
 			// This is used only to indicate that the current file does not have a terminal newline
 			if !bytes.Equal(lineBytes, []byte("\\ No newline at end of file")) {
-				return nil, false, fmt.Errorf("unexpected line in hunk: %s", string(lineBytes))
+				return nil, fmt.Errorf("unexpected line in hunk: %s", string(lineBytes))
 			}
 			// Technically this should be the end the file!
 			// FIXME: we should be putting a marker at the end of the file if there is no terminal new line
@@ -1170,27 +1169,23 @@ func parseHunks(ctx context.Context, curFile *DiffFile, maxLines, maxLineCharact
 			curSection.Lines = append(curSection.Lines, diffLine)
 		default:
 			// This is unexpected
-			return nil, false, fmt.Errorf("unexpected line in hunk: %s", string(lineBytes))
+			return nil, fmt.Errorf("unexpected line in hunk: %s", string(lineBytes))
 		}
 
+		curLine := curSection.Lines[len(curSection.Lines)-1]
+		curLine.IsTruncated = truncated
+
 		line := string(lineBytes)
-		if isFragment {
+		curFile.HasTruncatedLines = curFile.HasTruncatedLines || truncated
-			curFile.IsIncomplete = true
-			curFile.IsIncompleteLineTooLong = true
-			for isFragment {
-				lineBytes, isFragment, err = input.ReadLine()
-				if err != nil {
-					// Now by the definition of ReadLine this cannot be io.EOF
-					return lineBytes, isFragment, fmt.Errorf("unable to ReadLine: %w", err)
-				}
-			}
-		}
 		if len(line) > maxLineCharacters {
-			curFile.IsIncomplete = true
-			curFile.IsIncompleteLineTooLong = true
 			line = line[:maxLineCharacters]
+			curLine.IsTruncated = true
+		}
+		curLine.Content = line
+		if curLine.IsTruncated {
+			curFile.HasTruncatedLines = true
+			curLine.Content = tryFixTruncatedString(curLine.Content)
 		}
-		curSection.Lines[len(curSection.Lines)-1].Content = line
 
 		// handle LFS
 		if line[1:] == lfs.MetaFileIdentifier {
@@ -1489,7 +1484,7 @@ func highlightCodeLines(name, lang string, sections []*DiffSection, isLeft bool,
 			}
 			if lineIdx >= 1 {
 				idx := lineIdx - 1
-				if idx < len(unsafeLines) {
+				if idx < len(unsafeLines) && !ln.IsTruncated {
 					lines[idx] = template.HTML(util.UnsafeBytesToString(unsafeLines[idx]))
 				}
 			}

services/gitdiff/gitdiff_excerpt.go

@@ -4,12 +4,12 @@
 package gitdiff
 
 import (
-	"bufio"
 	"bytes"
 	"fmt"
 	"html/template"
 	"io"
 
+	"gitea.dev/modules/git"
 	"gitea.dev/modules/setting"
 
 	"github.com/alecthomas/chroma/v2"
@@ -29,7 +29,7 @@ type BlobExcerptOptions struct {
 
 func (diffSection *DiffSection) fillExcerptLines(reader io.Reader, leftStart, rightStart, chunkSize int) error {
 	buf := &bytes.Buffer{}
-	scanner := bufio.NewScanner(reader)
+	scanner := git.NewGitDiffScanner(reader)
 	var diffLines []*DiffLine
 	for rightLineIdx := 1; rightLineIdx < rightStart+chunkSize; rightLineIdx++ {
 		if ok := scanner.Scan(); !ok {

