Rework file highlight rendering and fix yaml copy-paste (#19967)

* Rework file highlight rendering and fix yaml copy-paste

* use Split+Trim to replace tag parser

* remove unnecessary bytes.Count

* remove newLineInHTML = "
"

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

2 files changed, 164 insertions, 130 deletions

diff --git a/modules/highlight/highlight.go b/modules/highlight/highlight.go
index 6832207c0f..af3376e8d7 100644
--- a/modules/highlight/highlight.go
+++ b/modules/highlight/highlight.go
@@ -10,6 +10,7 @@ import (
 	"bytes"
 	"fmt"
 	gohtml "html"
+	"io"
 	"path/filepath"
 	"strings"
 	"sync"
@@ -26,7 +27,7 @@ import (
 )
 
 // don't index files larger than this many bytes for performance purposes
-const sizeLimit = 1000000
+const sizeLimit = 1024 * 1024
 
 var (
 	// For custom user mapping
@@ -46,7 +47,6 @@ func NewContext() {
 				highlightMapping[keys[i].Name()] = keys[i].Value()
 			}
 		}
-
 		// The size 512 is simply a conservative rule of thumb
 		c, err := lru.New2Q(512)
 		if err != nil {
@@ -60,7 +60,7 @@ func NewContext() {
 func Code(fileName, language, code string) string {
 	NewContext()
 
-	// diff view newline will be passed as empty, change to literal \n so it can be copied
+	// diff view newline will be passed as empty, change to literal '\n' so it can be copied
 	// preserve literal newline in blame view
 	if code == "" || code == "\n" {
 		return "\n"
@@ -128,36 +128,32 @@ func CodeFromLexer(lexer chroma.Lexer, code string) string {
 		return code
 	}
 
-	htmlw.Flush()
+	_ = htmlw.Flush()
 	// Chroma will add newlines for certain lexers in order to highlight them properly
-	// Once highlighted, strip them here so they don't cause copy/paste trouble in HTML output
+	// Once highlighted, strip them here, so they don't cause copy/paste trouble in HTML output
 	return strings.TrimSuffix(htmlbuf.String(), "\n")
 }
 
-// File returns a slice of chroma syntax highlighted lines of code
-func File(numLines int, fileName, language string, code []byte) []string {
+// File returns a slice of chroma syntax highlighted HTML lines of code
+func File(fileName, language string, code []byte) ([]string, error) {
 	NewContext()
 
 	if len(code) > sizeLimit {
-		return plainText(string(code), numLines)
+		return PlainText(code), nil
 	}
+
 	formatter := html.New(html.WithClasses(true),
 		html.WithLineNumbers(false),
 		html.PreventSurroundingPre(true),
 	)
 
-	if formatter == nil {
-		log.Error("Couldn't create chroma formatter")
-		return plainText(string(code), numLines)
-	}
-
-	htmlbuf := bytes.Buffer{}
-	htmlw := bufio.NewWriter(&htmlbuf)
+	htmlBuf := bytes.Buffer{}
+	htmlWriter := bufio.NewWriter(&htmlBuf)
 
 	var lexer chroma.Lexer
 
 	// provided language overrides everything
-	if len(language) > 0 {
+	if language != "" {
 		lexer = lexers.Get(language)
 	}
 
@@ -168,9 +164,9 @@ func File(numLines int, fileName, language string, code []byte) []string {
 	}
 
 	if lexer == nil {
-		language := analyze.GetCodeLanguage(fileName, code)
+		guessLanguage := analyze.GetCodeLanguage(fileName, code)
 
-		lexer = lexers.Get(language)
+		lexer = lexers.Get(guessLanguage)
 		if lexer == nil {
 			lexer = lexers.Match(fileName)
 			if lexer == nil {
@@ -181,54 +177,43 @@ func File(numLines int, fileName, language string, code []byte) []string {
 
 	iterator, err := lexer.Tokenise(nil, string(code))
 	if err != nil {
-		log.Error("Can't tokenize code: %v", err)
-		return plainText(string(code), numLines)
+		return nil, fmt.Errorf("can't tokenize code: %w", err)
 	}
 
