oxidize_pdf/parser/
lexer.rs

1//! PDF Lexer
2//!
3//! Tokenizes PDF syntax according to ISO 32000-1 Section 7.2
4
5use super::{ParseError, ParseResult};
6use std::io::Read;
7
8/// PDF Token types
9#[derive(Debug, Clone, PartialEq)]
10pub enum Token {
11    /// Boolean: true or false
12    Boolean(bool),
13
14    /// Integer number
15    Integer(i64),
16
17    /// Real number
18    Real(f64),
19
20    /// String (literal or hexadecimal)
21    String(Vec<u8>),
22
23    /// Name object (e.g., /Type)
24    Name(String),
25
26    /// Left square bracket [
27    ArrayStart,
28
29    /// Right square bracket ]
30    ArrayEnd,
31
32    /// Dictionary start <<
33    DictStart,
34
35    /// Dictionary end >>
36    DictEnd,
37
38    /// Stream keyword
39    Stream,
40
41    /// Endstream keyword
42    EndStream,
43
44    /// Obj keyword
45    Obj,
46
47    /// Endobj keyword
48    EndObj,
49
50    /// StartXRef keyword
51    StartXRef,
52
53    /// Reference (e.g., 1 0 R)
54    Reference(u32, u16),
55
56    /// Null object
57    Null,
58
59    /// Comment (usually ignored)
60    Comment(String),
61
62    /// End of file
63    Eof,
64}
65
66/// PDF Lexer for tokenizing PDF content
67pub struct Lexer<R: Read> {
68    reader: std::io::BufReader<R>,
69    #[allow(dead_code)]
70    buffer: Vec<u8>,
71    position: usize,
72    peek_buffer: Option<u8>,
73    token_buffer: Vec<Token>,
74}
75
76impl<R: Read> Lexer<R> {
77    /// Create a new lexer from a reader
78    pub fn new(reader: R) -> Self {
79        Self {
80            reader: std::io::BufReader::new(reader),
81            buffer: Vec::with_capacity(1024),
82            position: 0,
83            peek_buffer: None,
84            token_buffer: Vec::new(),
85        }
86    }
87
88    /// Get the next token
89    pub fn next_token(&mut self) -> ParseResult<Token> {
90        // Check if we have a pushed-back token
91        if let Some(token) = self.token_buffer.pop() {
92            return Ok(token);
93        }
94
95        self.skip_whitespace()?;
96
97        let ch = match self.peek_char()? {
98            Some(ch) => ch,
99            None => return Ok(Token::Eof),
100        };
101
102        match ch {
103            b'%' => self.read_comment(),
104            b'/' => self.read_name(),
105            b'(' => self.read_literal_string(),
106            b'<' => self.read_angle_bracket(),
107            b'>' => {
108                self.consume_char()?;
109                if self.peek_char()? == Some(b'>') {
110                    self.consume_char()?;
111                    Ok(Token::DictEnd)
112                } else {
113                    Err(ParseError::SyntaxError {
114                        position: self.position,
115                        message: "Expected '>' after '>'".to_string(),
116                    })
117                }
118            }
119            b'[' => {
120                self.consume_char()?;
121                Ok(Token::ArrayStart)
122            }
123            b']' => {
124                self.consume_char()?;
125                Ok(Token::ArrayEnd)
126            }
127            b't' | b'f' => self.read_boolean(),
128            b'n' => self.read_null(),
129            b'+' | b'-' | b'0'..=b'9' | b'.' => self.read_number(),
130            b'R' => {
131                // R could be a keyword (for references)
132                self.consume_char()?;
133                Ok(Token::Name("R".to_string()))
134            }
135            _ if ch.is_ascii_alphabetic() => self.read_keyword(),
136            _ => Err(ParseError::SyntaxError {
137                position: self.position,
138                message: format!("Unexpected character: {}", ch as char),
139            }),
140        }
141    }
142
143    /// Peek at the next character without consuming it
144    fn peek_char(&mut self) -> ParseResult<Option<u8>> {
145        if let Some(ch) = self.peek_buffer {
146            return Ok(Some(ch));
147        }
148
149        let mut buf = [0u8; 1];
150        match self.reader.read_exact(&mut buf) {
151            Ok(_) => {
152                self.peek_buffer = Some(buf[0]);
153                Ok(Some(buf[0]))
154            }
155            Err(e) if e.kind() == std::io::ErrorKind::UnexpectedEof => Ok(None),
156            Err(e) => Err(e.into()),
157        }
158    }
159
160    /// Consume the next character
161    fn consume_char(&mut self) -> ParseResult<Option<u8>> {
162        let ch = self.peek_char()?;
163        if ch.is_some() {
164            self.peek_buffer = None;
165            self.position += 1;
166        }
167        Ok(ch)
168    }
169
170    /// Skip whitespace and return the number of bytes skipped
171    pub(crate) fn skip_whitespace(&mut self) -> ParseResult<usize> {
172        let mut count = 0;
173        while let Some(ch) = self.peek_char()? {
174            if ch.is_ascii_whitespace() {
175                self.consume_char()?;
176                count += 1;
177            } else {
178                break;
179            }
180        }
181        Ok(count)
182    }
183
184    /// Read a comment (from % to end of line)
185    fn read_comment(&mut self) -> ParseResult<Token> {
186        self.consume_char()?; // consume '%'
187        let mut comment = String::new();
188
189        while let Some(ch) = self.peek_char()? {
190            if ch == b'\n' || ch == b'\r' {
191                break;
192            }
193            self.consume_char()?;
194            comment.push(ch as char);
195        }
196
197        Ok(Token::Comment(comment))
198    }
199
200    /// Read a name object (e.g., /Type)
201    fn read_name(&mut self) -> ParseResult<Token> {
202        self.consume_char()?; // consume '/'
203        let mut name = String::new();
204
205        while let Some(ch) = self.peek_char()? {
206            if ch.is_ascii_whitespace()
207                || matches!(ch, b'/' | b'<' | b'>' | b'[' | b']' | b'(' | b')' | b'%')
208            {
209                break;
210            }
211            self.consume_char()?;
212
213            // Handle hex codes in names (e.g., /A#20B means /A B)
214            if ch == b'#' {
215                let hex1 = self
216                    .consume_char()?
