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, ParseOptions, ParseResult, ParseWarning};
6use std::io::{Read, Seek, SeekFrom};
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> {
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    options: ParseOptions,
75    warnings: Vec<ParseWarning>,
76}
77
78impl<R: Read> Lexer<R> {
79    /// Create a new lexer from a reader with default options
80    pub fn new(reader: R) -> Self {
81        Self::new_with_options(reader, ParseOptions::default())
82    }
83
84    /// Create a new lexer from a reader with custom options
85    pub fn new_with_options(reader: R, options: ParseOptions) -> Self {
86        Self {
87            reader: std::io::BufReader::new(reader),
88            buffer: Vec::with_capacity(1024),
89            position: 0,
90            peek_buffer: None,
91            token_buffer: Vec::new(),
92            options,
93            warnings: Vec::new(),
94        }
95    }
96
97    /// Get warnings collected during lexing (if enabled)
98    pub fn warnings(&self) -> &[ParseWarning] {
99        &self.warnings
100    }
101
102    /// Get the next token
103    pub fn next_token(&mut self) -> ParseResult<Token> {
104        // Check if we have a pushed-back token
105        if let Some(token) = self.token_buffer.pop() {
106            return Ok(token);
107        }
108
109        self.skip_whitespace()?;
110
111        let ch = match self.peek_char()? {
112            Some(ch) => ch,
113            None => return Ok(Token::Eof),
114        };
115
116        match ch {
117            b'%' => self.read_comment(),
118            b'/' => self.read_name(),
119            b'(' => self.read_literal_string(),
120            b'<' => self.read_angle_bracket(),
121            b'>' => {
122                self.consume_char()?;
123                if self.peek_char()? == Some(b'>') {
124                    self.consume_char()?;
125                    Ok(Token::DictEnd)
126                } else {
127                    Err(ParseError::SyntaxError {
128                        position: self.position,
129                        message: "Expected '>' after '>'".to_string(),
130                    })
131                }
132            }
133            b'[' => {
134                self.consume_char()?;
135                Ok(Token::ArrayStart)
136            }
137            b']' => {
138                self.consume_char()?;
139                Ok(Token::ArrayEnd)
140            }
141            b't' | b'f' => self.read_boolean(),
142            b'n' => self.read_null(),
143            b'+' | b'-' | b'0'..=b'9' | b'.' => self.read_number(),
144            b'R' => {
145                // R could be a keyword (for references)
146                self.consume_char()?;
147                Ok(Token::Name("R".to_string()))
148            }
149            _ if ch.is_ascii_alphabetic() => self.read_keyword(),
150            b';' => {
151                // Skip semicolons (corrupted PDF recovery)
152                self.consume_char()?;
153                self.next_token() // Recursively get next valid token
154            }
155            _ => {
156                // Check if this is a problematic encoding character
157                if self.is_problematic_encoding_char(ch) {
158                    self.handle_encoding_char_in_token_stream(ch)
159                } else if self.options.lenient_syntax {
160                    // In lenient mode, skip unexpected characters with a warning
161                    if self.options.collect_warnings {
162                        eprintln!(
163                            "Warning: Skipping unexpected character '{}' at position {}",
164                            ch as char, self.position
165                        );
166                    }
167                    self.consume_char()?;
168                    self.next_token() // Continue with next token
169                } else {
170                    Err(ParseError::SyntaxError {
171                        position: self.position,
172                        message: format!("Unexpected character: {}", ch as char),
173                    })
174                }
175            }
176        }
177    }
178
179    /// Peek at the next character without consuming it
180    fn peek_char(&mut self) -> ParseResult<Option<u8>> {
181        if let Some(ch) = self.peek_buffer {
182            return Ok(Some(ch));
183        }
184
185        let mut buf = [0u8; 1];
186        match self.reader.read_exact(&mut buf) {
187            Ok(_) => {
188                self.peek_buffer = Some(buf[0]);
189                Ok(Some(buf[0]))
190            }
191            Err(e) if e.kind() == std::io::ErrorKind::UnexpectedEof => Ok(None),
192            Err(e) => Err(e.into()),
193        }
194    }
195
196    /// Consume the next character
197    fn consume_char(&mut self) -> ParseResult<Option<u8>> {
198        let ch = self.peek_char()?;
199        if ch.is_some() {
200            self.peek_buffer = None;
201            self.position += 1;
202        }
203        Ok(ch)
204    }
205
206    /// Skip whitespace and return the number of bytes skipped
207    pub(crate) fn skip_whitespace(&mut self) -> ParseResult<usize> {
208        let mut count = 0;
209        while let Some(ch) = self.peek_char()? {
210            if ch.is_ascii_whitespace() {
211                self.consume_char()?;
212                count += 1;
213            } else {
214                break;
215            }
216        }
217        Ok(count)
218    }
219
220    /// Read a comment (from % to end of line)
221    fn read_comment(&mut self) -> ParseResult<Token> {
222        self.consume_char()?; // consume '%'
223        let mut comment = String::new();
224
225        while let Some(ch) = self.peek_char()? {
226            if ch == b'\n' || ch == b'\r' {
227                break;
228            }
229            self.consume_char()?;
230            comment.push(ch as char);
231        }
232
233        Ok(Token::Comment(comment))
234    }
235
236    /// Read a name object (e.g., /Type)
237    fn read_name(&mut self) -> ParseResult<Token> {
238        self.consume_char()?; // consume '/'
239        let mut name = String::new();
240
241        while let Some(ch) = self.peek_char()? {
242            if ch.is_ascii_whitespace()
243                || matches!(ch, b'/' | b'<' | b'>' | b'[' | b']' | b'(' | b')' | b'%')
244            {
245                break;
246            }
247            self.consume_char()?;
248
249            // Handle hex codes in names (e.g., /A#20B means /A B)
250            if ch == b'#' {
251                let hex1 = self
252                    .consume_char()?
253                    .ok_or_else(|| ParseError::SyntaxError {
254                        position: self.position,
255                        message: "Incomplete hex code in name".to_string(),
256                    })?;
257                let hex2 = self
258                    .consume_char()?
259                    .ok_or_else(|| ParseError::SyntaxError {
260                        position: self.position,
261                        message: "Incomplete hex code in name".to_string(),
262                    })?;
263
264                let value = u8::from_str_radix(&format!("{}{}", hex1 as char, hex2 as char), 16)
265                    .map_err(|_| ParseError::SyntaxError {
266                        position: self.position,
267                        message: "Invalid hex code in name".to_string(),
268                    })?;
269
270                name.push(value as char);
271            } else {
272                name.push(ch as char);
273            }
274        }
275
276        Ok(Token::Name(name))
277    }
278
279    /// Read a literal string (parentheses)
280    fn read_literal_string(&mut self) -> ParseResult<Token> {
281        self.consume_char()?; // consume '('
282        let mut string = Vec::new();
283        let mut paren_depth = 1;
284        let mut escape = false;
285
286        while paren_depth > 0 {
287            let ch = match self.consume_char()? {
288                Some(c) => c,
289                None => {
290                    if self.options.lenient_syntax {
291                        // In lenient mode, return what we have so far
292                        if self.options.collect_warnings {
293                            self.warnings.push(ParseWarning::SyntaxErrorRecovered {
294                                position: self.position,
295                                expected: "closing parenthesis".to_string(),
296                                found: "EOF".to_string(),
297                                recovery_action: "returned partial string content".to_string(),
298                            });
299                        }
300                        break;
301                    } else {
302                        return Err(ParseError::SyntaxError {
303                            position: self.position,
304                            message: "Unterminated string".to_string(),
305                        });
306                    }
307                }
308            };
309
310            if escape {
311                let escaped = match ch {
312                    b'n' => b'\n',
313                    b'r' => b'\r',
314                    b't' => b'\t',
315                    b'b' => b'\x08',
316                    b'f' => b'\x0C',
317                    b'(' => b'(',
318                    b')' => b')',
319                    b'\\' => b'\\',
320                    b'0'..=b'7' => {
321                        // Octal escape sequence
322                        let mut value = ch - b'0';
323                        for _ in 0..2 {
324                            if let Some(next) = self.peek_char()? {
325                                if matches!(next, b'0'..=b'7') {
326                                    self.consume_char()?;
327                                    value = value * 8 + (next - b'0');
328                                } else {
329                                    break;
330                                }
331                            }
332                        }
333                        value
334                    }
335                    _ => ch, // Unknown escape, use literal
336                };
337                string.push(escaped);
338                escape = false;
339            } else {
340                match ch {
341                    b'\\' => escape = true,
342                    b'(' => {
343                        string.push(ch);
344                        paren_depth += 1;
345                    }
346                    b')' => {
347                        paren_depth -= 1;
348                        if paren_depth > 0 {
349                            string.push(ch);
350                        }
351                    }
352                    _ => string.push(ch),
353                }
354            }
355        }
356
357        // Apply character encoding recovery if enabled
358        let processed_string = if self.options.lenient_encoding {
359            self.process_string_with_encoding_recovery(&string)?