services/gitdiff/gitdiff_parse_test.go

@@ -0,0 +1,48 @@
+// Copyright 2026 The Gitea Authors. All rights reserved.
+// SPDX-License-Identifier: MIT
+
+package gitdiff
+
+import (
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+)
+
+func TestTryFixTruncatedString(t *testing.T) {
+	assert.Equal(t, "", tryFixTruncatedString(""))
+
+	t.Run("UTF8", func(t *testing.T) {
+		s := "🌞a🌛"
+		assert.Equal(t, s[:1], tryFixTruncatedString(s[:1]))
+		assert.Equal(t, s[:2], tryFixTruncatedString(s[:2]))
+		assert.Equal(t, s[:3], tryFixTruncatedString(s[:3]))
+		assert.Equal(t, "🌞", tryFixTruncatedString(s[:4]))
+		assert.Equal(t, "🌞a", tryFixTruncatedString(s[:5]))
+		assert.Equal(t, "🌞a", tryFixTruncatedString(s[:6]))
+		assert.Equal(t, "🌞a", tryFixTruncatedString(s[:7]))
+		assert.Equal(t, "🌞a", tryFixTruncatedString(s[:8]))
+		assert.Equal(t, "🌞a🌛", tryFixTruncatedString(s[:9]))
+
+		s = "a🌞🌛"
+		assert.Equal(t, "a", tryFixTruncatedString(s[:1]))
+		assert.Equal(t, "a", tryFixTruncatedString(s[:2]))
+		assert.Equal(t, "a", tryFixTruncatedString(s[:3]))
+		assert.Equal(t, "a", tryFixTruncatedString(s[:4]))
+		assert.Equal(t, "a🌞", tryFixTruncatedString(s[:5]))
+		assert.Equal(t, "a🌞", tryFixTruncatedString(s[:6]))
+	})
+
+	t.Run("Non-UTF8", func(t *testing.T) {
+		s := "\xff\xee\xff\xeeb\xff\xee\xff\xee"
+		assert.Equal(t, "\xff", tryFixTruncatedString(s[:1]))
+		assert.Equal(t, "\xff\xee", tryFixTruncatedString(s[:2]))
+		assert.Equal(t, "\xff\xee\xff", tryFixTruncatedString(s[:3]))
+		assert.Equal(t, "\xff\xee\xff\xee", tryFixTruncatedString(s[:4]))
+		assert.Equal(t, "\xff\xee\xff\xeeb", tryFixTruncatedString(s[:5]))
+		assert.Equal(t, "\xff\xee\xff\xeeb", tryFixTruncatedString(s[:6]))
+		assert.Equal(t, "\xff\xee\xff\xeeb", tryFixTruncatedString(s[:7]))
+		assert.Equal(t, "\xff\xee\xff\xeeb", tryFixTruncatedString(s[:8]))
+		assert.Equal(t, "\xff\xee\xff\xeeb\xff\xee\xff\xee", tryFixTruncatedString(s[:9]))
+	})
+}

services/gitdiff/gitdiff_test.go

@@ -19,7 +19,6 @@ import (
 	"gitea.dev/modules/json"
 	"gitea.dev/modules/setting"
 	"gitea.dev/modules/translation"
-	"gitea.dev/modules/util"
 
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
@@ -203,7 +202,7 @@ func TestParsePatch_singlefile(t *testing.T) {
 
 	tests := []testcase{
 		{
-			name: "readme.md2readme.md",
+			name: "same name",
 			gitdiff: `diff --git "\\a/README.md" "\\b/README.md"
 --- "\\a/README.md"
 +++ "\\b/README.md"
@@ -222,7 +221,7 @@ func TestParsePatch_singlefile(t *testing.T) {
 			oldFilename: "README.md",
 		},
 		{
-			name: "A \\ B",
+			name: `A \ B`,
 			gitdiff: `diff --git "a/A \\ B" "b/A \\ B"
 --- "a/A \\ B"
 +++ "b/A \\ B"
@@ -236,8 +235,8 @@ func TestParsePatch_singlefile(t *testing.T) {
 + cut off`,
 			addition:    4,
 			deletion:    1,
-			filename:    "A \\ B",
+			filename:    `A \ B`,
-			oldFilename: "A \\ B",
+			oldFilename: `A \ B`,
 		},
 		{
 			name: "really weird filename",
@@ -327,7 +326,7 @@ index 0000000..92e798b
 			deletion:    0,
 		},
 		{
-			name: "rename",
+			name: "ambiguous rename 1",
 			gitdiff: `diff --git a/b b/b b/b b/b b/b b/b
 similarity index 100%
 rename from b b/b b/b b/b b/b
@@ -337,17 +336,7 @@ rename to b
 			filename:    "b",
 		},
 		{
-			name: "ambiguous 1",
+			name: "ambiguous rename 2",
-			gitdiff: `diff --git a/b b/b b/b b/b b/b b/b
-similarity index 100%
-rename from b b/b b/b b/b b/b
-rename to b
-`,
-			oldFilename: "b b/b b/b b/b b/b",
-			filename:    "b",
-		},
-		{
-			name: "ambiguous 2",
 			gitdiff: `diff --git a/b b/b b/b b/b b/b b/b
 similarity index 100%
 rename from b b/b b/b b/b
@@ -442,8 +431,8 @@ index 0000000..6bb8f39
 	if err != nil {
 		t.Errorf("There should not be an error: %v", err)
 	}
-	if !result.Files[0].IsIncomplete {
+	if result.Files[0].IsIncomplete {
-		t.Errorf("Files should be incomplete! %v", result.Files[0])
+		t.Errorf("Files should not be incomplete! %v", result.Files[0])
 	}
 