217                    .ok_or_else(|| ParseError::SyntaxError {
218                        position: self.position,
219                        message: "Incomplete hex code in name".to_string(),
220                    })?;
221                let hex2 = self
222                    .consume_char()?
223                    .ok_or_else(|| ParseError::SyntaxError {
224                        position: self.position,
225                        message: "Incomplete hex code in name".to_string(),
226                    })?;
227
228                let value = u8::from_str_radix(&format!("{}{}", hex1 as char, hex2 as char), 16)
229                    .map_err(|_| ParseError::SyntaxError {
230                        position: self.position,
231                        message: "Invalid hex code in name".to_string(),
232                    })?;
233
234                name.push(value as char);
235            } else {
236                name.push(ch as char);
237            }
238        }
239
240        Ok(Token::Name(name))
241    }
242
243    /// Read a literal string (parentheses)
244    fn read_literal_string(&mut self) -> ParseResult<Token> {
245        self.consume_char()?; // consume '('
246        let mut string = Vec::new();
247        let mut paren_depth = 1;
248        let mut escape = false;
249
250        while paren_depth > 0 {
251            let ch = self
252                .consume_char()?
253                .ok_or_else(|| ParseError::SyntaxError {
254                    position: self.position,
255                    message: "Unterminated string".to_string(),
256                })?;
257
258            if escape {
259                let escaped = match ch {
260                    b'n' => b'\n',
261                    b'r' => b'\r',
262                    b't' => b'\t',
263                    b'b' => b'\x08',
264                    b'f' => b'\x0C',
265                    b'(' => b'(',
266                    b')' => b')',
267                    b'\\' => b'\\',
268                    b'0'..=b'7' => {
269                        // Octal escape sequence
270                        let mut value = ch - b'0';
271                        for _ in 0..2 {
272                            if let Some(next) = self.peek_char()? {
273                                if matches!(next, b'0'..=b'7') {
274                                    self.consume_char()?;
275                                    value = value * 8 + (next - b'0');
276                                } else {
277                                    break;
278                                }
279                            }
280                        }
281                        value
282                    }
283                    _ => ch, // Unknown escape, use literal
284                };
285                string.push(escaped);
286                escape = false;
287            } else {
288                match ch {
289                    b'\\' => escape = true,
290                    b'(' => {
291                        string.push(ch);
292                        paren_depth += 1;
293                    }
294                    b')' => {
295                        paren_depth -= 1;
296                        if paren_depth > 0 {
297                            string.push(ch);
298                        }
299                    }
300                    _ => string.push(ch),
301                }
302            }
303        }
304
305        Ok(Token::String(string))
306    }
307
308    /// Read angle bracket tokens (hex strings or dict markers)
309    fn read_angle_bracket(&mut self) -> ParseResult<Token> {
310        self.consume_char()?; // consume '<'
311
312        if self.peek_char()? == Some(b'<') {
313            self.consume_char()?;
314            Ok(Token::DictStart)
315        } else {
316            // Hex string
317            let mut hex_chars = String::new();
318            let mut found_end = false;
319
320            while let Some(ch) = self.peek_char()? {
321                if ch == b'>' {
322                    self.consume_char()?;
323                    found_end = true;
324                    break;
325                }
326                self.consume_char()?;
327                if ch.is_ascii_hexdigit() {
328                    hex_chars.push(ch as char);
329                } else if !ch.is_ascii_whitespace() {
330                    return Err(ParseError::SyntaxError {
331                        position: self.position,
332                        message: "Invalid character in hex string".to_string(),
333                    });
334                }
335            }
336
337            if !found_end {
338                return Err(ParseError::SyntaxError {
339                    position: self.position,
340                    message: "Unterminated hex string".to_string(),
341                });
342            }
343
344            // Pad with 0 if odd number of digits
345            if hex_chars.len() % 2 != 0 {
346                hex_chars.push('0');
347            }
348
349            // Convert hex to bytes
350            let mut bytes = Vec::new();
351            for chunk in hex_chars.as_bytes().chunks(2) {
352                let hex_str = std::str::from_utf8(chunk).unwrap();
353                let byte =
354                    u8::from_str_radix(hex_str, 16).map_err(|_| ParseError::SyntaxError {
355                        position: self.position,
356                        message: "Invalid hex string".to_string(),
357                    })?;
358                bytes.push(byte);
359            }
360
361            Ok(Token::String(bytes))
362        }
363    }
364