360        } else {
361            string
362        };
363
364        Ok(Token::String(processed_string))
365    }
366
367    /// Read angle bracket tokens (hex strings or dict markers)
368    fn read_angle_bracket(&mut self) -> ParseResult<Token> {
369        self.consume_char()?; // consume '<'
370
371        if self.peek_char()? == Some(b'<') {
372            self.consume_char()?;
373            Ok(Token::DictStart)
374        } else {
375            // Hex string
376            let mut hex_chars = String::new();
377            let mut found_end = false;
378
379            while let Some(ch) = self.peek_char()? {
380                if ch == b'>' {
381                    self.consume_char()?;
382                    found_end = true;
383                    break;
384                }
385                self.consume_char()?;
386                if ch.is_ascii_hexdigit() {
387                    hex_chars.push(ch as char);
388                } else if !ch.is_ascii_whitespace() {
389                    if self.options.lenient_syntax {
390                        // In lenient mode, skip invalid characters
391                        if self.options.collect_warnings {
392                            self.warnings.push(ParseWarning::SyntaxErrorRecovered {
393                                position: self.position,
394                                expected: "hex digit".to_string(),
395                                found: format!("'{}'", ch as char),
396                                recovery_action: "skipped invalid character".to_string(),
397                            });
398                        }
399                    } else {
400                        return Err(ParseError::SyntaxError {
401                            position: self.position,
402                            message: "Invalid character in hex string".to_string(),
403                        });
404                    }
405                }
406            }
407
408            if !found_end {
409                if self.options.lenient_syntax {
410                    // In lenient mode, return what we have so far
411                    if self.options.collect_warnings {
412                        self.warnings.push(ParseWarning::SyntaxErrorRecovered {
413                            position: self.position,
414                            expected: ">".to_string(),
415                            found: "EOF".to_string(),
416                            recovery_action: "returned partial hex string".to_string(),
417                        });
418                    }
419                } else {
420                    return Err(ParseError::SyntaxError {
421                        position: self.position,
422                        message: "Unterminated hex string".to_string(),
423                    });
424                }
425            }
426
427            // Pad with 0 if odd number of digits
428            if hex_chars.len() % 2 != 0 {
429                hex_chars.push('0');
430            }
431
432            // Convert hex to bytes
433            let mut bytes = Vec::new();
434            for chunk in hex_chars.as_bytes().chunks(2) {
435                let hex_str = std::str::from_utf8(chunk).map_err(|_| ParseError::SyntaxError {
436                    position: self.position,
437                    message: "Invalid UTF-8 in hex string".to_string(),
438                })?;
439                let byte =
440                    u8::from_str_radix(hex_str, 16).map_err(|_| ParseError::SyntaxError {
441                        position: self.position,
442                        message: "Invalid hex string".to_string(),
443                    })?;
444                bytes.push(byte);
445            }
446
447            Ok(Token::String(bytes))
448        }
449    }
450
451    /// Read boolean (true/false)
452    fn read_boolean(&mut self) -> ParseResult<Token> {
453        let word = self.read_word()?;
454        match word.as_str() {
455            "true" => Ok(Token::Boolean(true)),
456            "false" => Ok(Token::Boolean(false)),
457            _ => {
458                // Not a boolean, might be a keyword
459                self.process_keyword(word)
460            }
461        }
462    }
463
464    /// Read null
465    fn read_null(&mut self) -> ParseResult<Token> {
466        let word = self.read_word()?;
467        if word == "null" {
468            Ok(Token::Null)
469        } else {
470            // Not null, might be a keyword
471            self.process_keyword(word)
472        }
473    }
474
475    /// Read a number (integer or real)
476    fn read_number(&mut self) -> ParseResult<Token> {
477        let mut number_str = String::new();
478        let mut has_dot = false;
479
480        // Handle sign - consume it first
481        if let Some(ch) = self.peek_char()? {
482            if ch == b'+' || ch == b'-' {
483                self.consume_char()?;
484                number_str.push(ch as char);
485
486                // After sign, we must have at least one digit
487                if let Some(next) = self.peek_char()? {
488                    if !next.is_ascii_digit() && next != b'.' {
489                        return Err(ParseError::SyntaxError {
490                            position: self.position,
491                            message: "Expected digit after sign".to_string(),
492                        });
493                    }
494                }
495            }
496        }
497
498        // Read digits and decimal point
499        while let Some(ch) = self.peek_char()? {
500            match ch {
501                b'0'..=b'9' => {
502                    self.consume_char()?;
503                    number_str.push(ch as char);
504                }
505                b'.' if !has_dot => {
506                    self.consume_char()?;
507                    number_str.push(ch as char);
508                    has_dot = true;
509                }
510                _ => break,
511            }
512        }
513
514        // Handle scientific notation (e/E)
515        if let Some(ch) = self.peek_char()? {
516            if ch == b'e' || ch == b'E' {
517                self.consume_char()?;
518                number_str.push(ch as char);
519
520                // Check for optional sign after e/E
521                if let Some(sign_ch) = self.peek_char()? {
522                    if sign_ch == b'+' || sign_ch == b'-' {
523                        self.consume_char()?;
524                        number_str.push(sign_ch as char);
525                    }
526                }
527
528                // Read exponent digits
529                while let Some(digit_ch) = self.peek_char()? {
530                    if digit_ch.is_ascii_digit() {
531                        self.consume_char()?;
532                        number_str.push(digit_ch as char);
533                    } else {
534                        break;
535                    }
536                }
537
538                // Scientific notation always results in a real number
539                has_dot = true;
540            }
541        }
542
543        // Don't try to parse references here - let the parser handle it
544        // References are just "num num R" and can be handled at a higher level
545
546        // Parse as number
547        if has_dot {
548            let value = number_str
549                .parse::<f64>()
550                .map_err(|_| ParseError::SyntaxError {
551                    position: self.position,
552                    message: format!("Invalid real number: '{number_str}'"),
553                })?;
554            Ok(Token::Real(value))
555        } else {
556            let value = number_str
557                .parse::<i64>()
558                .map_err(|_| ParseError::SyntaxError {
559                    position: self.position,
560                    message: format!("Invalid integer: '{number_str}'"),
561                })?;
562            Ok(Token::Integer(value))
563        }
564    }
565
566    /// Read a keyword
567    fn read_keyword(&mut self) -> ParseResult<Token> {
568        let word = self.read_word()?;
569        self.process_keyword(word)
570    }
571
572    /// Process a word as a keyword
573    fn process_keyword(&self, word: String) -> ParseResult<Token> {
574        match word.as_str() {
575            "stream" => Ok(Token::Stream),
576            "endstream" => Ok(Token::EndStream),
577            "obj" => Ok(Token::Obj),
578            "endobj" => Ok(Token::EndObj),
579            "startxref" => Ok(Token::StartXRef),
580            _ => Err(ParseError::SyntaxError {
581                position: self.position,
582                message: format!("Unknown keyword: {word}"),
583            }),
584        }
585    }
586
587    /// Read a word (sequence of non-delimiter characters)