 	// Test max characters
@@ -480,8 +469,8 @@ index 0000000..6bb8f39
 	if err != nil {
 		t.Errorf("There should not be an error: %v", err)
 	}
-	if !result.Files[0].IsIncomplete {
+	if result.Files[0].IsIncomplete {
-		t.Errorf("Files should be incomplete! %v", result.Files[0])
+		t.Errorf("Files should not be incomplete! %v", result.Files[0])
 	}
 
 	diff = `diff --git "a/README.md" "b/README.md"
@@ -549,6 +538,65 @@ index 0000000..6bb8f39
 	}
 }
 
+func TestParsePatchLongLines(t *testing.T) {
+	overSizedContent := strings.Repeat("a", defaultDiffLineBufferSize*2)
+	for _, test := range []struct {
+		name, full string
+		limit      int
+		result     string
+	}{
+		{name: "below", limit: 8, full: "short", result: "short"},
+		{name: "exact", limit: 8, full: "1234567", result: "1234567"},
+		{name: "above", limit: 8, full: "12345678", result: "1234567"},
+		{name: "multiple fragments", limit: 101, full: overSizedContent, result: overSizedContent[:100]},
+		{name: "cutoff-truncate", limit: 8, full: "a🙂b🙂", result: "a🙂b"},
+		{name: "cutoff-no-truncate", limit: 12, full: "a🙂b🙂", result: "a🙂b🙂"},
+	} {
+		for _, eol := range []string{"\n", "\r\n"} {
+			t.Run(test.name+"/"+strconv.Quote(eol), func(t *testing.T) {
+				patch := strings.Join([]string{
+					"diff --git a/first b/first",
+					"--- a/first",
+					"+++ b/first",
+					"@@ -1,3 +1,3 @@",
+					"-" + test.full,
+					"+" + test.full,
+					" " + test.full,
+					" short",
+					"@@ -10 +10 @@",
+					" next",
+					"diff --git a/second b/second",
+					"--- a/second", "+++ b/second",
+					"@@ -1 +1 @@",
+					" final",
+					"",
+				}, eol)
+				maxLines, maxFiles, skipToFile := 20, 10, ""
+				diff, err := ParsePatch(t.Context(), maxLines, test.limit, maxFiles, strings.NewReader(patch), skipToFile)
+				require.NoError(t, err)
+				require.Len(t, diff.Files, 2)
+				file := diff.Files[0]
+				assert.False(t, file.IsIncomplete)
+				assert.Equal(t, test.result != test.full, file.HasTruncatedLines)
+				require.Len(t, file.Sections, 2)
+				lines := file.Sections[0].Lines
+				require.Len(t, lines, 5)
+				for i, marker := range []string{"-", "+", " "} {
+					assert.Equal(t, marker+test.result, lines[i+1].Content)
+					assert.Equal(t, test.result != test.full, lines[i+1].IsTruncated)
+				}
+				assert.Equal(t, 2, lines[1].Match)
+				assert.Equal(t, 1, lines[2].Match)
+				assert.Equal(t, &DiffLine{Type: DiffLinePlain, LeftIdx: 3, RightIdx: 3, Content: " short"}, lines[4])
+				assert.Equal(t, &DiffLine{Type: DiffLinePlain, LeftIdx: 10, RightIdx: 10, Content: " next"}, file.Sections[1].Lines[1])
+				assert.False(t, diff.Files[1].IsIncomplete)
+				assert.False(t, diff.Files[1].HasTruncatedLines)
+				assert.Equal(t, &DiffLine{Type: DiffLinePlain, LeftIdx: 1, RightIdx: 1, Content: " final"}, diff.Files[1].Sections[0].Lines[1])
+			})
+		}
+	}
+}
+
 func TestParsePatchExactLineLimit(t *testing.T) {