-	err = formatter.Format(htmlw, styles.GitHub, iterator)
+	err = formatter.Format(htmlWriter, styles.GitHub, iterator)
 	if err != nil {
-		log.Error("Can't format code: %v", err)
-		return plainText(string(code), numLines)
+		return nil, fmt.Errorf("can't format code: %w", err)
 	}
 
-	htmlw.Flush()
-	finalNewLine := false
-	if len(code) > 0 {
-		finalNewLine = code[len(code)-1] == '\n'
-	}
+	_ = htmlWriter.Flush()
 
-	m := make([]string, 0, numLines)
-	for _, v := range strings.SplitN(htmlbuf.String(), "\n", numLines) {
-		content := v
-		// need to keep lines that are only \n so copy/paste works properly in browser
-		if content == "" {
-			content = "\n"
-		} else if content == `</span><span class="w">` {
-			content += "\n</span>"
-		} else if content == `</span></span><span class="line"><span class="cl">` {
-			content += "\n"
-		}
-		content = strings.TrimSuffix(content, `<span class="w">`)
-		content = strings.TrimPrefix(content, `</span>`)
-		m = append(m, content)
+	// at the moment, Chroma generates stable output `<span class="line"><span class="cl">...\n</span></span>` for each line
+	htmlStr := htmlBuf.String()
+	lines := strings.Split(htmlStr, `<span class="line"><span class="cl">`)
+	m := make([]string, 0, len(lines))
+	for i := 1; i < len(lines); i++ {
+		line := lines[i]
+		line = strings.TrimSuffix(line, "</span></span>")
+		m = append(m, line)
 	}
-	if finalNewLine {
-		m = append(m, "<span class=\"w\">\n</span>")
-	}
-
-	return m
+	return m, nil
 }
 
-// return unhiglighted map
-func plainText(code string, numLines int) []string {
-	m := make([]string, 0, numLines)
-	for _, v := range strings.SplitN(code, "\n", numLines) {
-		content := v
-		// need to keep lines that are only \n so copy/paste works properly in browser
-		if content == "" {
-			content = "\n"
+// PlainText returns non-highlighted HTML for code
+func PlainText(code []byte) []string {
+	r := bufio.NewReader(bytes.NewReader(code))
+	m := make([]string, 0, bytes.Count(code, []byte{'\n'})+1)
+	for {
+		content, err := r.ReadString('\n')
+		if err != nil && err != io.EOF {
+			log.Error("failed to read string from buffer: %v", err)
+			break
+		}
+		if content == "" && err == io.EOF {
+			break
 		}
-		m = append(m, gohtml.EscapeString(content))
+		s := gohtml.EscapeString(content)
+		m = append(m, s)
 	}
 	return m
 }
diff --git a/modules/highlight/highlight_test.go b/modules/highlight/highlight_test.go
index e5dfedd2b3..8f83f4a2f6 100644
--- a/modules/highlight/highlight_test.go
+++ b/modules/highlight/highlight_test.go
@@ -8,97 +8,146 @@ import (
 	"strings"
 	"testing"
 
-	"code.gitea.io/gitea/modules/setting"
-	"code.gitea.io/gitea/modules/util"
-
 	"github.com/stretchr/testify/assert"
-	"gopkg.in/ini.v1"
 )
 