588    fn read_word(&mut self) -> ParseResult<String> {
589        let mut word = String::new();
590
591        while let Some(ch) = self.peek_char()? {
592            if ch.is_ascii_whitespace()
593                || matches!(ch, b'/' | b'<' | b'>' | b'[' | b']' | b'(' | b')' | b'%')
594            {
595                break;
596            }
597            self.consume_char()?;
598            word.push(ch as char);
599        }
600
601        Ok(word)
602    }
603
604    /// Read a sequence of digits
605    #[allow(dead_code)]
606    fn read_digits(&mut self) -> ParseResult<String> {
607        let mut digits = String::new();
608
609        while let Some(ch) = self.peek_char()? {
610            if ch.is_ascii_digit() {
611                self.consume_char()?;
612                digits.push(ch as char);
613            } else {
614                break;
615            }
616        }
617
618        Ok(digits)
619    }
620
621    /// Read a newline sequence (CR, LF, or CRLF)
622    pub fn read_newline(&mut self) -> ParseResult<()> {
623        match self.peek_char()? {
624            Some(b'\r') => {
625                self.consume_char()?;
626                // Check for CRLF
627                if self.peek_char()? == Some(b'\n') {
628                    self.consume_char()?;
629                }
630                Ok(())
631            }
632            Some(b'\n') => {
633                self.consume_char()?;
634                Ok(())
635            }
636            _ => Err(ParseError::SyntaxError {
637                position: self.position,
638                message: "Expected newline".to_string(),
639            }),
640        }
641    }
642
643    /// Read exactly n bytes
644    /// Peek at the next byte without consuming it
645    pub fn peek_byte(&mut self) -> ParseResult<u8> {
646        match self.peek_char()? {
647            Some(b) => Ok(b),
648            None => Err(ParseError::UnexpectedToken {
649                expected: "byte".to_string(),
650                found: "EOF".to_string(),
651            }),
652        }
653    }
654
655    /// Read a single byte
656    pub fn read_byte(&mut self) -> ParseResult<u8> {
657        match self.consume_char()? {
658            Some(b) => Ok(b),
659            None => Err(ParseError::UnexpectedToken {
660                expected: "byte".to_string(),
661                found: "EOF".to_string(),
662            }),
663        }
664    }
665
666    /// Seek to a specific position
667    pub fn seek(&mut self, pos: u64) -> ParseResult<()>
668    where
669        R: Seek,
670    {
671        self.reader.seek(SeekFrom::Start(pos))?;
672        self.position = pos as usize;
673        Ok(())
674    }
675
676    pub fn read_bytes(&mut self, n: usize) -> ParseResult<Vec<u8>> {
677        let mut bytes = Vec::with_capacity(n);
678
679        // First consume any peeked byte to avoid duplication
680        if self.peek_buffer.is_some() && n > 0 {
681            bytes.push(self.consume_char()?.unwrap());
682        }
683
684        // Read remaining bytes directly
685        let remaining = n - bytes.len();
686        if remaining > 0 {
687            let mut rest = vec![0u8; remaining];
688            self.reader.read_exact(&mut rest)?;
689            self.position += remaining;
690            bytes.extend_from_slice(&rest);
691        }
692
693        Ok(bytes)
694    }
695
696    /// Read until a specific byte sequence is found
697    pub fn read_until_sequence(&mut self, sequence: &[u8]) -> ParseResult<Vec<u8>> {
698        let mut result = Vec::new();
699        let mut match_pos = 0;
700
701        while let Some(ch) = self.consume_char()? {
702            result.push(ch);
703
704            if ch == sequence[match_pos] {
705                match_pos += 1;
706                if match_pos == sequence.len() {
707                    // Found the sequence, remove it from result
708                    result.truncate(result.len() - sequence.len());
709                    break;
710                }
711            } else if ch == sequence[0] {
712                match_pos = 1;
713            } else {
714                match_pos = 0;
715            }
716        }
717
718        if match_pos < sequence.len() {
719            return Err(ParseError::SyntaxError {
720                position: self.position,
721                message: format!("Sequence {sequence:?} not found"),
722            });
723        }
724
725        Ok(result)
726    }
727
728    /// Get current position
729    pub fn position(&self) -> usize {
730        self.position
731    }
732
733    /// Push back a token to be returned by the next call to next_token
734    pub fn push_token(&mut self, token: Token) {
735        self.token_buffer.push(token);
736    }
737
738    /// Expect a specific keyword token
739    pub fn expect_keyword(&mut self, keyword: &str) -> ParseResult<()> {
740        let token = self.next_token()?;
741        match (keyword, &token) {
742            ("endstream", Token::EndStream) => Ok(()),
743            ("stream", Token::Stream) => Ok(()),
744            ("endobj", Token::EndObj) => Ok(()),
745            ("obj", Token::Obj) => Ok(()),
746            ("startxref", Token::StartXRef) => Ok(()),
747            _ => Err(ParseError::UnexpectedToken {
748                expected: format!("keyword '{keyword}'"),
749                found: format!("{token:?}"),
750            }),
751        }
752    }
753
754    /// Find a keyword ahead in the stream without consuming bytes
755    /// Returns the number of bytes until the keyword is found
756    pub fn find_keyword_ahead(
757        &mut self,
758        keyword: &str,
759        max_bytes: usize,
760    ) -> ParseResult<Option<usize>>
761    where
762        R: Seek,
763    {
764        use std::io::{Read, Seek, SeekFrom};
765
766        // Save current position
767        let current_pos = self.reader.stream_position()?;
768        let start_buffer_state = self.peek_buffer;
769
770        let keyword_bytes = keyword.as_bytes();
771        let mut bytes_read = 0;
772        let mut match_buffer = Vec::new();
773
774        // Search for the keyword
775        while bytes_read < max_bytes {
776            let mut byte = [0u8; 1];
777            match self.reader.read_exact(&mut byte) {
778                Ok(_) => {
779                    bytes_read += 1;
780                    match_buffer.push(byte[0]);
781
782                    // Keep only the last keyword.len() bytes in match_buffer
783                    if match_buffer.len() > keyword_bytes.len() {
784                        match_buffer.remove(0);
785                    }
786
787                    // Check if we found the keyword
788                    if match_buffer.len() == keyword_bytes.len() && match_buffer == keyword_bytes {
789                        // Restore position
790                        self.reader.seek(SeekFrom::Start(current_pos))?;
791                        self.peek_buffer = start_buffer_state;
792                        return Ok(Some(bytes_read - keyword_bytes.len()));
793                    }
794                }
795                Err(_) => break, // EOF or error
796            }
797        }
798
799        // Restore position
800        self.reader.seek(SeekFrom::Start(current_pos))?;
801        self.peek_buffer = start_buffer_state;
802        Ok(None)
803    }
804
805    /// Peek ahead n bytes without consuming them
806    pub fn peek_ahead(&mut self, n: usize) -> ParseResult<Vec<u8>>
807    where
808        R: Seek,
809    {
810        use std::io::{Read, Seek, SeekFrom};
811
812        // Save current position
813        let current_pos = self.reader.stream_position()?;
814        let start_buffer_state = self.peek_buffer;
815
816        // Read n bytes
817        let mut bytes = vec![0u8; n];
818        let bytes_read = self.reader.read(&mut bytes)?;
819        bytes.truncate(bytes_read);
820
821        // Restore position
822        self.reader.seek(SeekFrom::Start(current_pos))?;
823        self.peek_buffer = start_buffer_state;
824
825        Ok(bytes)
826    }
827
828    /// Save the current position for later restoration
829    pub fn save_position(&mut self) -> ParseResult<(u64, Option<u8>)>
830    where
831        R: Seek,
832    {
833        use std::io::Seek;
834        let pos = self.reader.stream_position()?;
835        Ok((pos, self.peek_buffer))
836    }
837
838    /// Restore a previously saved position
839    pub fn restore_position(&mut self, saved: (u64, Option<u8>)) -> ParseResult<()>
840    where
841        R: Seek,
842    {
843        use std::io::{Seek, SeekFrom};
844        self.reader.seek(SeekFrom::Start(saved.0))?;
845        self.peek_buffer = saved.1;
846        self.position = saved.0 as usize;
847        Ok(())
848    }
849
850    /// Peek the next token without consuming it
