import static java.util.Objects.requireNonNull;
import com.google.common.annotations.Beta;
+import com.google.common.escape.Escaper;
+import com.google.common.escape.Escapers;
import javax.xml.XMLConstants;
import org.eclipse.jdt.annotation.NonNull;
import org.eclipse.jdt.annotation.Nullable;
@Beta
public abstract class AbstractStringInstanceIdentifierCodec extends AbstractNamespaceCodec<YangInstanceIdentifier>
implements InstanceIdentifierCodec<String> {
+ // Escaper as per https://www.rfc-editor.org/rfc/rfc7950#section-6.1.3
+ private static final Escaper DQUOT_ESCAPER = Escapers.builder()
+ .addEscape('\n', "\\n")
+ .addEscape('\t', "\\t")
+ .addEscape('"', "\\\"")
+ .addEscape('\\', "\\\\")
+ .build();
+
@Override
protected final String serializeImpl(final YangInstanceIdentifier data) {
final StringBuilder sb = new StringBuilder();
if (arg instanceof NodeIdentifierWithPredicates nip) {
for (var entry : nip.entrySet()) {
- appendQName(sb.append('['), entry.getKey(), lastModule);
- sb.append("='").append(String.valueOf(entry.getValue())).append("']");
+ final var keyName = entry.getKey();
+ appendQName(sb.append('['), keyName, lastModule).append('=');
+ appendValue(sb, keyName.getModule(), entry.getValue()).append(']');
}
} else if (arg instanceof NodeWithValue<?> val) {
- sb.append("[.='").append(val.getValue()).append("']");
+ appendValue(sb.append("[.="), lastModule, val.getValue()).append(']');
}
}
return sb.toString();
}
+ private static StringBuilder appendValue(final StringBuilder sb, final QNameModule currentModule,
+ final Object value) {
+ final var str = String.valueOf(value);
+
+ // We have two specifications here: Section 6.1.3 of both RFC6020 and RFC7950:
+ //
+ // RFC6020 Section 6.1.3:
+ // If a string contains any space or tab characters, a semicolon (";"),
+ // braces ("{" or "}"), or comment sequences ("//", "/*", or "*/"), then
+ // it MUST be enclosed within double or single quotes.
+ //
+ // RFC7950 Section 6.1.3:
+ // An unquoted string is any sequence of characters that does not
+ // contain any space, tab, carriage return, or line feed characters, a
+ // single or double quote character, a semicolon (";"), braces ("{" or
+ // "}"), or comment sequences ("//", "/*", or "*/").
+ //
+ // Plus the common part:
+ // A single-quoted string (enclosed within ' ') preserves each character
+ // within the quotes. A single quote character cannot occur in a
+ // single-quoted string, even when preceded by a backslash.
+ //
+ // Unquoted strings are not interesting, as we are embedding the value in a string, not a YANG document, hence
+ // we have to use quotes. Single-quoted case is simpler, as it does not involve any escaping. The only case
+ // where we cannot use it is when the value itself has a single-quote in itself -- then we call back to
+ // double-quoting.
+
+ return str.indexOf('\'') == -1
+ // No escaping needed, use single quotes
+ ? sb.append('\'').append(str).append('\'')
+ // Escaping needed: use double quotes
+ : sb.append('"').append(DQUOT_ESCAPER.escape(str)).append('"');
+ }
+
/**
* Returns DataSchemaContextTree associated with SchemaContext for which
* serialization / deserialization occurs.
private static final CharMatcher IDENTIFIER = IDENTIFIER_FIRST_CHAR.or(CharMatcher.inRange('0', '9'))
.or(CharMatcher.anyOf(".-")).precomputed();
- private static final CharMatcher QUOTE = CharMatcher.anyOf("'\"");
-
private static final char SLASH = '/';
+ private static final char BACKSLASH = '\\';
private static final char COLON = ':';
private static final char DOT = '.';
private static final char EQUALS = '=';
private static final char PRECONDITION_START = '[';
private static final char PRECONDITION_END = ']';
+ private static final char SQUOT = '\'';
+ private static final char DQUOT = '"';
private final List<PathArgument> product = new ArrayList<>();
private final AbstractStringInstanceIdentifierCodec codec;
*/
private void checkValid(final boolean condition, final String errorMsg, final Object... attributes) {
if (!condition) {
- throw new IllegalArgumentException(String.format(
- "Could not parse Instance Identifier '%s'. Offset: %s : Reason: %s", data, offset,
- String.format(errorMsg, attributes)));
+ throw iae(errorMsg, attributes);
}
}
+ private @NonNull IllegalArgumentException iae(final String errorMsg, final Object... attributes) {
+ return new IllegalArgumentException("Could not parse Instance Identifier '%s'. Offset: %s : Reason: %s"
+ .formatted(data, offset, errorMsg.formatted(attributes)));
+ }
+
/**
* Returns following value of quoted literal (without quotes) and sets offset after literal.
*
* @return String literal
*/
private String nextQuotedValue() {
- final char quoteChar = currentChar();
- checkValid(QUOTE.matches(quoteChar), "Value must be qoute escaped with ''' or '\"'.");
+ return switch (currentChar()) {
+ case SQUOT -> nextSingleQuotedValue();
+ case DQUOT -> nextDoubleQuotedValue();
+ default -> throw iae("Value must be quote escaped with ''' or '\"'.");
+ };
+ }
+
+ // Simple: just look for the matching single quote and return substring
+ private String nextSingleQuotedValue() {
skipCurrentChar();
- final int valueStart = offset;
- final int endQoute = data.indexOf(quoteChar, offset);
- final String value = data.substring(valueStart, endQoute);
- offset = endQoute;
+ final int start = offset;
+ final int end = data.indexOf(SQUOT, start);
+ checkValid(end != -1, "Closing single quote not found");
+ offset = end;
skipCurrentChar();
- return value;
+ return data.substring(start, end);
+ }
+
+ // Complicated: we need to potentially un-escape
+ private String nextDoubleQuotedValue() {
+ skipCurrentChar();
+
+ final int maxIndex = data.length() - 1;
+ final var sb = new StringBuilder();
+ while (true) {
+ final int nextStart = offset;
+
+ // Find next double quotes
+ final int nextEnd = data.indexOf(DQUOT, nextStart);
+ checkValid(nextEnd != -1, "Closing double quote not found");
+ offset = nextEnd;
+
+ // Find next backslash
+ final int nextBackslash = data.indexOf(BACKSLASH, nextStart);
+ if (nextBackslash == -1 || nextBackslash > nextEnd) {
+ // No backslash between nextStart and nextEnd -- just copy characters and terminate
+ offset = nextEnd;
+ skipCurrentChar();
+ return sb.append(data, nextStart, nextEnd).toString();
+ }
+
+ // Validate escape completeness and append buffer
+ checkValid(nextBackslash != maxIndex, "Incomplete escape");
+ sb.append(data, nextStart, nextBackslash);
+
+ // Adjust offset before potentially referencing it and
+ offset = nextBackslash;
+ sb.append(unescape(data.charAt(nextBackslash + 1)));
+
+ // Rinse and repeat
+ offset = nextBackslash + 2;
+ }
+ }
+
+ // As per https://www.rfc-editor.org/rfc/rfc7950#section-6.1.3
+ private char unescape(final char escape) {
+ return switch (escape) {
+ case 'n' -> '\n';
+ case 't' -> '\t';
+ case DQUOT -> DQUOT;
+ case BACKSLASH -> BACKSLASH;
+ default -> throw iae("Unrecognized escape");
+ };
}
/**
Code Review / yangtools.git / commitdiff