parser/keyvalue: handle escaped quotes when parsing

Currently, the keyvalue parser does handle escaped quotes and instead
will parse this as independent '\', '"' characters. This results in
unexpected breakages between fields for strings like:
> <key>="...\" ..."
Here, the backslash will be appended to the result pair, while the
(now un-)escaped quotation will result in the pair being terminated early.

Add handling of escaped quotation marks (for both ", ') in the
keyvalue parser along with a testcase to exercise this functionality.

Signed-off-by: Ethan Adams <[email protected]>

Author: jadams41

internal/coreinternal/parseutils/parser.go

@@ -18,6 +18,7 @@ func SplitString(input, delimiter string) ([]string, error) {
 	current := ""
 	delimiterLength := len(delimiter)
 	quoteChar := "" // "" means we are not in quotes
+	escaped := false
 
 	for i := 0; i < len(input); i++ {
 		if quoteChar == "" && i+delimiterLength <= len(input) && input[i:i+delimiterLength] == delimiter { // delimiter
@@ -31,13 +32,19 @@ func SplitString(input, delimiter string) ([]string, error) {
 			continue
 		}
 
-		if quoteChar == "" && (input[i] == '"' || input[i] == '\'') { // start of quote
-			quoteChar = string(input[i])
-			continue
-		}
-		if string(input[i]) == quoteChar { // end of quote
-			quoteChar = ""
-			continue
+		if !escaped { // consider quote termination so long as previous character wasn't backslash
+			if quoteChar == "" && (input[i] == '"' || input[i] == '\'') { // start of quote
+				quoteChar = string(input[i])
+				continue
+			}
+			if string(input[i]) == quoteChar { // end of quote
+				quoteChar = ""
+				continue
+			}
+			// Only if we weren't escaped could the next character result in escaped state
+			escaped = input[i] == '\\' // potentially escaping next character
+		} else {
+			escaped = false
 		}
 
 		current += string(input[i])

pkg/stanza/operator/parser/keyvalue/parser_test.go

@@ -688,6 +688,24 @@ key=value`,
 			true,
 			false,
 		},
+		{
+			"containerd output",
+			func(_ *Config) {},
+			&entry.Entry{
+				Body: `time="2024-11-01T12:38:17.992190505Z" level=warning msg="cleanup warnings time='2024-11-01T12:38:17Z' level=debug msg=\"starting signal loop\" namespace=moby-10000.10000 pid=1608080 runtime=io.containerd.runc.v2" namespace=moby-10000.10000`,
+			},
+			&entry.Entry{
+				Attributes: map[string]any{
+					"time":      "2024-11-01T12:38:17.992190505Z",
+					"level":     "warning",
+					"msg":       `cleanup warnings time='2024-11-01T12:38:17Z' level=debug msg=\"starting signal loop\" namespace=moby-10000.10000 pid=1608080 runtime=io.containerd.runc.v2`,
+					"namespace": "moby-10000.10000",
+				},
+				Body: `time="2024-11-01T12:38:17.992190505Z" level=warning msg="cleanup warnings time='2024-11-01T12:38:17Z' level=debug msg=\"starting signal loop\" namespace=moby-10000.10000 pid=1608080 runtime=io.containerd.runc.v2" namespace=moby-10000.10000`,
+			},
+			false,
+			false,
+		},
 	}
 
 	for _, tc := range cases {