851    pub fn peek_token(&mut self) -> ParseResult<Token>
852    where
853        R: Seek,
854    {
855        let saved_pos = self.save_position()?;
856        let token = self.next_token()?;
857        self.restore_position(saved_pos)?;
858        Ok(token)
859    }
860
861    /// Process string bytes with enhanced character encoding recovery
862    fn process_string_with_encoding_recovery(
863        &mut self,
864        string_bytes: &[u8],
865    ) -> ParseResult<Vec<u8>> {
866        use super::encoding::{CharacterDecoder, EncodingOptions, EncodingType, EnhancedDecoder};
867
868        // First check for common problematic bytes that need special handling
869        let has_problematic_chars = string_bytes.iter().any(|&b| {
870            // Control characters and Latin-1 supplement range that often cause issues
871            (0x80..=0x9F).contains(&b)
872                || b == 0x07
873                || (b <= 0x1F && b != 0x09 && b != 0x0A && b != 0x0D)
874        });
875
876        let decoder = EnhancedDecoder::new();
877
878        // Use more aggressive encoding options if problematic characters detected
879        let encoding_options = if has_problematic_chars {
880            EncodingOptions {
881                lenient_mode: true, // Always use lenient mode for problematic chars
882                preferred_encoding: Some(EncodingType::Windows1252), // Try Windows-1252 first for control chars
883                max_replacements: std::cmp::max(100, string_bytes.len() / 10), // More generous replacement limit
884                log_issues: self.options.collect_warnings,
885            }
886        } else {
887            EncodingOptions {
888                lenient_mode: self.options.lenient_encoding,
889                preferred_encoding: self.options.preferred_encoding,
890                max_replacements: 50,
891                log_issues: self.options.collect_warnings,
892            }
893        };
894
895        match decoder.decode(string_bytes, &encoding_options) {
896            Ok(result) => {
897                // Log warning if replacements were made or problematic chars detected
898                if (result.replacement_count > 0 || has_problematic_chars)
899                    && self.options.collect_warnings
900                {
901                    self.warnings.push(ParseWarning::InvalidEncoding {
902                        position: self.position,
903                        recovered_text: if result.text.len() > 50 {
904                            // Ultra-safe character boundary truncation
905                            let truncate_at = result
906                                .text
907                                .char_indices()
908                                .map(|(i, _)| i)
909                                .nth(47)
910                                .unwrap_or_else(|| {
911                                    // Fallback: find last valid boundary within first 47 bytes
912                                    let limit = result.text.len().min(47);
913                                    let mut pos = limit;
914                                    while pos > 0 && !result.text.is_char_boundary(pos) {
915                                        pos -= 1;
916                                    }
917                                    pos
918                                });
919
920                            // Double-check the boundary before slicing
921                            let safe_text = if truncate_at <= result.text.len()
922                                && result.text.is_char_boundary(truncate_at)
923                            {
924                                result.text[..truncate_at].to_string()
925                            } else {
926                                // Emergency fallback: use chars().take() for absolute safety
927                                result.text.chars().take(47).collect::<String>()
928                            };
929
930                            format!(
931                                "{}... (truncated, {} chars total)",
932                                safe_text,
933                                result.text.chars().count()
934                            )
935                        } else {
936                            result.text.clone()
937                        },
938                        encoding_used: result.detected_encoding,
939                        replacement_count: result.replacement_count,
940                    });
941                }
942
943                // Convert back to bytes
944                Ok(result.text.into_bytes())
945            }
946            Err(encoding_error) => {
947                if self.options.lenient_encoding {
948                    // Enhanced fallback strategy
949                    let fallback_result = self.apply_fallback_encoding_strategy(string_bytes);
950
951                    if self.options.collect_warnings {
952                        self.warnings.push(ParseWarning::InvalidEncoding {
953                            position: self.position,
954                            recovered_text: format!(
955                                "Fallback strategy applied: {} -> {} chars",
956                                string_bytes.len(),
957                                fallback_result.len()
958                            ),
959                            encoding_used: None,
960                            replacement_count: string_bytes.len(),
961                        });
962                    }
963                    Ok(fallback_result)
964                } else {
965                    Err(ParseError::CharacterEncodingError {
966                        position: self.position,
967                        message: format!(
968                            "Failed to decode string with any supported encoding: {encoding_error}"
969                        ),
970                    })
971                }
972            }
973        }
974    }
975
976    /// Apply fallback encoding strategy for severely corrupted strings
977    fn apply_fallback_encoding_strategy(&self, string_bytes: &[u8]) -> Vec<u8> {
978        let mut result = Vec::with_capacity(string_bytes.len());
979
980        for &byte in string_bytes {
981            match byte {
982                // Replace common problematic control characters with safe alternatives
983                0x00..=0x08 | 0x0B | 0x0C | 0x0E..=0x1F => {
984                    result.push(b' '); // Replace control chars with space
985                }
986                0x80..=0x9F => {
987                    // Windows-1252 control character range - try to map to reasonable alternatives
988                    let replacement = match byte {
989                        0x80 => b'E',  // Euro sign -> E
990                        0x81 => b' ',  // Undefined -> space
991                        0x82 => b',',  // Single low-9 quotation mark -> comma
992                        0x83 => b'f',  // Latin small letter f with hook -> f
993                        0x84 => b'"',  // Double low-9 quotation mark -> quote
994                        0x85 => b'.',  // Horizontal ellipsis -> period
995                        0x86 => b'+',  // Dagger -> plus
996                        0x87 => b'+',  // Double dagger -> plus
997                        0x88 => b'^',  // Modifier letter circumflex accent -> caret
998                        0x89 => b'%',  // Per mille sign -> percent
999                        0x8A => b'S',  // Latin capital letter S with caron -> S
1000                        0x8B => b'<',  // Single left-pointing angle quotation mark
1001                        0x8C => b'O',  // Latin capital ligature OE -> O
1002                        0x8D => b' ',  // Undefined -> space
1003                        0x8E => b'Z',  // Latin capital letter Z with caron -> Z
1004                        0x8F => b' ',  // Undefined -> space
1005                        0x90 => b' ',  // Undefined -> space
1006                        0x91 => b'\'', // Left single quotation mark
1007                        0x92 => b'\'', // Right single quotation mark
1008                        0x93 => b'"',  // Left double quotation mark
1009                        0x94 => b'"',  // Right double quotation mark
1010                        0x95 => b'*',  // Bullet -> asterisk
1011                        0x96 => b'-',  // En dash -> hyphen
1012                        0x97 => b'-',  // Em dash -> hyphen
1013                        0x98 => b'~',  // Small tilde
1014                        0x99 => b'T',  // Trade mark sign -> T
1015                        0x9A => b's',  // Latin small letter s with caron -> s
1016                        0x9B => b'>',  // Single right-pointing angle quotation mark
