feat(inputs.tail): Allow handling of quoted strings spanning multiple lines (#11762)

plugins/inputs/tail/README.md

@@ -95,6 +95,15 @@ See the [CONFIGURATION.md][CONFIGURATION.md] for more details.
     ## If true, a message not matching the pattern will constitute a match of the multiline filter and the what will be applied. (vice-versa is also true)
     #invert_match = false
 
+    ## The handling method for quoted text (defaults to 'ignore').
+    ## The following methods are available:
+    ##   ignore  -- do not consider quotation (default)
+    ##   single-quotes -- consider text quoted by single quotes (')
+    ##   double-quotes -- consider text quoted by double quotes (")
+    ##   backticks     -- consider text quoted by backticks (`)
+    ## When handling quotes, escaped quotes (e.g. \") are handled correctly.
+    #quotation = "ignore"
+
     #After the specified timeout, this plugin sends the multiline event even if no new pattern is found to start a new event. The default is 5s.
     #timeout = 5s
 ```
@@ -103,3 +112,7 @@ See the [CONFIGURATION.md][CONFIGURATION.md] for more details.
 
 Metrics are produced according to the `data_format` option.  Additionally a
 tag labeled `path` is added to the metric containing the filename being tailed.
+
+## Example Output
+
+There is no predefined metric format, so output depends on plugin input.

plugins/inputs/tail/multiline.go

@@ -2,6 +2,7 @@ package tail
 
 import (
 	"bytes"
+	"errors"
 	"fmt"
 	"regexp"
 	"strings"
@@ -17,13 +18,16 @@ type Multiline struct {
 	config        *MultilineConfig
 	enabled       bool
 	patternRegexp *regexp.Regexp
+	quote         byte
+	inQuote       bool
 }
 
 type MultilineConfig struct {
-	Pattern        string
+	Pattern        string                  `toml:"pattern"`
 	MatchWhichLine MultilineMatchWhichLine `toml:"match_which_line"`
-	InvertMatch    bool
+	InvertMatch    bool                    `toml:"invert_match"`
-	Timeout        *config.Duration
+	Quotation      string                  `toml:"quotation"`
+	Timeout        *config.Duration        `toml:"timeout"`
 }
 
 const (
@@ -34,25 +38,41 @@ const (
 )
 
 func (m *MultilineConfig) NewMultiline() (*Multiline, error) {
-	enabled := false
 	var r *regexp.Regexp
-	var err error
 
 	if m.Pattern != "" {
-		enabled = true
+		var err error
 		if r, err = regexp.Compile(m.Pattern); err != nil {
 			return nil, err
 		}
+	}
+
+	var quote byte
+	switch m.Quotation {
+	case "", "ignore":
+		m.Quotation = "ignore"
+	case "single-quotes":
+		quote = '\''
+	case "double-quotes":
+		quote = '"'
+	case "backticks":
+		quote = '`'
+	default:
+		return nil, errors.New("invalid 'quotation' setting")
+	}
+
+	enabled := m.Pattern != "" || quote != 0
 	if m.Timeout == nil || time.Duration(*m.Timeout).Nanoseconds() == int64(0) {
 		d := config.Duration(5 * time.Second)
 		m.Timeout = &d
 	}
-	}
 
 	return &Multiline{
 		config:        m,
 		enabled:       enabled,
-		patternRegexp: r}, nil
+		patternRegexp: r,
+		quote:         quote,
+	}, nil
 }
 
 func (m *Multiline) IsEnabled() bool {
@@ -60,10 +80,14 @@ func (m *Multiline) IsEnabled() bool {
 }
 
 func (m *Multiline) ProcessLine(text string, buffer *bytes.Buffer) string {
+	if m.matchQuotation(text) {
+		// Ignore the returned error as we cannot do anything about it anyway
+		_, _ = buffer.WriteString(text + "\n")
+		return ""
+	}
 	if m.matchString(text) {
 		// Ignore the returned error as we cannot do anything about it anyway
-		//nolint:errcheck,revive
+		_, _ = buffer.WriteString(text)
-		buffer.WriteString(text)
 		return ""
 	}
 
@@ -97,9 +121,40 @@ func (m *Multiline) Flush(buffer *bytes.Buffer) string {
 	return text
 }
 
+func (m *Multiline) matchQuotation(text string) bool {
+	if m.config.Quotation == "ignore" {
+		return false
+	}
+	escaped := 0
+	count := 0
+	for i := 0; i < len(text); i++ {
+		if text[i] == '\\' {
+			escaped++
+			continue
+		}
+
+		// If we do encounter a backslash-quote combination, we interpret this
+		// as an escaped-quoted and should not count the quote. However,
+		// backslash-backslash combinations (or any even number of backslashes)
+		// are interpreted as a literal backslash not escaping the quote.