 	for _, test := range []struct {
 		name, hunk   string
@@ -563,6 +611,7 @@ func TestParsePatchExactLineLimit(t *testing.T) {
 		{name: "deletion", limit: 1, lines: 1, hunk: "@@ -1 +0,0 @@\n-one\n"},
 		{name: "marker has no cost", limit: 1, lines: 1, hunk: "@@ -1 +1 @@\n line\n\\ No newline at end of file\n"},
 		{name: "hunk at capacity", limit: 1, lines: 1, hunk: "@@ -1 +1 @@\n one\n@@ -3 +3 @@\n three\n", incomplete: true},
+		{name: "long line after capacity", limit: 1, lines: 1, hunk: "@@ -0,0 +1,3 @@\n+one\n+" + strings.Repeat("x", 10000) + "\n+three\n", incomplete: true},
 	} {
 		t.Run(test.name, func(t *testing.T) {
 			patch := "diff --git a/file b/file\n--- a/file\n+++ b/file\n" + test.hunk
@@ -570,17 +619,12 @@ func TestParsePatchExactLineLimit(t *testing.T) {
 			require.NoError(t, err)
 			require.Len(t, diff.Files, 1)
 			diffFile := diff.Files[0]
+			assert.Equal(t, test.incomplete, diffFile.IsIncomplete)
 			if test.limit == 0 {
 				require.Len(t, diffFile.Sections, 0)
 			} else {
 				require.Len(t, diffFile.Sections, 1)
-				diffSection := diffFile.Sections[0]
+				assert.Len(t, diffFile.Sections[0].Lines, test.lines+1) // lines with a section
-				lineSecCount := 0
-				for _, line := range diffSection.Lines {
-					lineSecCount += util.Iif(line.Type == DiffLineSection, 1, 0)
-				}
-				assert.Equal(t, test.lines, len(diffSection.Lines)-lineSecCount) // actual diff lines
-				assert.Equal(t, test.incomplete, diffFile.IsIncomplete)
 			}
 		})
 	}
@@ -1165,6 +1209,22 @@ func TestDiffSection_GetComputedInlineDiffFor(t *testing.T) {
 		assert.True(t, diffInline.EscapeStatus.Escaped)
 		assert.Equal(t, `@@ -1,3 +1,3 @@ func <span class="escaped-code-point" data-escaped="[U+202E]"><span class="char">`+"\u202e"+`</span></span>name() &lt;b&gt;`, string(diffInline.Content))
 	})
+	t.Run("ShortLineUseHighlight", func(t *testing.T) {
+		diffFile := &DiffFile{}
+		diffFile.highlightedRightLines.value = map[int]template.HTML{0: "highlighted short line"}
+		diffLine := &DiffLine{Type: DiffLinePlain, RightIdx: 1, Content: " short line", IsTruncated: false}
+		inline := newDiffSectionForDiffFile(diffFile).GetComputedInlineDiffFor(diffLine, translation.MockLocale{})
+		assert.Equal(t, template.HTML("highlighted short line"), inline.Content)
+		assert.False(t, inline.IsTruncated)
+	})
+	t.Run("LongLineTruncated", func(t *testing.T) {
+		diffFile := &DiffFile{}
+		diffFile.highlightedRightLines.value = map[int]template.HTML{0: "highlighted line"}
+		diffLine := &DiffLine{Type: DiffLinePlain, RightIdx: 1, Content: " truncated line", IsTruncated: true}
+		inline := newDiffSectionForDiffFile(diffFile).GetComputedInlineDiffFor(diffLine, translation.MockLocale{})
+		assert.Equal(t, template.HTML(`truncated line<span class="ui label diff-line-truncated">repo.diff.line_truncated</span>`), inline.Content)
+		assert.True(t, inline.IsTruncated)
+	})
 }
 