365    /// Read boolean (true/false)
366    fn read_boolean(&mut self) -> ParseResult<Token> {
367        let word = self.read_word()?;
368        match word.as_str() {
369            "true" => Ok(Token::Boolean(true)),
370            "false" => Ok(Token::Boolean(false)),
371            _ => {
372                // Not a boolean, might be a keyword
373                self.process_keyword(word)
374            }
375        }
376    }
377
378    /// Read null
379    fn read_null(&mut self) -> ParseResult<Token> {
380        let word = self.read_word()?;
381        if word == "null" {
382            Ok(Token::Null)
383        } else {
384            // Not null, might be a keyword
385            self.process_keyword(word)
386        }
387    }
388
389    /// Read a number (integer or real)
390    fn read_number(&mut self) -> ParseResult<Token> {
391        let mut number_str = String::new();
392        let mut has_dot = false;
393
394        // Handle sign - consume it first
395        if let Some(ch) = self.peek_char()? {
396            if ch == b'+' || ch == b'-' {
397                self.consume_char()?;
398                number_str.push(ch as char);
399
400                // After sign, we must have at least one digit
401                if let Some(next) = self.peek_char()? {
402                    if !next.is_ascii_digit() && next != b'.' {
403                        return Err(ParseError::SyntaxError {
404                            position: self.position,
405                            message: "Expected digit after sign".to_string(),
406                        });
407                    }
408                }
409            }
410        }
411
412        // Read digits and decimal point
413        while let Some(ch) = self.peek_char()? {
414            match ch {
415                b'0'..=b'9' => {
416                    self.consume_char()?;
417                    number_str.push(ch as char);
418                }
419                b'.' if !has_dot => {
420                    self.consume_char()?;
421                    number_str.push(ch as char);
422                    has_dot = true;
423                }
424                _ => break,
425            }
426        }
427
428        // Handle scientific notation (e/E)
429        if let Some(ch) = self.peek_char()? {
430            if ch == b'e' || ch == b'E' {
431                self.consume_char()?;
432                number_str.push(ch as char);
433                
434                // Check for optional sign after e/E
435                if let Some(sign_ch) = self.peek_char()? {
436                    if sign_ch == b'+' || sign_ch == b'-' {
437                        self.consume_char()?;
438                        number_str.push(sign_ch as char);
439                    }
440                }
441                
442                // Read exponent digits
443                while let Some(digit_ch) = self.peek_char()? {
444                    if digit_ch.is_ascii_digit() {
445                        self.consume_char()?;
446                        number_str.push(digit_ch as char);
447                    } else {
448                        break;
449                    }
450                }
451                
452                // Scientific notation always results in a real number
453                has_dot = true;
454            }
455        }
456
457        // Don't try to parse references here - let the parser handle it
458        // References are just "num num R" and can be handled at a higher level
459
460        // Parse as number
461        if has_dot {
462            let value = number_str
463                .parse::<f64>()
464                .map_err(|_| ParseError::SyntaxError {
465                    position: self.position,
466                    message: format!("Invalid real number: '{number_str}'"),
467                })?;
468            Ok(Token::Real(value))
469        } else {
470            let value = number_str
471                .parse::<i64>()
472                .map_err(|_| ParseError::SyntaxError {
473                    position: self.position,
474                    message: format!("Invalid integer: '{number_str}'"),
475                })?;
476            Ok(Token::Integer(value))
477        }
478    }
479
480    /// Read a keyword
481    fn read_keyword(&mut self) -> ParseResult<Token> {
482        let word = self.read_word()?;
483        self.process_keyword(word)
484    }
485
486    /// Process a word as a keyword
487    fn process_keyword(&self, word: String) -> ParseResult<Token> {
488        match word.as_str() {
489            "stream" => Ok(Token::Stream),
490            "endstream" => Ok(Token::EndStream),
491            "obj" => Ok(Token::Obj),
492            "endobj" => Ok(Token::EndObj),
493            "startxref" => Ok(Token::StartXRef),
494            _ => Err(ParseError::SyntaxError {
495                position: self.position,
496                message: format!("Unknown keyword: {word}"),
497            }),
498        }
499    }
500
501    /// Read a word (sequence of non-delimiter characters)
502    fn read_word(&mut self) -> ParseResult<String> {
503        let mut word = String::new();
504
505        while let Some(ch) = self.peek_char()? {
506            if ch.is_ascii_whitespace()
507                || matches!(ch, b'/' | b'<' | b'>' | b'[' | b']' | b'(' | b')' | b'%')
508            {
509                break;
510            }
511            self.consume_char()?;
512            word.push(ch as char);
513        }
514
515        Ok(word)
516    }
517
518    /// Read a sequence of digits
519    #[allow(dead_code)]
