fix: ignore blank lines and decode UTF-16 in the runner env files (#1084)

Fixes #496
Fixes #552

Reviewed-on: https://gitea.com/gitea/runner/pulls/1084
Reviewed-by: Zettat123 <39446+zettat123@noreply.gitea.com>

act/container/parse_env_file.go

@@ -13,6 +13,9 @@ import (
 	"strings"
 
 	"gitea.com/gitea/runner/act/common"
+
+	"golang.org/x/text/encoding/unicode"
+	"golang.org/x/text/transform"
 )
 
 func parseEnvFile(e Container, srcPath string, env *map[string]string) common.Executor {
@@ -28,11 +31,19 @@ func parseEnvFile(e Container, srcPath string, env *map[string]string) common.Ex
 		if err != nil && err != io.EOF {
 			return err
 		}
-		s := bufio.NewScanner(reader)
+		// Decode by BOM: Windows PowerShell 5.1 redirection writes UTF-16, and some
+		// tools emit a UTF-8 BOM. Without a BOM the file is read as UTF-8, as before.
+		decoded := transform.NewReader(reader, unicode.BOMOverride(unicode.UTF8.NewDecoder()))
+
+		s := bufio.NewScanner(decoded)
 		// Default 64 KiB max token size is too small for realistic env-file lines; allow up to 16 MiB.
 		s.Buffer(make([]byte, 0, 64*1024), 16*1024*1024)
 		for s.Scan() {
 			line := s.Text()
+			// GitHub's runner ignores blank lines
+			if strings.TrimSpace(line) == "" {
+				continue
+			}
 			singleLineEnv := strings.Index(line, "=")
 			multiLineEnv := strings.Index(line, "<<")
 			if singleLineEnv != -1 && (multiLineEnv == -1 || singleLineEnv < multiLineEnv) {

act/container/parse_env_file_test.go

@@ -13,6 +13,8 @@ import (
 
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
+	"golang.org/x/text/encoding"
+	"golang.org/x/text/encoding/unicode"
 )
 
 func newTestHostEnv(t *testing.T) (*HostEnvironment, string) {
@@ -64,6 +66,63 @@ func TestParseEnvFileLineExceedsBufferReportsScannerError(t *testing.T) {
 	assert.Contains(t, err.Error(), "reading env file")
 }
 
+// Regression test: a blank line used to fail the job at "Complete Job", after
+// every step had already been recorded as successful.
+func TestParseEnvFileBlankLines(t *testing.T) {
+	e, envPath := newTestHostEnv(t)
+	require.NoError(t, os.WriteFile(envPath, []byte("\nFOO=bar\n\n   \nBAZ=qux\n\n"), 0o600))
+
+	env := map[string]string{}
+	require.NoError(t, parseEnvFile(e, envPath, &env)(context.Background()))
+	assert.Equal(t, "bar", env["FOO"])
+	assert.Equal(t, "qux", env["BAZ"])
+}
+
+// blank lines inside a heredoc value are content, not separators
+func TestParseEnvFileMultiLineKeepsBlankLines(t *testing.T) {
+	e, envPath := newTestHostEnv(t)
+	require.NoError(t, os.WriteFile(envPath, []byte("FOO<<EOF\nline1\n\nline2\nEOF\n"), 0o600))
+
+	env := map[string]string{}
+	require.NoError(t, parseEnvFile(e, envPath, &env)(context.Background()))
+	assert.Equal(t, "line1\n\nline2", env["FOO"])
+}
+
+func TestParseEnvFileUTF8BOM(t *testing.T) {
+	e, envPath := newTestHostEnv(t)
+	content := append([]byte{0xEF, 0xBB, 0xBF}, []byte("FOO=bar\n")...)
+	require.NoError(t, os.WriteFile(envPath, content, 0o600))
+
+	env := map[string]string{}
+	require.NoError(t, parseEnvFile(e, envPath, &env)(context.Background()))
+	assert.Equal(t, "bar", env["FOO"])
+}
+
+// Windows host mode: PowerShell 5.1 redirection writes UTF-16, which used to be
+// unrecognisable as KEY=VALUE, so the writes were silently ignored.
+func TestParseEnvFileUTF16(t *testing.T) {
+	tests := []struct {
+		name    string
+		encoder *encoding.Encoder
+	}{
+		{"little endian", unicode.UTF16(unicode.LittleEndian, unicode.UseBOM).NewEncoder()},
+		{"big endian", unicode.UTF16(unicode.BigEndian, unicode.UseBOM).NewEncoder()},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			e, envPath := newTestHostEnv(t)
+			content, err := tt.encoder.Bytes([]byte("FOO=bar\r\nMULTI<<EOF\r\nline1\r\nEOF\r\n"))
+			require.NoError(t, err)
+			require.NoError(t, os.WriteFile(envPath, content, 0o600))
+
+			env := map[string]string{}
+			require.NoError(t, parseEnvFile(e, envPath, &env)(context.Background()))
+			assert.Equal(t, "bar", env["FOO"])
+			assert.Equal(t, "line1", env["MULTI"])
+		})
+	}
+}
+
 func TestParseEnvFileMissingDelimiter(t *testing.T) {
 	e, envPath := newTestHostEnv(t)
 	require.NoError(t, os.WriteFile(envPath, []byte("FOO<<EOF\nline1\nline2\n"), 0o600))