+		if text[i] == m.quote && escaped%2 == 0 {
+			count++
+		}
+		// If we encounter any non-quote, non-backslash character we can
+		// safely reset the escape state.
+		escaped = 0
+	}
+	even := count%2 == 0
+	m.inQuote = (m.inQuote && even) || (!m.inQuote && !even)
+	return m.inQuote
+}
+
 func (m *Multiline) matchString(text string) bool {
+	if m.patternRegexp != nil {
 		return m.patternRegexp.MatchString(text) != m.config.InvertMatch
 	}
+	return false
+}
 
 func (w MultilineMatchWhichLine) String() string {
 	switch w {

plugins/inputs/tail/multiline_test.go

@@ -1,7 +1,11 @@
 package tail
 
 import (
+	"bufio"
 	"bytes"
+	"fmt"
+	"os"
+	"path/filepath"
 	"testing"
 	"time"
 
@@ -234,3 +238,169 @@ func TestMultilineWhat(t *testing.T) {
 	require.Error(t, w7.UnmarshalTOML([]byte(`nope`)))
 	require.Equal(t, MultilineMatchWhichLine(-1), w7)
 }
+
+func TestMultiLineQuoted(t *testing.T) {
+	tests := []struct {
+		name      string
+		quotation string
+		quote     string
+		filename  string
+	}{
+		{
+			name:      "single-quotes",
+			quotation: "single-quotes",
+			quote:     `'`,
+			filename:  "multiline_quoted_single.csv",
+		},
+		{
+			name:      "double-quotes",
+			quotation: "double-quotes",
+			quote:     `"`,
+			filename:  "multiline_quoted_double.csv",
+		},
+		{
+			name:      "backticks",
+			quotation: "backticks",
+			quote:     "`",
+			filename:  "multiline_quoted_backticks.csv",
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			expected := []string{
+				`1660819827410,1,some text without quotes,A`,
+				fmt.Sprintf("1660819827411,1,%ssome text all quoted%s,A", tt.quote, tt.quote),
+				fmt.Sprintf("1660819827412,1,%ssome text all quoted\nbut wrapped%s,A", tt.quote, tt.quote),
+				fmt.Sprintf("1660819827420,2,some text with %squotes%s,B", tt.quote, tt.quote),
+				"1660819827430,3,some text with 'multiple \"quotes\" in `one` line',C",
+				fmt.Sprintf("1660819827440,4,some multiline text with %squotes\n", tt.quote) +
+					fmt.Sprintf("spanning \\%smultiple\\%s\n", tt.quote, tt.quote) +
+					fmt.Sprintf("lines%s but do not %send\ndirectly%s,D", tt.quote, tt.quote, tt.quote),
+				fmt.Sprintf("1660819827450,5,all of %sthis%s should %sbasically%s work...,E", tt.quote, tt.quote, tt.quote, tt.quote),
+			}
+
+			c := &MultilineConfig{
+				MatchWhichLine: Next,
+				Quotation:      tt.quotation,
+			}
+			m, err := c.NewMultiline()
+			require.NoError(t, err)
+
+			f, err := os.Open(filepath.Join("testdata", tt.filename))
+			require.NoError(t, err)
+
+			scanner := bufio.NewScanner(f)
+
+			var buffer bytes.Buffer
+			var result []string
+			for scanner.Scan() {
+				line := scanner.Text()
+
+				text := m.ProcessLine(line, &buffer)
+				if text == "" {
+					continue
+				}
+				result = append(result, text)
+			}
+
+			require.EqualValues(t, expected, result)
+		})
+	}
+}
+
+func TestMultiLineQuotedError(t *testing.T) {
+	tests := []struct {
+		name      string
+		filename  string
+		quotation string
+		quote     string
+		expected  []string
+	}{
+		{
+			name:      "messed up quoting",
+			filename:  "multiline_quoted_messed_up.csv",
+			quotation: "single-quotes",
+			quote:     `'`,
+			expected: []string{
+				"1660819827410,1,some text without quotes,A",
+				"1660819827411,1,'some text all quoted,A\n1660819827412,1,'some text all quoted",
+				"but wrapped,A"},
+		},
+		{
+			name:      "missing closing quote",
+			filename:  "multiline_quoted_missing_close.csv",
+			quotation: "single-quotes",
+			quote:     `'`,
+			expected:  nil,
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			c := &MultilineConfig{
+				MatchWhichLine: Next,
+				Quotation:      tt.quotation,
+			}
+			m, err := c.NewMultiline()
+			require.NoError(t, err)
+
+			f, err := os.Open(filepath.Join("testdata", tt.filename))
+			require.NoError(t, err)
+
+			scanner := bufio.NewScanner(f)
+
+			var buffer bytes.Buffer
+			var result []string
+			for scanner.Scan() {
+				line := scanner.Text()
+
+				text := m.ProcessLine(line, &buffer)
+				if text == "" {
+					continue
+				}
+				result = append(result, text)
+			}
+			require.EqualValues(t, tt.expected, result)
+		})
+	}
+}
+
+func TestMultiLineQuotedAndPattern(t *testing.T) {
+	c := &MultilineConfig{
+		Pattern:        "=>$",
+		MatchWhichLine: Next,
+		Quotation:      "double-quotes",
+	}
+	m, err := c.NewMultiline()
+	require.NoError(t, err, "Configuration was OK.")
