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

oxidize_pdf/parser/
lexer.rs