 func TestHighlightCodeLines(t *testing.T) {
@@ -1214,6 +1274,19 @@ func TestHighlightCodeLines(t *testing.T) {
 		assert.Equal(t, "a␍b\n", string(ret[0]))
 		assert.Equal(t, `c`, string(ret[1]))
 	})
+
+	t.Run("LongLineTruncated", func(t *testing.T) {
+		diffFile := &DiffFile{
+			Name: "a.c",
+			Sections: []*DiffSection{
+				{
+					Lines: []*DiffLine{{LeftIdx: 1, IsTruncated: true}},
+				},
+			},
+		}
+		ret := highlightCodeLinesForDiffFile(diffFile, true, []byte("// anything"))
+		assert.Empty(t, ret)
+	})
 }
 
 func TestSyncUserSpecificDiff_UpdatedFiles(t *testing.T) {

templates/repo/diff/box.tmpl

@@ -90,7 +90,7 @@
 					{{/*notice: the index of Diff.Files should not be used for element ID, because the index will be restarted from 0 when doing load-more for PRs with a lot of files*/}}
 					{{$isImage:= $file.IsBlobTypeImage}}
 					{{$isCsv := $file.IsBlobTypeCsv}}
-					{{$showFileViewToggle := or $isImage (and (not $file.IsIncomplete) $isCsv)}}
+					{{$showFileViewToggle := $file.CanShowFileViewToggle}}
 					{{$isReviewFile := and $.IsSigned $.PageIsPullFiles (not $.Repository.IsArchived) $.IsShowingAllCommits}}
 					<div class="diff-file-box file-content {{TabSizeClass $.Editorconfig $file.Name}} tw-mt-0" id="diff-{{$file.NameHash}}" data-old-filename="{{$file.OldName}}" data-new-filename="{{$file.Name}}" {{if $file.ShouldBeHidden}}data-folded="true"{{end}}>
 						<div class="diff-file-header sticky-2nd-row ui top attached header">
@@ -173,12 +173,8 @@
 								{{if or $file.IsIncomplete $file.IsBin}}
 									<div class="tw-p-3">
 										{{if $file.IsIncomplete}}
-											{{if $file.IsIncompleteLineTooLong}}
-												{{ctx.Locale.Tr "repo.diff.file_suppressed_line_too_long"}}
-											{{else}}
 											{{ctx.Locale.Tr "repo.diff.file_suppressed"}}
 											<a class="ui basic tiny button" data-global-click="diffLoadFileBody" data-href="?file-only=true&files={{$file.Name}}&files={{$file.OldName}}">{{ctx.Locale.Tr "repo.diff.load"}}</a>
-											{{end}}
 										{{else}}
 											{{ctx.Locale.Tr "repo.diff.bin_not_shown"}}
 										{{end}}

tests/integration/pull_diff_test.go

@@ -5,12 +5,19 @@ package integration
 
 import (
 	"net/http"
+	"strings"
 	"testing"
 
+	repo_model "gitea.dev/models/repo"
+	"gitea.dev/models/unittest"
+	"gitea.dev/modules/git"
+	"gitea.dev/modules/setting"
+	"gitea.dev/modules/test"
 	"gitea.dev/tests"
 