520    fn read_digits(&mut self) -> ParseResult<String> {
521        let mut digits = String::new();
522
523        while let Some(ch) = self.peek_char()? {
524            if ch.is_ascii_digit() {
525                self.consume_char()?;
526                digits.push(ch as char);
527            } else {
528                break;
529            }
530        }
531
532        Ok(digits)
533    }
534
535    /// Read a newline sequence (CR, LF, or CRLF)
536    pub fn read_newline(&mut self) -> ParseResult<()> {
537        match self.peek_char()? {
538            Some(b'\r') => {
539                self.consume_char()?;
540                // Check for CRLF
541                if self.peek_char()? == Some(b'\n') {
542                    self.consume_char()?;
543                }
544                Ok(())
545            }
546            Some(b'\n') => {
547                self.consume_char()?;
548                Ok(())
549            }
550            _ => Err(ParseError::SyntaxError {
551                position: self.position,
552                message: "Expected newline".to_string(),
553            }),
554        }
555    }
556
557    /// Read exactly n bytes
558    pub fn read_bytes(&mut self, n: usize) -> ParseResult<Vec<u8>> {
559        let mut bytes = vec![0u8; n];
560        self.reader.read_exact(&mut bytes)?;
561        self.position += n;
562        Ok(bytes)
563    }
564
565    /// Read until a specific byte sequence is found
566    pub fn read_until_sequence(&mut self, sequence: &[u8]) -> ParseResult<Vec<u8>> {
567        let mut result = Vec::new();
568        let mut match_pos = 0;
569
570        while let Some(ch) = self.consume_char()? {
571            result.push(ch);
572
573            if ch == sequence[match_pos] {
574                match_pos += 1;
575                if match_pos == sequence.len() {
576                    // Found the sequence, remove it from result
577                    result.truncate(result.len() - sequence.len());
578                    break;
579                }
580            } else if ch == sequence[0] {
581                match_pos = 1;
582            } else {
583                match_pos = 0;
584            }
585        }
586
587        if match_pos < sequence.len() {
588            return Err(ParseError::SyntaxError {
589                position: self.position,
590                message: format!("Sequence {sequence:?} not found"),
591            });
592        }
593
594        Ok(result)
595    }
596
597    /// Get current position
598    pub fn position(&self) -> usize {
599        self.position
600    }
601
602    /// Push back a token to be returned by the next call to next_token
603    pub fn push_token(&mut self, token: Token) {
604        self.token_buffer.push(token);
605    }
606}
607
608#[cfg(test)]
609mod tests {
610    use super::*;
611    use std::io::Cursor;
612
613    #[test]
614    fn test_lexer_basic_tokens() {
615        // Test positive and negative numbers
616        let input = b"123 -456 3.14 true false null /Name";
617        let mut lexer = Lexer::new(Cursor::new(input));
618
619        assert_eq!(lexer.next_token().unwrap(), Token::Integer(123));
620        assert_eq!(lexer.next_token().unwrap(), Token::Integer(-456));
621        assert_eq!(lexer.next_token().unwrap(), Token::Real(3.14));
622        assert_eq!(lexer.next_token().unwrap(), Token::Boolean(true));
623        assert_eq!(lexer.next_token().unwrap(), Token::Boolean(false));
624        assert_eq!(lexer.next_token().unwrap(), Token::Null);
625        assert_eq!(lexer.next_token().unwrap(), Token::Name("Name".to_string()));
626        assert_eq!(lexer.next_token().unwrap(), Token::Eof);
627    }
628
629    #[test]
630    fn test_lexer_negative_numbers() {
631        // Test negative numbers without space
632        let input = b"-123 -45.67";
633        let mut lexer = Lexer::new(Cursor::new(input));
634
635        assert_eq!(lexer.next_token().unwrap(), Token::Integer(-123));
636        assert_eq!(lexer.next_token().unwrap(), Token::Real(-45.67));
637    }
638
639    #[test]
640    fn test_lexer_strings() {
641        let input = b"(Hello World) <48656C6C6F>";
642        let mut lexer = Lexer::new(Cursor::new(input));
643
644        assert_eq!(
645            lexer.next_token().unwrap(),
646            Token::String(b"Hello World".to_vec())
647        );
648        assert_eq!(
649            lexer.next_token().unwrap(),
650            Token::String(b"Hello".to_vec())
651        );
652    }
653
654    #[test]
655    fn test_lexer_dictionaries() {
656        let input = b"<< /Type /Page >>";
657        let mut lexer = Lexer::new(Cursor::new(input));
658
659        assert_eq!(lexer.next_token().unwrap(), Token::DictStart);
660        assert_eq!(lexer.next_token().unwrap(), Token::Name("Type".to_string()));
661        assert_eq!(lexer.next_token().unwrap(), Token::Name("Page".to_string()));
662        assert_eq!(lexer.next_token().unwrap(), Token::DictEnd);
663    }
664
665    #[test]
666    fn test_lexer_arrays() {
667        let input = b"[1 2 3]";
668        let mut lexer = Lexer::new(Cursor::new(input));
669
670        assert_eq!(lexer.next_token().unwrap(), Token::ArrayStart);
671        assert_eq!(lexer.next_token().unwrap(), Token::Integer(1));
672        assert_eq!(lexer.next_token().unwrap(), Token::Integer(2));
673        assert_eq!(lexer.next_token().unwrap(), Token::Integer(3));
674        assert_eq!(lexer.next_token().unwrap(), Token::ArrayEnd);
675    }
676
677    #[test]
678    fn test_lexer_references() {
679        let input = b"1 0 R 25 1 R";
680        let mut lexer = Lexer::new(Cursor::new(input));
681