1017                        0x9C => b'o',  // Latin small ligature oe -> o
1018                        0x9D => b' ',  // Undefined -> space
1019                        0x9E => b'z',  // Latin small letter z with caron -> z
1020                        0x9F => b'Y',  // Latin capital letter Y with diaeresis -> Y
1021                        _ => b'?',     // Fallback
1022                    };
1023                    result.push(replacement);
1024                }
1025                _ => {
1026                    result.push(byte); // Keep valid bytes as-is
1027                }
1028            }
1029        }
1030
1031        result
1032    }
1033
1034    /// Check if a character is likely a problematic encoding character
1035    fn is_problematic_encoding_char(&self, ch: u8) -> bool {
1036        // Control characters and Latin-1 supplement range that often indicate encoding issues
1037        (0x80..=0x9F).contains(&ch) ||
1038        ch == 0x07 || // Bell character
1039        (ch <= 0x1F && ch != 0x09 && ch != 0x0A && ch != 0x0D) || // Control chars except tab, LF, CR
1040        // In lenient mode, also handle extended Latin-1 characters that may appear in corrupted streams
1041        (self.options.lenient_syntax && ch >= 0xA0) // Extended Latin-1 range (u8 max is 0xFF)
1042    }
1043
1044    /// Handle problematic encoding characters in the main token stream
1045    fn handle_encoding_char_in_token_stream(&mut self, ch: u8) -> ParseResult<Token> {
1046        if self.options.lenient_encoding {
1047            // Consume the problematic character and continue
1048            self.consume_char()?;
1049
1050            // Log warning about the character recovery
1051            if self.options.collect_warnings {
1052                let replacement_char = match ch {
1053                    0x07 => "bell",
1054                    0x00..=0x1F => "control",
1055                    0x80..=0x9F => "latin1-supplement",
1056                    _ => "unknown",
1057                };
1058
1059                self.warnings.push(ParseWarning::InvalidEncoding {
1060                    position: self.position,
1061                    recovered_text: format!(
1062                        "Skipped problematic {replacement_char} character (0x{ch:02X})"
1063                    ),
1064                    encoding_used: None,
1065                    replacement_count: 1,
1066                });
1067            }
1068
1069            // Skip this character and try to get the next token
1070            self.skip_whitespace()?;
1071            if let Ok(Some(_)) = self.peek_char() {
1072                self.next_token() // Recursively try next token
1073            } else {
1074                Err(ParseError::SyntaxError {
1075                    position: self.position,
1076                    message: "Unexpected end of file after problematic character".to_string(),
1077                })
1078            }
1079        } else {
1080            // In strict mode, generate a more descriptive error
1081            let char_description = match ch {
1082                0x07 => "Bell character (\\u{07})".to_string(),
1083                0x00..=0x1F => format!("Control character (\\u{{{ch:02X}}})"),
1084                0x80..=0x9F => format!("Latin-1 supplement character (\\u{{{ch:02X}}})"),
1085                _ => format!("Problematic character (\\u{{{ch:02X}}})"),
1086            };
1087
1088            Err(ParseError::CharacterEncodingError {
1089                position: self.position,
1090                message: format!(
1091                    "Unexpected character: {char_description} - Consider using lenient parsing mode"
1092                ),
1093            })
1094        }
1095    }
1096}
1097
1098#[cfg(test)]
1099mod tests {
1100    use super::*;
1101    use std::io::Cursor;
1102
1103    #[test]
1104    fn test_lexer_basic_tokens() {
1105        // Test positive and negative numbers
1106        let input = b"123 -456 3.14 true false null /Name";
1107        let mut lexer = Lexer::new(Cursor::new(input));
1108
1109        assert_eq!(lexer.next_token().unwrap(), Token::Integer(123));
1110        assert_eq!(lexer.next_token().unwrap(), Token::Integer(-456));
1111        assert_eq!(lexer.next_token().unwrap(), Token::Real(3.14));
1112        assert_eq!(lexer.next_token().unwrap(), Token::Boolean(true));
1113        assert_eq!(lexer.next_token().unwrap(), Token::Boolean(false));
1114        assert_eq!(lexer.next_token().unwrap(), Token::Null);
1115        assert_eq!(lexer.next_token().unwrap(), Token::Name("Name".to_string()));
1116        assert_eq!(lexer.next_token().unwrap(), Token::Eof);
1117    }
1118
1119    #[test]
1120    fn test_lexer_negative_numbers() {
1121        // Test negative numbers without space
1122        let input = b"-123 -45.67";
1123        let mut lexer = Lexer::new(Cursor::new(input));
1124
1125        assert_eq!(lexer.next_token().unwrap(), Token::Integer(-123));
1126        assert_eq!(lexer.next_token().unwrap(), Token::Real(-45.67));
1127    }
1128
1129    #[test]
1130    fn test_lexer_strings() {
1131        let input = b"(Hello World) <48656C6C6F>";
1132        let mut lexer = Lexer::new(Cursor::new(input));
1133
1134        assert_eq!(
1135            lexer.next_token().unwrap(),
1136            Token::String(b"Hello World".to_vec())
1137        );
1138        assert_eq!(
1139            lexer.next_token().unwrap(),
1140            Token::String(b"Hello".to_vec())
1141        );
1142    }
1143
1144    #[test]
1145    fn test_lexer_dictionaries() {
1146        let input = b"<< /Type /Page >>";
1147        let mut lexer = Lexer::new(Cursor::new(input));
1148
1149        assert_eq!(lexer.next_token().unwrap(), Token::DictStart);
1150        assert_eq!(lexer.next_token().unwrap(), Token::Name("Type".to_string()));
1151        assert_eq!(lexer.next_token().unwrap(), Token::Name("Page".to_string()));
1152        assert_eq!(lexer.next_token().unwrap(), Token::DictEnd);
1153    }
1154
1155    #[test]
1156    fn test_lexer_arrays() {
1157        let input = b"[1 2 3]";
1158        let mut lexer = Lexer::new(Cursor::new(input));
1159
1160        assert_eq!(lexer.next_token().unwrap(), Token::ArrayStart);
1161        assert_eq!(lexer.next_token().unwrap(), Token::Integer(1));
1162        assert_eq!(lexer.next_token().unwrap(), Token::Integer(2));
1163        assert_eq!(lexer.next_token().unwrap(), Token::Integer(3));
1164        assert_eq!(lexer.next_token().unwrap(), Token::ArrayEnd);
1165    }
1166
1167    #[test]
1168    fn test_lexer_references() {
1169        let input = b"1 0 R 25 1 R";
1170        let mut lexer = Lexer::new(Cursor::new(input));
1171
1172        // Now references are parsed as separate tokens
1173        assert_eq!(lexer.next_token().unwrap(), Token::Integer(1));
1174        assert_eq!(lexer.next_token().unwrap(), Token::Integer(0));
1175        // 'R' should be parsed as a keyword or name
1176        match lexer.next_token().unwrap() {
1177            Token::Name(s) if s == "R" => {} // Could be a name
1178            other => panic!("Expected R token, got {other:?}"),
1179        }
1180
1181        assert_eq!(lexer.next_token().unwrap(), Token::Integer(25));
1182        assert_eq!(lexer.next_token().unwrap(), Token::Integer(1));
1183        match lexer.next_token().unwrap() {
1184            Token::Name(s) if s == "R" => {} // Could be a name
1185            other => panic!("Expected R token, got {other:?}"),
1186        }
1187    }
1188
1189    #[test]
1190    fn test_lexer_comments() {
1191        let input = b"%PDF-1.7\n123";
1192        let mut lexer = Lexer::new(Cursor::new(input));
1193
1194        assert_eq!(
1195            lexer.next_token().unwrap(),
1196            Token::Comment("PDF-1.7".to_string())
1197        );
1198        assert_eq!(lexer.next_token().unwrap(), Token::Integer(123));
1199    }
1200
1201    // Comprehensive tests for Lexer
1202    mod comprehensive_tests {
1203        use super::*;
1204        use std::io::Cursor;
1205
1206        #[test]
1207        fn test_token_debug_trait() {
1208            let token = Token::Integer(42);
1209            let debug_str = format!("{token:?}");
1210            assert!(debug_str.contains("Integer"));
1211            assert!(debug_str.contains("42"));
1212        }
1213
1214        #[test]
1215        fn test_token_clone() {
1216            let token = Token::String(b"hello".to_vec());
1217            let cloned = token.clone();
1218            assert_eq!(token, cloned);
1219        }
1220
1221        #[test]
1222        fn test_token_equality() {
1223            assert_eq!(Token::Integer(42), Token::Integer(42));
1224            assert_ne!(Token::Integer(42), Token::Integer(43));
1225            assert_eq!(Token::Boolean(true), Token::Boolean(true));
1226            assert_ne!(Token::Boolean(true), Token::Boolean(false));