 	"github.com/PuerkitoBio/goquery"
 	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
 )
 
 func TestPullDiff(t *testing.T) {
@@ -65,3 +72,61 @@ func testPullDiffAssertPage(t *testing.T, prDiffURL string, reviewBtnDisabled bo
 	// Ensure the review button is enabled for full PR reviews
 	assert.Equal(t, reviewBtnDisabled, doc.Find(".js-btn-review").HasClass("disabled"))
 }
+
+func TestLongLineDiffRendering(t *testing.T) {
+	defer tests.PrepareTestEnv(t)()
+	defer test.MockVariableValue(&setting.Git.MaxGitDiffLineCharacters, 32)()
+
+	repo := unittest.AssertExistsAndLoadBean(t, &repo_model.Repository{ID: 1})
+	const suffix = "DISCARDED_SUFFIX"
+	const short = "following short line"
+	contextLine := strings.Repeat("context ", 8) + suffix + "\n"
+	before := git.FastImportCommit{Ref: "refs/heads/long-line-before"}
+	after := git.FastImportCommit{Ref: "refs/heads/long-line-after"}
+	for _, name := range []string{"long.txt", "long.csv"} {
+		before.Files = append(before.Files, git.FastImportFile{Path: name, Content: strings.Repeat("old ", 16) + suffix + "\n" + contextLine + short + "\n"})
+		after.Files = append(after.Files, git.FastImportFile{Path: name, Content: strings.Repeat("new ", 16) + suffix + "\n" + contextLine + short + "\n"})
+	}
+	outsideHunk := "header\n" + contextLine + strings.Repeat("unchanged\n", 6)
+	before.Files = append(before.Files, git.FastImportFile{Path: "outside.csv", Content: outsideHunk + "old\n"})
+	after.Files = append(after.Files, git.FastImportFile{Path: "outside.csv", Content: outsideHunk + "new\n"})
+	require.NoError(t, git.ForceFastImport(t.Context(), repo.CodeStorageRepo(), []git.FastImportCommit{before, after}))
+
+	oldPrefix := strings.Repeat("old ", 8)[:31]
+	newPrefix := strings.Repeat("new ", 8)[:31]
+	contextPrefix := contextLine[:31]
+	for style, want := range map[string][]string{
+		"unified": {oldPrefix, newPrefix, contextPrefix, short},
+		"split":   {oldPrefix, newPrefix, contextPrefix, contextPrefix, short, short},
+	} {
+		t.Run(style, func(t *testing.T) {
+			req := NewRequest(t, "GET", "/user2/repo1/compare/long-line-before..long-line-after?style="+style)
+			resp := MakeRequest(t, req, http.StatusOK)
+			doc := NewHTMLParser(t, resp.Body)
+			for _, filename := range []string{"long.txt", "long.csv"} {
+				t.Run(filename, func(t *testing.T) {
+					diffFileBox := doc.Find(`.diff-file-box[data-new-filename="` + filename + `"]`)
+					diffBody := diffFileBox.Find(".code-diff-" + style)
+					require.Equal(t, 1, diffBody.Length())
+					assert.False(t, diffBody.HasClass("tw-hidden"))
+					assert.Empty(t, diffFileBox.Find(".file-view-toggle, .data-table").Nodes)
+					assert.NotContains(t, diffFileBox.Text(), suffix)
+					got := diffBody.Find("tr:not(.tag-code) code.code-inner").Map(func(_ int, s *goquery.Selection) string {
+						markerText := s.Find(".diff-line-truncated").Text()
+						textContent := strings.TrimSpace(strings.TrimSuffix(strings.TrimSpace(s.Text()), markerText))
+						if textContent == short {
+							assert.Empty(t, markerText)
+						} else {
+							assert.Equal(t, "Line truncated", markerText)
+						}
+						return textContent
+					})
+					assert.Equal(t, want, got)
+				})
+			}
+			outside := doc.Find(`.diff-file-box[data-new-filename="outside.csv"]`)
+			assert.Equal(t, 2, outside.Find(".file-view-toggle").Length())
+			assert.Equal(t, 1, outside.Find(".data-table").Length())
+		})
+	}
+}

web_src/css/base.css

@@ -686,6 +686,17 @@ overflow-menu .ui.label:empty {
   line-height: inherit; /* needed for inline code preview in markup */
 }
 
+.code-inner:has(.diff-line-truncated) {
+  text-decoration: underline dashed 1px;
+  text-underline-offset: 4px;
+}
+
+.code-inner .diff-line-truncated {
+  margin-left: var(--gap-inline);
+  padding: 0 4px;
+  user-select: none;
+}
+
 .blame .code-inner {
   white-space: pre-wrap;
   overflow-wrap: anywhere;