682        // Now references are parsed as separate tokens
683        assert_eq!(lexer.next_token().unwrap(), Token::Integer(1));
684        assert_eq!(lexer.next_token().unwrap(), Token::Integer(0));
685        // 'R' should be parsed as a keyword or name
686        match lexer.next_token().unwrap() {
687            Token::Name(s) if s == "R" => {} // Could be a name
688            other => panic!("Expected R token, got {other:?}"),
689        }
690
691        assert_eq!(lexer.next_token().unwrap(), Token::Integer(25));
692        assert_eq!(lexer.next_token().unwrap(), Token::Integer(1));
693        match lexer.next_token().unwrap() {
694            Token::Name(s) if s == "R" => {} // Could be a name
695            other => panic!("Expected R token, got {other:?}"),
696        }
697    }
698
699    #[test]
700    fn test_lexer_comments() {
701        let input = b"%PDF-1.7\n123";
702        let mut lexer = Lexer::new(Cursor::new(input));
703
704        assert_eq!(
705            lexer.next_token().unwrap(),
706            Token::Comment("PDF-1.7".to_string())
707        );
708        assert_eq!(lexer.next_token().unwrap(), Token::Integer(123));
709    }
710
711    // Comprehensive tests for Lexer
712    mod comprehensive_tests {
713        use super::*;
714        use std::io::Cursor;
715
716        #[test]
717        fn test_token_debug_trait() {
718            let token = Token::Integer(42);
719            let debug_str = format!("{:?}", token);
720            assert!(debug_str.contains("Integer"));
721            assert!(debug_str.contains("42"));
722        }
723
724        #[test]
725        fn test_token_clone() {
726            let token = Token::String(b"hello".to_vec());
727            let cloned = token.clone();
728            assert_eq!(token, cloned);
729        }
730
731        #[test]
732        fn test_token_equality() {
733            assert_eq!(Token::Integer(42), Token::Integer(42));
734            assert_ne!(Token::Integer(42), Token::Integer(43));
735            assert_eq!(Token::Boolean(true), Token::Boolean(true));
736            assert_ne!(Token::Boolean(true), Token::Boolean(false));
737            assert_eq!(Token::Null, Token::Null);
738            assert_ne!(Token::Null, Token::Integer(0));
739        }
740
741        #[test]
742        fn test_lexer_empty_input() {
743            let input = b"";
744            let mut lexer = Lexer::new(Cursor::new(input));
745            assert_eq!(lexer.next_token().unwrap(), Token::Eof);
746        }
747
748        #[test]
749        fn test_lexer_whitespace_only() {
750            let input = b"   \t\n\r  ";
751            let mut lexer = Lexer::new(Cursor::new(input));
752            assert_eq!(lexer.next_token().unwrap(), Token::Eof);
753        }
754
755        #[test]
756        fn test_lexer_integer_edge_cases() {
757            let input = b"0 +123 -0 9876543210";
758            let mut lexer = Lexer::new(Cursor::new(input));
759
760            assert_eq!(lexer.next_token().unwrap(), Token::Integer(0));
761            assert_eq!(lexer.next_token().unwrap(), Token::Integer(123));
762            assert_eq!(lexer.next_token().unwrap(), Token::Integer(0));
763            assert_eq!(lexer.next_token().unwrap(), Token::Integer(9876543210));
764        }
765
766        #[test]
767        fn test_lexer_real_edge_cases() {
768            let input = b"0.0 +3.14 -2.71828 .5 5. 123.456789";
769            let mut lexer = Lexer::new(Cursor::new(input));
770
771            assert_eq!(lexer.next_token().unwrap(), Token::Real(0.0));
772            assert_eq!(lexer.next_token().unwrap(), Token::Real(3.14));
773            assert_eq!(lexer.next_token().unwrap(), Token::Real(-2.71828));
774            assert_eq!(lexer.next_token().unwrap(), Token::Real(0.5));
775            assert_eq!(lexer.next_token().unwrap(), Token::Real(5.0));
776            assert_eq!(lexer.next_token().unwrap(), Token::Real(123.456789));
777        }
778
779        #[test]
780        fn test_lexer_scientific_notation() {
781            let input = b"1.23e10 -4.56E-5 1e0 2E+3";
782            let mut lexer = Lexer::new(Cursor::new(input));
783
784            assert_eq!(lexer.next_token().unwrap(), Token::Real(1.23e10));
785            assert_eq!(lexer.next_token().unwrap(), Token::Real(-4.56e-5));
786            assert_eq!(lexer.next_token().unwrap(), Token::Real(1e0));
787            assert_eq!(lexer.next_token().unwrap(), Token::Real(2e3));
788        }
789
790        #[test]
791        fn test_lexer_string_literal_escapes() {
792            let input = b"(Hello\\nWorld) (Tab\\tChar) (Quote\\\"Mark) (Backslash\\\\)";
793            let mut lexer = Lexer::new(Cursor::new(input));
794
795            assert_eq!(lexer.next_token().unwrap(), Token::String(b"Hello\nWorld".to_vec()));
796            assert_eq!(lexer.next_token().unwrap(), Token::String(b"Tab\tChar".to_vec()));
797            assert_eq!(lexer.next_token().unwrap(), Token::String(b"Quote\"Mark".to_vec()));
798            assert_eq!(lexer.next_token().unwrap(), Token::String(b"Backslash\\".to_vec()));
799        }
800
801        #[test]
802        fn test_lexer_string_literal_nested_parens() {
803            let input = b"(Nested (parentheses) work)";
804            let mut lexer = Lexer::new(Cursor::new(input));
805
806            assert_eq!(lexer.next_token().unwrap(), Token::String(b"Nested (parentheses) work".to_vec()));
807        }
808
809        #[test]
810        fn test_lexer_string_literal_empty() {
811            let input = b"()";
812            let mut lexer = Lexer::new(Cursor::new(input));
813
814            assert_eq!(lexer.next_token().unwrap(), Token::String(b"".to_vec()));