1227            assert_eq!(Token::Null, Token::Null);
1228            assert_ne!(Token::Null, Token::Integer(0));
1229        }
1230
1231        #[test]
1232        fn test_lexer_empty_input() {
1233            let input = b"";
1234            let mut lexer = Lexer::new(Cursor::new(input));
1235            assert_eq!(lexer.next_token().unwrap(), Token::Eof);
1236        }
1237
1238        #[test]
1239        fn test_lexer_whitespace_only() {
1240            let input = b"   \t\n\r  ";
1241            let mut lexer = Lexer::new(Cursor::new(input));
1242            assert_eq!(lexer.next_token().unwrap(), Token::Eof);
1243        }
1244
1245        #[test]
1246        fn test_lexer_integer_edge_cases() {
1247            let input = b"0 +123 -0 9876543210";
1248            let mut lexer = Lexer::new(Cursor::new(input));
1249
1250            assert_eq!(lexer.next_token().unwrap(), Token::Integer(0));
1251            assert_eq!(lexer.next_token().unwrap(), Token::Integer(123));
1252            assert_eq!(lexer.next_token().unwrap(), Token::Integer(0));
1253            assert_eq!(lexer.next_token().unwrap(), Token::Integer(9876543210));
1254        }
1255
1256        #[test]
1257        fn test_lexer_real_edge_cases() {
1258            let input = b"0.0 +3.14 -2.71828 .5 5. 123.456789";
1259            let mut lexer = Lexer::new(Cursor::new(input));
1260
1261            assert_eq!(lexer.next_token().unwrap(), Token::Real(0.0));
1262            assert_eq!(lexer.next_token().unwrap(), Token::Real(3.14));
1263            assert_eq!(lexer.next_token().unwrap(), Token::Real(-2.71828));
1264            assert_eq!(lexer.next_token().unwrap(), Token::Real(0.5));
1265            assert_eq!(lexer.next_token().unwrap(), Token::Real(5.0));
1266            assert_eq!(lexer.next_token().unwrap(), Token::Real(123.456789));
1267        }
1268
1269        #[test]
1270        fn test_lexer_scientific_notation() {
1271            let input = b"1.23e10 -4.56E-5 1e0 2E+3";
1272            let mut lexer = Lexer::new(Cursor::new(input));
1273
1274            assert_eq!(lexer.next_token().unwrap(), Token::Real(1.23e10));
1275            assert_eq!(lexer.next_token().unwrap(), Token::Real(-4.56e-5));
1276            assert_eq!(lexer.next_token().unwrap(), Token::Real(1e0));
1277            assert_eq!(lexer.next_token().unwrap(), Token::Real(2e3));
1278        }
1279
1280        #[test]
1281        fn test_lexer_string_literal_escapes() {
1282            let input = b"(Hello\\nWorld) (Tab\\tChar) (Quote\\\"Mark) (Backslash\\\\)";
1283            let mut lexer = Lexer::new(Cursor::new(input));
1284
1285            assert_eq!(
1286                lexer.next_token().unwrap(),
1287                Token::String(b"Hello\nWorld".to_vec())
1288            );
1289            assert_eq!(
1290                lexer.next_token().unwrap(),
1291                Token::String(b"Tab\tChar".to_vec())
1292            );
1293            assert_eq!(
1294                lexer.next_token().unwrap(),
1295                Token::String(b"Quote\"Mark".to_vec())
1296            );
1297            assert_eq!(
1298                lexer.next_token().unwrap(),
1299                Token::String(b"Backslash\\".to_vec())
1300            );
1301        }
1302
1303        #[test]
1304        fn test_lexer_string_literal_nested_parens() {
1305            let input = b"(Nested (parentheses) work)";
1306            let mut lexer = Lexer::new(Cursor::new(input));
1307
1308            assert_eq!(
1309                lexer.next_token().unwrap(),
1310                Token::String(b"Nested (parentheses) work".to_vec())
1311            );
1312        }
1313
1314        #[test]
1315        fn test_lexer_string_literal_empty() {
1316            let input = b"()";
1317            let mut lexer = Lexer::new(Cursor::new(input));
1318
1319            assert_eq!(lexer.next_token().unwrap(), Token::String(b"".to_vec()));
1320        }
1321
1322        #[test]
1323        fn test_lexer_hexadecimal_strings() {
1324            let input = b"<48656C6C6F> <20576F726C64> <>";
1325            let mut lexer = Lexer::new(Cursor::new(input));
1326
1327            assert_eq!(
1328                lexer.next_token().unwrap(),
1329                Token::String(b"Hello".to_vec())
1330            );
1331            assert_eq!(
1332                lexer.next_token().unwrap(),
1333                Token::String(b" World".to_vec())
1334            );
1335            assert_eq!(lexer.next_token().unwrap(), Token::String(b"".to_vec()));
1336        }
1337
1338        #[test]
1339        fn test_lexer_hexadecimal_strings_odd_length() {
1340            let input = b"<48656C6C6F2> <1> <ABC>";
1341            let mut lexer = Lexer::new(Cursor::new(input));
1342
1343            // Odd length hex strings should pad with 0
1344            assert_eq!(
1345                lexer.next_token().unwrap(),
1346                Token::String(b"Hello ".to_vec())
1347            );
1348            assert_eq!(lexer.next_token().unwrap(), Token::String(b"\x10".to_vec()));
1349            assert_eq!(
1350                lexer.next_token().unwrap(),
1351                Token::String(b"\xAB\xC0".to_vec())
1352            );
1353        }
1354
1355        #[test]
1356        fn test_lexer_hexadecimal_strings_whitespace() {
1357            let input = b"<48 65 6C 6C 6F>";
1358            let mut lexer = Lexer::new(Cursor::new(input));
1359
1360            assert_eq!(
1361                lexer.next_token().unwrap(),
1362                Token::String(b"Hello".to_vec())
1363            );
1364        }
1365
1366        #[test]
1367        fn test_lexer_names() {
1368            let input = b"/Type /Page /Root /Kids /Count /MediaBox";
1369            let mut lexer = Lexer::new(Cursor::new(input));
1370
1371            assert_eq!(lexer.next_token().unwrap(), Token::Name("Type".to_string()));
1372            assert_eq!(lexer.next_token().unwrap(), Token::Name("Page".to_string()));
1373            assert_eq!(lexer.next_token().unwrap(), Token::Name("Root".to_string()));
1374            assert_eq!(lexer.next_token().unwrap(), Token::Name("Kids".to_string()));
1375            assert_eq!(
1376                lexer.next_token().unwrap(),
1377                Token::Name("Count".to_string())
1378            );
1379            assert_eq!(
1380                lexer.next_token().unwrap(),
1381                Token::Name("MediaBox".to_string())
1382            );
1383        }
1384
1385        #[test]
1386        fn test_lexer_names_with_special_chars() {
1387            let input = b"/Name#20with#20spaces /Name#2Fwith#2Fslashes";
1388            let mut lexer = Lexer::new(Cursor::new(input));
1389
1390            assert_eq!(
1391                lexer.next_token().unwrap(),
1392                Token::Name("Name with spaces".to_string())
1393            );
1394            assert_eq!(
1395                lexer.next_token().unwrap(),
1396                Token::Name("Name/with/slashes".to_string())
1397            );
1398        }
1399
1400        #[test]
1401        fn test_lexer_names_edge_cases() {
1402            let input = b"/ /A /123 /true /false /null";
1403            let mut lexer = Lexer::new(Cursor::new(input));
1404
1405            assert_eq!(lexer.next_token().unwrap(), Token::Name("".to_string()));
1406            assert_eq!(lexer.next_token().unwrap(), Token::Name("A".to_string()));
1407            assert_eq!(lexer.next_token().unwrap(), Token::Name("123".to_string()));
1408            assert_eq!(lexer.next_token().unwrap(), Token::Name("true".to_string()));
1409            assert_eq!(
1410                lexer.next_token().unwrap(),
1411                Token::Name("false".to_string())
1412            );
1413            assert_eq!(lexer.next_token().unwrap(), Token::Name("null".to_string()));
1414        }
1415
1416        #[test]
1417        fn test_lexer_nested_dictionaries() {
1418            let input = b"<< /Type /Page /Resources << /Font << /F1 123 0 R >> >> >>";
1419            let mut lexer = Lexer::new(Cursor::new(input));
1420
1421            assert_eq!(lexer.next_token().unwrap(), Token::DictStart);
1422            assert_eq!(lexer.next_token().unwrap(), Token::Name("Type".to_string()));
1423            assert_eq!(lexer.next_token().unwrap(), Token::Name("Page".to_string()));
1424            assert_eq!(
1425                lexer.next_token().unwrap(),
1426                Token::Name("Resources".to_string())
1427            );
1428            assert_eq!(lexer.next_token().unwrap(), Token::DictStart);
1429            assert_eq!(lexer.next_token().unwrap(), Token::Name("Font".to_string()));
1430            assert_eq!(lexer.next_token().unwrap(), Token::DictStart);
1431            assert_eq!(lexer.next_token().unwrap(), Token::Name("F1".to_string()));