+func lines(s string) []string {
+	return strings.Split(strings.ReplaceAll(strings.TrimSpace(s), `\n`, "\n"), "\n")
+}
+
 func TestFile(t *testing.T) {
-	setting.Cfg = ini.Empty()
 	tests := []struct {
-		name     string
-		numLines int
-		fileName string
-		code     string
-		want     string
+		name string
+		code string
+		want []string
 	}{
 		{
-			name:     ".drone.yml",
-			numLines: 12,
-			fileName: ".drone.yml",
-			code: util.Dedent(`
-				kind: pipeline
-				name: default
+			name: "empty.py",
+			code: "",
+			want: lines(""),
+		},
+		{
+			name: "tags.txt",
+			code: "<>",
+			want: lines("&lt;&gt;"),
+		},
+		{
+			name: "tags.py",
+			code: "<>",
+			want: lines(`<span class="o">&lt;</span><span class="o">&gt;</span>`),
+		},
+		{
+			name: "eol-no.py",
+			code: "a=1",
+			want: lines(`<span class="n">a</span><span class="o">=</span><span class="mi">1</span>`),
+		},
+		{
+			name: "eol-newline1.py",
+			code: "a=1\n",
+			want: lines(`<span class="n">a</span><span class="o">=</span><span class="mi">1</span>\n`),
+		},
+		{
+			name: "eol-newline2.py",
+			code: "a=1\n\n",
+			want: lines(`
+<span class="n">a</span><span class="o">=</span><span class="mi">1</span>\n
+\n
+			`,
+			),
+		},
+		{
+			name: "empty-line-with-space.py",
+			code: strings.ReplaceAll(strings.TrimSpace(`
+def:
+    a=1
 
-				steps:
-				- name: test
-					image: golang:1.13
-					environment:
-						GOPROXY: https://goproxy.cn
-					commands:
-					- go get -u
-					- go build -v
-					- go test -v -race -coverprofile=coverage.txt -covermode=atomic
-			`),
-			want: util.Dedent(`
-				<span class="line"><span class="cl"><span class="nt">kind</span><span class="p">:</span><span class="w"> </span><span class="l">pipeline</span>
-				</span></span><span class="line"><span class="cl"><span class="w"></span><span class="nt">name</span><span class="p">:</span><span class="w"> </span><span class="l">default</span>
-				</span></span><span class="line"><span class="cl">
-				</span></span><span class="line"><span class="cl"><span class="w"></span><span class="nt">steps</span><span class="p">:</span>
-				</span></span><span class="line"><span class="cl"><span class="w"></span>- <span class="nt">name</span><span class="p">:</span><span class="w"> </span><span class="l">test</span>
-				</span></span><span class="line"><span class="cl"><span class="w">	</span><span class="nt">image</span><span class="p">:</span><span class="w"> </span><span class="l">golang:1.13</span>
-				</span></span><span class="line"><span class="cl"><span class="w">	</span><span class="nt">environment</span><span class="p">:</span>
-				</span></span><span class="line"><span class="cl"><span class="w"></span><span class="w">		</span><span class="nt">GOPROXY</span><span class="p">:</span><span class="w"> </span><span class="l">https://goproxy.cn</span>
-				</span></span><span class="line"><span class="cl"><span class="w">	</span><span class="nt">commands</span><span class="p">:</span>
-				</span></span><span class="line"><span class="cl"><span class="w"></span><span class="w">	</span>- <span class="l">go get -u</span>
-				</span></span><span class="line"><span class="cl"><span class="w">	</span>- <span class="l">go build -v</span>
-				</span></span><span class="line"><span class="cl"><span class="w">	</span>- <span class="l">go test -v -race -coverprofile=coverage.txt -covermode=atomic</span></span></span>
+b=''
+{space}
+c=2
+			`), "{space}", "    "),
+			want: lines(`
+<span class="n">def</span><span class="p">:</span>\n
+    <span class="n">a</span><span class="o">=</span><span class="mi">1</span>\n
+\n
+<span class="n">b</span><span class="o">=</span><span class="sa"></span><span class="s1">&#39;</span><span class="s1">&#39;</span>\n
+    \n
+<span class="n">c</span><span class="o">=</span><span class="mi">2</span>`,
+			),
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			out, err := File(tt.name, "", []byte(tt.code))
+			assert.NoError(t, err)
+			expected := strings.Join(tt.want, "\n")
+			actual := strings.Join(out, "\n")
+			assert.Equal(t, strings.Count(actual, "<span"), strings.Count(actual, "</span>"))
+			assert.EqualValues(t, expected, actual)
+		})
+	}
+}
+
+func TestPlainText(t *testing.T) {
+	tests := []struct {
+		name string
+		code string
+		want []string
+	}{
+		{
+			name: "empty.py",
+			code: "",
+			want: lines(""),
+		},
+		{
+			name: "tags.py",
+			code: "<>",
+			want: lines("&lt;&gt;"),
+		},
+		{
+			name: "eol-no.py",
+			code: "a=1",
+			want: lines(`a=1`),
+		},
+		{
+			name: "eol-newline1.py",
+			code: "a=1\n",
+			want: lines(`a=1\n`),
+		},
+		{
+			name: "eol-newline2.py",
+			code: "a=1\n\n",
+			want: lines(`
+a=1\n
+\n
 			`),
 		},
 		{
-			name:     ".drone.yml - trailing space",
-			numLines: 13,
-			fileName: ".drone.yml",
-			code: strings.Replace(util.Dedent(`
-				kind: pipeline
-				name: default
+			name: "empty-line-with-space.py",
+			code: strings.ReplaceAll(strings.TrimSpace(`
+def:
+    a=1
 