815        }
816
817        #[test]
818        fn test_lexer_hexadecimal_strings() {
819            let input = b"<48656C6C6F> <20576F726C64> <>";
820            let mut lexer = Lexer::new(Cursor::new(input));
821
822            assert_eq!(lexer.next_token().unwrap(), Token::String(b"Hello".to_vec()));
823            assert_eq!(lexer.next_token().unwrap(), Token::String(b" World".to_vec()));
824            assert_eq!(lexer.next_token().unwrap(), Token::String(b"".to_vec()));
825        }
826
827        #[test]
828        fn test_lexer_hexadecimal_strings_odd_length() {
829            let input = b"<48656C6C6F2> <1> <ABC>";
830            let mut lexer = Lexer::new(Cursor::new(input));
831
832            // Odd length hex strings should pad with 0
833            assert_eq!(lexer.next_token().unwrap(), Token::String(b"Hello ".to_vec()));
834            assert_eq!(lexer.next_token().unwrap(), Token::String(b"\x10".to_vec()));
835            assert_eq!(lexer.next_token().unwrap(), Token::String(b"\xAB\xC0".to_vec()));
836        }
837
838        #[test]
839        fn test_lexer_hexadecimal_strings_whitespace() {
840            let input = b"<48 65 6C 6C 6F>";
841            let mut lexer = Lexer::new(Cursor::new(input));
842
843            assert_eq!(lexer.next_token().unwrap(), Token::String(b"Hello".to_vec()));
844        }
845
846        #[test]
847        fn test_lexer_names() {
848            let input = b"/Type /Page /Root /Kids /Count /MediaBox";
849            let mut lexer = Lexer::new(Cursor::new(input));
850
851            assert_eq!(lexer.next_token().unwrap(), Token::Name("Type".to_string()));
852            assert_eq!(lexer.next_token().unwrap(), Token::Name("Page".to_string()));
853            assert_eq!(lexer.next_token().unwrap(), Token::Name("Root".to_string()));
854            assert_eq!(lexer.next_token().unwrap(), Token::Name("Kids".to_string()));
855            assert_eq!(lexer.next_token().unwrap(), Token::Name("Count".to_string()));
856            assert_eq!(lexer.next_token().unwrap(), Token::Name("MediaBox".to_string()));
857        }
858
859        #[test]
860        fn test_lexer_names_with_special_chars() {
861            let input = b"/Name#20with#20spaces /Name#2Fwith#2Fslashes";
862            let mut lexer = Lexer::new(Cursor::new(input));
863
864            assert_eq!(lexer.next_token().unwrap(), Token::Name("Name with spaces".to_string()));
865            assert_eq!(lexer.next_token().unwrap(), Token::Name("Name/with/slashes".to_string()));
866        }
867
868        #[test]
869        fn test_lexer_names_edge_cases() {
870            let input = b"/ /A /123 /true /false /null";
871            let mut lexer = Lexer::new(Cursor::new(input));
872
873            assert_eq!(lexer.next_token().unwrap(), Token::Name("".to_string()));
874            assert_eq!(lexer.next_token().unwrap(), Token::Name("A".to_string()));
875            assert_eq!(lexer.next_token().unwrap(), Token::Name("123".to_string()));
876            assert_eq!(lexer.next_token().unwrap(), Token::Name("true".to_string()));
877            assert_eq!(lexer.next_token().unwrap(), Token::Name("false".to_string()));
878            assert_eq!(lexer.next_token().unwrap(), Token::Name("null".to_string()));
879        }
880
881        #[test]
882        fn test_lexer_nested_dictionaries() {
883            let input = b"<< /Type /Page /Resources << /Font << /F1 123 0 R >> >> >>";
884            let mut lexer = Lexer::new(Cursor::new(input));
885
886            assert_eq!(lexer.next_token().unwrap(), Token::DictStart);
887            assert_eq!(lexer.next_token().unwrap(), Token::Name("Type".to_string()));
888            assert_eq!(lexer.next_token().unwrap(), Token::Name("Page".to_string()));
889            assert_eq!(lexer.next_token().unwrap(), Token::Name("Resources".to_string()));
890            assert_eq!(lexer.next_token().unwrap(), Token::DictStart);
891            assert_eq!(lexer.next_token().unwrap(), Token::Name("Font".to_string()));
892            assert_eq!(lexer.next_token().unwrap(), Token::DictStart);
893            assert_eq!(lexer.next_token().unwrap(), Token::Name("F1".to_string()));
894            assert_eq!(lexer.next_token().unwrap(), Token::Integer(123));
895            assert_eq!(lexer.next_token().unwrap(), Token::Integer(0));
896            assert_eq!(lexer.next_token().unwrap(), Token::Name("R".to_string()));
897            assert_eq!(lexer.next_token().unwrap(), Token::DictEnd);
898            assert_eq!(lexer.next_token().unwrap(), Token::DictEnd);
899            assert_eq!(lexer.next_token().unwrap(), Token::DictEnd);
900        }
901
902        #[test]
903        fn test_lexer_nested_arrays() {
904            let input = b"[[1 2] [3 4] [5 [6 7]]]";
905            let mut lexer = Lexer::new(Cursor::new(input));
906
907            assert_eq!(lexer.next_token().unwrap(), Token::ArrayStart);
908            assert_eq!(lexer.next_token().unwrap(), Token::ArrayStart);
909            assert_eq!(lexer.next_token().unwrap(), Token::Integer(1));
910            assert_eq!(lexer.next_token().unwrap(), Token::Integer(2));
911            assert_eq!(lexer.next_token().unwrap(), Token::ArrayEnd);
912            assert_eq!(lexer.next_token().unwrap(), Token::ArrayStart);
913            assert_eq!(lexer.next_token().unwrap(), Token::Integer(3));
914            assert_eq!(lexer.next_token().unwrap(), Token::Integer(4));
915            assert_eq!(lexer.next_token().unwrap(), Token::ArrayEnd);
916            assert_eq!(lexer.next_token().unwrap(), Token::ArrayStart);
917            assert_eq!(lexer.next_token().unwrap(), Token::Integer(5));