1432            assert_eq!(lexer.next_token().unwrap(), Token::Integer(123));
1433            assert_eq!(lexer.next_token().unwrap(), Token::Integer(0));
1434            assert_eq!(lexer.next_token().unwrap(), Token::Name("R".to_string()));
1435            assert_eq!(lexer.next_token().unwrap(), Token::DictEnd);
1436            assert_eq!(lexer.next_token().unwrap(), Token::DictEnd);
1437            assert_eq!(lexer.next_token().unwrap(), Token::DictEnd);
1438        }
1439
1440        #[test]
1441        fn test_lexer_nested_arrays() {
1442            let input = b"[[1 2] [3 4] [5 [6 7]]]";
1443            let mut lexer = Lexer::new(Cursor::new(input));
1444
1445            assert_eq!(lexer.next_token().unwrap(), Token::ArrayStart);
1446            assert_eq!(lexer.next_token().unwrap(), Token::ArrayStart);
1447            assert_eq!(lexer.next_token().unwrap(), Token::Integer(1));
1448            assert_eq!(lexer.next_token().unwrap(), Token::Integer(2));
1449            assert_eq!(lexer.next_token().unwrap(), Token::ArrayEnd);
1450            assert_eq!(lexer.next_token().unwrap(), Token::ArrayStart);
1451            assert_eq!(lexer.next_token().unwrap(), Token::Integer(3));
1452            assert_eq!(lexer.next_token().unwrap(), Token::Integer(4));
1453            assert_eq!(lexer.next_token().unwrap(), Token::ArrayEnd);
1454            assert_eq!(lexer.next_token().unwrap(), Token::ArrayStart);
1455            assert_eq!(lexer.next_token().unwrap(), Token::Integer(5));
1456            assert_eq!(lexer.next_token().unwrap(), Token::ArrayStart);
1457            assert_eq!(lexer.next_token().unwrap(), Token::Integer(6));
1458            assert_eq!(lexer.next_token().unwrap(), Token::Integer(7));
1459            assert_eq!(lexer.next_token().unwrap(), Token::ArrayEnd);
1460            assert_eq!(lexer.next_token().unwrap(), Token::ArrayEnd);
1461            assert_eq!(lexer.next_token().unwrap(), Token::ArrayEnd);
1462        }
1463
1464        #[test]
1465        fn test_lexer_mixed_content() {
1466            let input = b"<< /Type /Page /MediaBox [0 0 612 792] /Resources << /Font << /F1 << /Type /Font /Subtype /Type1 >> >> >> >>";
1467            let mut lexer = Lexer::new(Cursor::new(input));
1468
1469            // Just test that we can parse this without errors
1470            let mut tokens = Vec::new();
1471            loop {
1472                match lexer.next_token().unwrap() {
1473                    Token::Eof => break,
1474                    token => tokens.push(token),
1475                }
1476            }
1477            assert!(tokens.len() > 10);
1478        }
1479
1480        #[test]
1481        fn test_lexer_keywords() {
1482            let input = b"obj endobj stream endstream startxref";
1483            let mut lexer = Lexer::new(Cursor::new(input));
1484
1485            assert_eq!(lexer.next_token().unwrap(), Token::Obj);
1486            assert_eq!(lexer.next_token().unwrap(), Token::EndObj);
1487            assert_eq!(lexer.next_token().unwrap(), Token::Stream);
1488            assert_eq!(lexer.next_token().unwrap(), Token::EndStream);
1489            assert_eq!(lexer.next_token().unwrap(), Token::StartXRef);
1490        }
1491
1492        #[test]
1493        fn test_lexer_multiple_comments() {
1494            let input = b"%First comment\n%Second comment\n123";
1495            let mut lexer = Lexer::new(Cursor::new(input));
1496
1497            assert_eq!(
1498                lexer.next_token().unwrap(),
1499                Token::Comment("First comment".to_string())
1500            );
1501            assert_eq!(
1502                lexer.next_token().unwrap(),
1503                Token::Comment("Second comment".to_string())
1504            );
1505            assert_eq!(lexer.next_token().unwrap(), Token::Integer(123));
1506        }
1507
1508        #[test]
1509        fn test_lexer_comment_without_newline() {
1510            let input = b"%Comment at end";
1511            let mut lexer = Lexer::new(Cursor::new(input));
1512
1513            assert_eq!(
1514                lexer.next_token().unwrap(),
1515                Token::Comment("Comment at end".to_string())
1516            );
1517            assert_eq!(lexer.next_token().unwrap(), Token::Eof);
1518        }
1519
1520        #[test]
1521        fn test_lexer_special_characters_in_streams() {
1522            let input = b"<< /Length 5 >> stream\nHello endstream";
1523            let mut lexer = Lexer::new(Cursor::new(input));
1524
1525            assert_eq!(lexer.next_token().unwrap(), Token::DictStart);
1526            assert_eq!(
1527                lexer.next_token().unwrap(),
1528                Token::Name("Length".to_string())
1529            );
1530            assert_eq!(lexer.next_token().unwrap(), Token::Integer(5));
1531            assert_eq!(lexer.next_token().unwrap(), Token::DictEnd);
1532            assert_eq!(lexer.next_token().unwrap(), Token::Stream);
1533            // The actual stream content would be handled by a higher-level parser
1534        }
1535
1536        #[test]
1537        fn test_lexer_push_token() {
1538            let input = b"123 456";
1539            let mut lexer = Lexer::new(Cursor::new(input));
1540
1541            let token1 = lexer.next_token().unwrap();
1542            assert_eq!(token1, Token::Integer(123));
1543
1544            let token2 = lexer.next_token().unwrap();
1545            assert_eq!(token2, Token::Integer(456));
1546
1547            // Push token2 back
1548            lexer.push_token(token2.clone());
1549
1550            // Should get token2 again
1551            let token3 = lexer.next_token().unwrap();
1552            assert_eq!(token3, token2);
1553
1554            // Should get EOF
1555            let token4 = lexer.next_token().unwrap();
1556            assert_eq!(token4, Token::Eof);
1557        }
1558
1559        #[test]
1560        fn test_lexer_push_multiple_tokens() {
1561            let input = b"123";
1562            let mut lexer = Lexer::new(Cursor::new(input));
1563
1564            let original_token = lexer.next_token().unwrap();
1565            assert_eq!(original_token, Token::Integer(123));
1566
1567            // Push multiple tokens
1568            lexer.push_token(Token::Boolean(true));
1569            lexer.push_token(Token::Boolean(false));
1570            lexer.push_token(Token::Null);
1571
1572            // Should get them back in reverse order (stack behavior)
1573            assert_eq!(lexer.next_token().unwrap(), Token::Null);
1574            assert_eq!(lexer.next_token().unwrap(), Token::Boolean(false));
1575            assert_eq!(lexer.next_token().unwrap(), Token::Boolean(true));
1576            assert_eq!(lexer.next_token().unwrap(), Token::Eof);
1577        }
1578
1579        #[test]
1580        fn test_lexer_read_newline() {
1581            let input = b"123\n456\r\n789";
1582            let mut lexer = Lexer::new(Cursor::new(input));
1583
1584            // Read first digits
1585            let digits1 = lexer.read_digits().unwrap();
1586            assert_eq!(digits1, "123");
1587            assert!(lexer.read_newline().is_ok());
1588
1589            // Read second digits
1590            let digits2 = lexer.read_digits().unwrap();
1591            assert_eq!(digits2, "456");
1592            assert!(lexer.read_newline().is_ok());
1593
1594            // Read final digits
1595            let digits3 = lexer.read_digits().unwrap();
1596            assert_eq!(digits3, "789");
1597        }
1598
1599        #[test]
1600        fn test_lexer_read_bytes() {
1601            let input = b"Hello World";
1602            let mut lexer = Lexer::new(Cursor::new(input));
1603
1604            let bytes = lexer.read_bytes(5).unwrap();
1605            assert_eq!(bytes, b"Hello");
1606
1607            let bytes = lexer.read_bytes(6).unwrap();
1608            assert_eq!(bytes, b" World");
1609        }
1610
1611        #[test]
1612        fn test_lexer_read_until_sequence() {
1613            let input = b"Hello endstream World";
1614            let mut lexer = Lexer::new(Cursor::new(input));
1615
1616            let result = lexer.read_until_sequence(b"endstream").unwrap();
1617            assert_eq!(result, b"Hello ");
1618
1619            // Continue reading after the sequence
1620            let rest = lexer.read_digits().unwrap();
1621            assert_eq!(rest, ""); // read_digits only reads digits, " World" has no digits
1622        }
1623
1624        #[test]
1625        fn test_lexer_read_until_sequence_not_found() {
1626            let input = b"Hello World";
1627            let mut lexer = Lexer::new(Cursor::new(input));
1628
1629            let result = lexer.read_until_sequence(b"notfound");
1630            assert!(result.is_err());
1631        }
1632
1633        #[test]
1634        fn test_lexer_position_tracking() {