+	var buffer bytes.Buffer
+
+	text := m.ProcessLine("1=>", &buffer)
+	require.Empty(t, text)
+	require.NotZero(t, buffer.Len())
+
+	text = m.ProcessLine("2=>", &buffer)
+	require.Empty(t, text)
+	require.NotZero(t, buffer.Len())
+
+	text = m.ProcessLine(`"a quoted`, &buffer)
+	require.Empty(t, text)
+	require.NotZero(t, buffer.Len())
+
+	text = m.ProcessLine(`multiline string"=>`, &buffer)
+	require.Empty(t, text)
+	require.NotZero(t, buffer.Len())
+
+	text = m.ProcessLine("3=>", &buffer)
+	require.Empty(t, text)
+	require.NotZero(t, buffer.Len())
+
+	text = m.ProcessLine("4", &buffer)
+	require.Equal(t, "1=>2=>\"a quoted\nmultiline string\"=>3=>4", text)
+	require.Zero(t, buffer.Len())
+
+	text = m.ProcessLine("5", &buffer)
+	require.Equal(t, "5", text)
+	require.Zero(t, buffer.Len())
+}

plugins/inputs/tail/sample.conf

@@ -62,5 +62,14 @@
     ## If true, a message not matching the pattern will constitute a match of the multiline filter and the what will be applied. (vice-versa is also true)
     #invert_match = false
 
+    ## The handling method for quoted text (defaults to 'ignore').
+    ## The following methods are available:
+    ##   ignore  -- do not consider quotation (default)
+    ##   single-quotes -- consider text quoted by single quotes (')
+    ##   double-quotes -- consider text quoted by double quotes (")
+    ##   backticks     -- consider text quoted by backticks (`)
+    ## When handling quotes, escaped quotes (e.g. \") are handled correctly.
+    #quotation = "ignore"
+
     #After the specified timeout, this plugin sends the multiline event even if no new pattern is found to start a new event. The default is 5s.
     #timeout = 5s

plugins/inputs/tail/testdata/multiline_quoted_backticks.csv

@@ -0,0 +1,12 @@
+1660819827410,1,some text without quotes,A
+1660819827411,1,`some text all quoted`,A
+1660819827412,1,`some text all quoted
+but wrapped`,A
+1660819827420,2,some text with `quotes`,B
+1660819827430,3,some text with 'multiple "quotes" in `one` line',C
+1660819827440,4,some multiline text with `quotes
+spanning \`multiple\`
+lines` but do not `end
+directly`,D
+1660819827450,5,all of `this` should `basically` work...,E
+

plugins/inputs/tail/testdata/multiline_quoted_double.csv

@@ -0,0 +1,12 @@
+1660819827410,1,some text without quotes,A
+1660819827411,1,"some text all quoted",A
+1660819827412,1,"some text all quoted
+but wrapped",A
+1660819827420,2,some text with "quotes",B
+1660819827430,3,some text with 'multiple "quotes" in `one` line',C
+1660819827440,4,some multiline text with "quotes
+spanning \"multiple\"
+lines" but do not "end
+directly",D
+1660819827450,5,all of "this" should "basically" work...,E
+

plugins/inputs/tail/testdata/multiline_quoted_messed_up.csv

@@ -0,0 +1,4 @@
+1660819827410,1,some text without quotes,A
+1660819827411,1,'some text all quoted,A
+1660819827412,1,'some text all quoted
+but wrapped,A

plugins/inputs/tail/testdata/multiline_quoted_missing_close.csv

@@ -0,0 +1,2 @@
+1660819827411,2,'some text all quoted,B
+1660819827410,1,some text without quotes,A

plugins/inputs/tail/testdata/multiline_quoted_single.csv

@@ -0,0 +1,12 @@
+1660819827410,1,some text without quotes,A
+1660819827411,1,'some text all quoted',A
+1660819827412,1,'some text all quoted
+but wrapped',A
+1660819827420,2,some text with 'quotes',B
+1660819827430,3,some text with 'multiple "quotes" in `one` line',C
+1660819827440,4,some multiline text with 'quotes
+spanning \'multiple\'
+lines' but do not 'end
+directly',D
+1660819827450,5,all of 'this' should 'basically' work...,E
+