918            assert_eq!(lexer.next_token().unwrap(), Token::ArrayStart);
919            assert_eq!(lexer.next_token().unwrap(), Token::Integer(6));
920            assert_eq!(lexer.next_token().unwrap(), Token::Integer(7));
921            assert_eq!(lexer.next_token().unwrap(), Token::ArrayEnd);
922            assert_eq!(lexer.next_token().unwrap(), Token::ArrayEnd);
923            assert_eq!(lexer.next_token().unwrap(), Token::ArrayEnd);
924        }
925
926        #[test]
927        fn test_lexer_mixed_content() {
928            let input = b"<< /Type /Page /MediaBox [0 0 612 792] /Resources << /Font << /F1 << /Type /Font /Subtype /Type1 >> >> >> >>";
929            let mut lexer = Lexer::new(Cursor::new(input));
930
931            // Just test that we can parse this without errors
932            let mut tokens = Vec::new();
933            loop {
934                match lexer.next_token().unwrap() {
935                    Token::Eof => break,
936                    token => tokens.push(token),
937                }
938            }
939            assert!(tokens.len() > 10);
940        }
941
942        #[test]
943        fn test_lexer_keywords() {
944            let input = b"obj endobj stream endstream startxref";
945            let mut lexer = Lexer::new(Cursor::new(input));
946
947            assert_eq!(lexer.next_token().unwrap(), Token::Obj);
948            assert_eq!(lexer.next_token().unwrap(), Token::EndObj);
949            assert_eq!(lexer.next_token().unwrap(), Token::Stream);
950            assert_eq!(lexer.next_token().unwrap(), Token::EndStream);
951            assert_eq!(lexer.next_token().unwrap(), Token::StartXRef);
952        }
953
954        #[test]
955        fn test_lexer_multiple_comments() {
956            let input = b"%First comment\n%Second comment\n123";
957            let mut lexer = Lexer::new(Cursor::new(input));
958
959            assert_eq!(lexer.next_token().unwrap(), Token::Comment("First comment".to_string()));
960            assert_eq!(lexer.next_token().unwrap(), Token::Comment("Second comment".to_string()));
961            assert_eq!(lexer.next_token().unwrap(), Token::Integer(123));
962        }
963
964        #[test]
965        fn test_lexer_comment_without_newline() {
966            let input = b"%Comment at end";
967            let mut lexer = Lexer::new(Cursor::new(input));
968
969            assert_eq!(lexer.next_token().unwrap(), Token::Comment("Comment at end".to_string()));
970            assert_eq!(lexer.next_token().unwrap(), Token::Eof);
971        }
972
973        #[test]
974        fn test_lexer_special_characters_in_streams() {
975            let input = b"<< /Length 5 >> stream\nHello endstream";
976            let mut lexer = Lexer::new(Cursor::new(input));
977
978            assert_eq!(lexer.next_token().unwrap(), Token::DictStart);
979            assert_eq!(lexer.next_token().unwrap(), Token::Name("Length".to_string()));
980            assert_eq!(lexer.next_token().unwrap(), Token::Integer(5));
981            assert_eq!(lexer.next_token().unwrap(), Token::DictEnd);
982            assert_eq!(lexer.next_token().unwrap(), Token::Stream);
983            // The actual stream content would be handled by a higher-level parser
984        }
985
986        #[test]
987        fn test_lexer_push_token() {
988            let input = b"123 456";
989            let mut lexer = Lexer::new(Cursor::new(input));
990
991            let token1 = lexer.next_token().unwrap();
992            assert_eq!(token1, Token::Integer(123));
993
994            let token2 = lexer.next_token().unwrap();
995            assert_eq!(token2, Token::Integer(456));
996
997            // Push token2 back
998            lexer.push_token(token2.clone());
999
1000            // Should get token2 again
1001            let token3 = lexer.next_token().unwrap();
1002            assert_eq!(token3, token2);
1003
1004            // Should get EOF
1005            let token4 = lexer.next_token().unwrap();
1006            assert_eq!(token4, Token::Eof);
1007        }
1008
1009        #[test]
1010        fn test_lexer_push_multiple_tokens() {
1011            let input = b"123";
1012            let mut lexer = Lexer::new(Cursor::new(input));
1013
1014            let original_token = lexer.next_token().unwrap();
1015            assert_eq!(original_token, Token::Integer(123));
1016
1017            // Push multiple tokens
1018            lexer.push_token(Token::Boolean(true));
1019            lexer.push_token(Token::Boolean(false));
1020            lexer.push_token(Token::Null);
1021
1022            // Should get them back in reverse order (stack behavior)
1023            assert_eq!(lexer.next_token().unwrap(), Token::Null);
1024            assert_eq!(lexer.next_token().unwrap(), Token::Boolean(false));
1025            assert_eq!(lexer.next_token().unwrap(), Token::Boolean(true));
1026            assert_eq!(lexer.next_token().unwrap(), Token::Eof);
1027        }
1028
1029        #[test]
1030        fn test_lexer_read_newline() {
1031            let input = b"123\n456\r\n789";
1032            let mut lexer = Lexer::new(Cursor::new(input));
1033
1034            // Read first digits
1035            let digits1 = lexer.read_digits().unwrap();
1036            assert_eq!(digits1, "123");
1037            assert!(lexer.read_newline().is_ok());
1038            
1039            // Read second digits
1040            let digits2 = lexer.read_digits().unwrap();
1041            assert_eq!(digits2, "456");
1042            assert!(lexer.read_newline().is_ok());
1043            
1044            // Read final digits