1635            let input = b"123 456";
1636            let mut lexer = Lexer::new(Cursor::new(input));
1637
1638            let initial_pos = lexer.position();
1639            assert_eq!(initial_pos, 0);
1640
1641            lexer.next_token().unwrap(); // "123"
1642            let pos_after_first = lexer.position();
1643            assert!(pos_after_first > initial_pos);
1644
1645            lexer.next_token().unwrap(); // "456"
1646            let pos_after_second = lexer.position();
1647            assert!(pos_after_second > pos_after_first);
1648        }
1649
1650        #[test]
1651        fn test_lexer_large_numbers() {
1652            let input = b"2147483647 -2147483648 9223372036854775807 -9223372036854775808";
1653            let mut lexer = Lexer::new(Cursor::new(input));
1654
1655            assert_eq!(lexer.next_token().unwrap(), Token::Integer(2147483647));
1656            assert_eq!(lexer.next_token().unwrap(), Token::Integer(-2147483648));
1657            assert_eq!(
1658                lexer.next_token().unwrap(),
1659                Token::Integer(9223372036854775807)
1660            );
1661            assert_eq!(
1662                lexer.next_token().unwrap(),
1663                Token::Integer(-9223372036854775808)
1664            );
1665        }
1666
1667        #[test]
1668        fn test_lexer_very_long_string() {
1669            let long_str = "A".repeat(1000);
1670            let input = format!("({long_str})");
1671            let mut lexer = Lexer::new(Cursor::new(input.as_bytes()));
1672
1673            if let Token::String(s) = lexer.next_token().unwrap() {
1674                assert_eq!(s.len(), 1000);
1675                assert_eq!(s, long_str.as_bytes());
1676            } else {
1677                panic!("Expected string token");
1678            }
1679        }
1680
1681        #[test]
1682        fn test_lexer_very_long_name() {
1683            let long_name = "A".repeat(500);
1684            let input = format!("/{long_name}");
1685            let mut lexer = Lexer::new(Cursor::new(input.as_bytes()));
1686
1687            if let Token::Name(name) = lexer.next_token().unwrap() {
1688                assert_eq!(name.len(), 500);
1689                assert_eq!(name, long_name);
1690            } else {
1691                panic!("Expected name token");
1692            }
1693        }
1694
1695        #[test]
1696        fn test_lexer_error_handling_invalid_hex() {
1697            let input = b"<48656C6C6FG>";
1698            let mut lexer = Lexer::new(Cursor::new(input));
1699
1700            // Should handle invalid hex gracefully
1701            let result = lexer.next_token();
1702            assert!(result.is_ok() || result.is_err()); // Either works or fails gracefully
1703        }
1704
1705        #[test]
1706        fn test_lexer_all_token_types() {
1707            let input = b"true false null 123 -456 3.14 (string) <48656C6C6F> /Name [ ] << >> obj endobj stream endstream startxref % comment\n";
1708            let mut lexer = Lexer::new(Cursor::new(input));
1709
1710            let mut token_types = Vec::new();
1711            loop {
1712                match lexer.next_token().unwrap() {
1713                    Token::Eof => break,
1714                    token => token_types.push(std::mem::discriminant(&token)),
1715                }
1716            }
1717
1718            // Should have multiple different token types
1719            assert!(token_types.len() > 10);
1720        }
1721
1722        #[test]
1723        fn test_lexer_performance() {
1724            let input = "123 456 789 ".repeat(1000);
1725            let mut lexer = Lexer::new(Cursor::new(input.as_bytes()));
1726
1727            let start_time = std::time::Instant::now();
1728            let mut count = 0;
1729            loop {
1730                match lexer.next_token().unwrap() {
1731                    Token::Eof => break,
1732                    _ => count += 1,
1733                }
1734            }
1735            let elapsed = start_time.elapsed();
1736
1737            assert_eq!(count, 3000); // 1000 repetitions * 3 tokens each
1738            assert!(elapsed.as_millis() < 1000); // Should complete within 1 second
1739        }
1740    }
1741
1742    #[test]
1743    fn test_lexer_find_keyword_ahead() {
1744        let input = b"some data here endstream more data";
1745        let mut lexer = Lexer::new(Cursor::new(input));
1746
1747        // Find endstream keyword
1748        let result = lexer.find_keyword_ahead("endstream", 100);
1749        assert!(result.is_ok());
1750        assert_eq!(result.unwrap(), Some(15)); // Position of endstream
1751
1752        // Try to find non-existent keyword
1753        let result2 = lexer.find_keyword_ahead("notfound", 100);
1754        assert!(result2.is_ok());
1755        assert_eq!(result2.unwrap(), None);
1756
1757        // Test with limited search distance
1758        let result3 = lexer.find_keyword_ahead("endstream", 10);
1759        assert!(result3.is_ok());
1760        assert_eq!(result3.unwrap(), None); // Not found within 10 bytes
1761    }
1762
1763    #[test]
1764    fn test_lexer_peek_token() {
1765        let input = b"123 456 /Name";
1766        let mut lexer = Lexer::new(Cursor::new(input));
1767
1768        // Peek first token
1769        let peeked = lexer.peek_token();
1770        assert!(peeked.is_ok());
1771        assert_eq!(peeked.unwrap(), Token::Integer(123));
1772
1773        // Verify peek doesn't consume
1774        let next = lexer.next_token();
1775        assert!(next.is_ok());
1776        assert_eq!(next.unwrap(), Token::Integer(123));
1777
1778        // Peek and consume next tokens
1779        assert_eq!(lexer.peek_token().unwrap(), Token::Integer(456));
1780        assert_eq!(lexer.next_token().unwrap(), Token::Integer(456));
1781
1782        assert_eq!(lexer.peek_token().unwrap(), Token::Name("Name".to_string()));
1783        assert_eq!(lexer.next_token().unwrap(), Token::Name("Name".to_string()));
1784    }
1785
1786    #[test]
1787    fn test_lexer_expect_keyword() {
1788        let input = b"endstream obj endobj";
1789        let mut lexer = Lexer::new(Cursor::new(input));
1790
1791        // Expect correct keyword
1792        assert!(lexer.expect_keyword("endstream").is_ok());
1793
1794        // Expect another correct keyword
1795        assert!(lexer.expect_keyword("obj").is_ok());
1796
1797        // Expect wrong keyword (should fail)
1798        let result = lexer.expect_keyword("stream");
1799        assert!(result.is_err());
1800        match result {
1801            Err(ParseError::UnexpectedToken { expected, found }) => {
1802                assert!(expected.contains("stream"));
1803                assert!(found.contains("EndObj"));
1804            }
1805            _ => panic!("Expected UnexpectedToken error"),
1806        }
1807    }
1808
1809    #[test]
1810    fn test_lexer_save_restore_position() {
1811        let input = b"123 456 789";
1812        let mut lexer = Lexer::new(Cursor::new(input));
1813
1814        // Read first token
1815        assert_eq!(lexer.next_token().unwrap(), Token::Integer(123));
1816
1817        // Save position
1818        let saved = lexer.save_position();
1819        assert!(saved.is_ok());
1820        let saved_pos = saved.unwrap();
1821
1822        // Read more tokens
1823        assert_eq!(lexer.next_token().unwrap(), Token::Integer(456));
1824        assert_eq!(lexer.next_token().unwrap(), Token::Integer(789));
1825
1826        // Restore position
1827        assert!(lexer.restore_position(saved_pos).is_ok());
1828
1829        // Should be back at second token
1830        assert_eq!(lexer.next_token().unwrap(), Token::Integer(456));
1831    }
1832
1833    #[test]
1834    fn test_lexer_character_encoding_recovery() {
1835        // Test string with encoding issues (Windows-1252 bytes)
1836        let input = b"(Caf\x80 \x91Hello\x92)"; // "Café 'Hello'"
1837        let options = ParseOptions::lenient();
1838        let mut lexer = Lexer::new_with_options(Cursor::new(input), options);
1839
1840        match lexer.next_token().unwrap() {
1841            Token::String(bytes) => {
1842                // Should contain the text, potentially with encoding recovery
1843                let text = String::from_utf8_lossy(&bytes);
1844                println!("Recovered text: {text}");
1845                assert!(!text.is_empty()); // Should not be empty
1846            }
1847            other => panic!("Expected String token, got {other:?}"),
1848        }
1849
1850        // Check that warnings were collected
1851        let warnings = lexer.warnings();
1852        if !warnings.is_empty() {
1853            println!("Encoding warnings: {warnings:?}");
1854        }
1855    }
1856}
oxidize_pdf/parser/lexer.rs

oxidize_pdf/parser/
lexer.rs