-				steps:
-				- name: test
-					image: golang:1.13
-					environment:
-						GOPROXY: https://goproxy.cn
-					commands:
-					- go get -u
-					- go build -v
-					- go test -v -race -coverprofile=coverage.txt -covermode=atomic
-			`)+"\n", "name: default", "name: default  ", 1),
-			want: util.Dedent(`
-				<span class="line"><span class="cl"><span class="nt">kind</span><span class="p">:</span><span class="w"> </span><span class="l">pipeline</span>
-				</span></span><span class="line"><span class="cl"><span class="w"></span><span class="nt">name</span><span class="p">:</span><span class="w"> </span><span class="l">default  </span>
-				</span></span><span class="line"><span class="cl">
-				</span></span><span class="line"><span class="cl"><span class="w"></span><span class="nt">steps</span><span class="p">:</span>
-				</span></span><span class="line"><span class="cl"><span class="w"></span>- <span class="nt">name</span><span class="p">:</span><span class="w"> </span><span class="l">test</span>
-				</span></span><span class="line"><span class="cl"><span class="w">	</span><span class="nt">image</span><span class="p">:</span><span class="w"> </span><span class="l">golang:1.13</span>
-				</span></span><span class="line"><span class="cl"><span class="w">	</span><span class="nt">environment</span><span class="p">:</span>
-				</span></span><span class="line"><span class="cl"><span class="w"></span><span class="w">		</span><span class="nt">GOPROXY</span><span class="p">:</span><span class="w"> </span><span class="l">https://goproxy.cn</span>
-				</span></span><span class="line"><span class="cl"><span class="w">	</span><span class="nt">commands</span><span class="p">:</span>
-				</span></span><span class="line"><span class="cl"><span class="w"></span><span class="w">	</span>- <span class="l">go get -u</span>
-				</span></span><span class="line"><span class="cl"><span class="w">	</span>- <span class="l">go build -v</span>
-				</span></span><span class="line"><span class="cl"><span class="w">	</span>- <span class="l">go test -v -race -coverprofile=coverage.txt -covermode=atomic</span>
-				</span></span>
-				<span class="w">
-				</span>
-			`),
+b=''
+{space}
+c=2
+			`), "{space}", "    "),
+			want: lines(`
+def:\n
+    a=1\n
+\n
+b=&#39;&#39;\n
+    \n
+c=2`),
 		},
 	}
 
 	for _, tt := range tests {
 		t.Run(tt.name, func(t *testing.T) {
-			got := strings.Join(File(tt.numLines, tt.fileName, "", []byte(tt.code)), "\n")
-			assert.Equal(t, tt.want, got)
+			out := PlainText([]byte(tt.code))
+			expected := strings.Join(tt.want, "\n")
+			actual := strings.Join(out, "\n")
+			assert.EqualValues(t, expected, actual)
 		})
 	}
 }