1045            let digits3 = lexer.read_digits().unwrap();
1046            assert_eq!(digits3, "789");
1047        }
1048
1049        #[test]
1050        fn test_lexer_read_bytes() {
1051            let input = b"Hello World";
1052            let mut lexer = Lexer::new(Cursor::new(input));
1053
1054            let bytes = lexer.read_bytes(5).unwrap();
1055            assert_eq!(bytes, b"Hello");
1056
1057            let bytes = lexer.read_bytes(6).unwrap();
1058            assert_eq!(bytes, b" World");
1059        }
1060
1061        #[test]
1062        fn test_lexer_read_until_sequence() {
1063            let input = b"Hello endstream World";
1064            let mut lexer = Lexer::new(Cursor::new(input));
1065
1066            let result = lexer.read_until_sequence(b"endstream").unwrap();
1067            assert_eq!(result, b"Hello ");
1068
1069            // Continue reading after the sequence
1070            let rest = lexer.read_digits().unwrap();
1071            assert_eq!(rest, "");  // read_digits only reads digits, " World" has no digits
1072        }
1073
1074        #[test]
1075        fn test_lexer_read_until_sequence_not_found() {
1076            let input = b"Hello World";
1077            let mut lexer = Lexer::new(Cursor::new(input));
1078
1079            let result = lexer.read_until_sequence(b"notfound");
1080            assert!(result.is_err());
1081        }
1082
1083        #[test]
1084        fn test_lexer_position_tracking() {
1085            let input = b"123 456";
1086            let mut lexer = Lexer::new(Cursor::new(input));
1087
1088            let initial_pos = lexer.position();
1089            assert_eq!(initial_pos, 0);
1090
1091            lexer.next_token().unwrap(); // "123"
1092            let pos_after_first = lexer.position();
1093            assert!(pos_after_first > initial_pos);
1094
1095            lexer.next_token().unwrap(); // "456"
1096            let pos_after_second = lexer.position();
1097            assert!(pos_after_second > pos_after_first);
1098        }
1099
1100        #[test]
1101        fn test_lexer_large_numbers() {
1102            let input = b"2147483647 -2147483648 9223372036854775807 -9223372036854775808";
1103            let mut lexer = Lexer::new(Cursor::new(input));
1104
1105            assert_eq!(lexer.next_token().unwrap(), Token::Integer(2147483647));
1106            assert_eq!(lexer.next_token().unwrap(), Token::Integer(-2147483648));
1107            assert_eq!(lexer.next_token().unwrap(), Token::Integer(9223372036854775807));
1108            assert_eq!(lexer.next_token().unwrap(), Token::Integer(-9223372036854775808));
1109        }
1110
1111        #[test]
1112        fn test_lexer_very_long_string() {
1113            let long_str = "A".repeat(1000);
1114            let input = format!("({})", long_str);
1115            let mut lexer = Lexer::new(Cursor::new(input.as_bytes()));
1116
1117            if let Token::String(s) = lexer.next_token().unwrap() {
1118                assert_eq!(s.len(), 1000);
1119                assert_eq!(s, long_str.as_bytes());
1120            } else {
1121                panic!("Expected string token");
1122            }
1123        }
1124
1125        #[test]
1126        fn test_lexer_very_long_name() {
1127            let long_name = "A".repeat(500);
1128            let input = format!("/{}", long_name);
1129            let mut lexer = Lexer::new(Cursor::new(input.as_bytes()));
1130
1131            if let Token::Name(name) = lexer.next_token().unwrap() {
1132                assert_eq!(name.len(), 500);
1133                assert_eq!(name, long_name);
1134            } else {
1135                panic!("Expected name token");
1136            }
1137        }
1138
1139        #[test]
1140        fn test_lexer_error_handling_invalid_hex() {
1141            let input = b"<48656C6C6FG>";
1142            let mut lexer = Lexer::new(Cursor::new(input));
1143
1144            // Should handle invalid hex gracefully
1145            let result = lexer.next_token();
1146            assert!(result.is_ok() || result.is_err()); // Either works or fails gracefully
1147        }
1148
1149        #[test]
1150        fn test_lexer_all_token_types() {
1151            let input = b"true false null 123 -456 3.14 (string) <48656C6C6F> /Name [ ] << >> obj endobj stream endstream startxref % comment\n";
1152            let mut lexer = Lexer::new(Cursor::new(input));
1153
1154            let mut token_types = Vec::new();
1155            loop {
1156                match lexer.next_token().unwrap() {
1157                    Token::Eof => break,
1158                    token => token_types.push(std::mem::discriminant(&token)),
1159                }
1160            }
1161
1162            // Should have multiple different token types
1163            assert!(token_types.len() > 10);
1164        }
1165
1166        #[test]
1167        fn test_lexer_performance() {
1168            let input = "123 456 789 ".repeat(1000);
1169            let mut lexer = Lexer::new(Cursor::new(input.as_bytes()));
1170
1171            let start_time = std::time::Instant::now();
1172            let mut count = 0;
1173            loop {
1174                match lexer.next_token().unwrap() {
1175                    Token::Eof => break,
1176                    _ => count += 1,
1177                }
1178            }
1179            let elapsed = start_time.elapsed();
1180
1181            assert_eq!(count, 3000); // 1000 repetitions * 3 tokens each
1182            assert!(elapsed.as_millis() < 1000); // Should complete within 1 second
1183        }
1184    }
1185}
oxidize_pdf/parser/lexer.rs

oxidize_pdf/parser/
lexer.rs
