1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
use super::super::blocks::block::BlockHeader;
use super::super::blocks::block::BlockType;
use super::super::blocks::literals_section::LiteralsSection;
use super::super::blocks::literals_section::LiteralsSectionType;
use super::super::blocks::sequence_section::SequencesHeader;
use super::literals_section_decoder::{decode_literals, DecompressLiteralsError};
use super::sequence_execution::ExecuteSequencesError;
use super::sequence_section_decoder::decode_sequences;
use super::sequence_section_decoder::DecodeSequenceError;
use crate::blocks::literals_section::LiteralsSectionParseError;
use crate::blocks::sequence_section::SequencesHeaderParseError;
use crate::decoding::scratch::DecoderScratch;
use crate::decoding::sequence_execution::execute_sequences;
use crate::io::{self, Read};

pub struct BlockDecoder {
    header_buffer: [u8; 3],
    internal_state: DecoderState,
}

enum DecoderState {
    ReadyToDecodeNextHeader,
    ReadyToDecodeNextBody,
    #[allow(dead_code)]
    Failed, //TODO put "self.internal_state = DecoderState::Failed;" everywhere an unresolvable error occurs
}

#[derive(Debug, derive_more::Display, derive_more::From)]
#[cfg_attr(feature = "std", derive(derive_more::Error))]
#[non_exhaustive]
pub enum BlockHeaderReadError {
    #[display(fmt = "Error while reading the block header")]
    #[from]
    ReadError(io::Error),
    #[display(fmt = "Reserved block occured. This is considered corruption by the documentation")]
    FoundReservedBlock,
    #[display(fmt = "Error getting block type: {_0}")]
    #[from]
    BlockTypeError(BlockTypeError),
    #[display(fmt = "Error getting block content size: {_0}")]
    #[from]
    BlockSizeError(BlockSizeError),
}

#[derive(Debug, derive_more::Display)]
#[cfg_attr(feature = "std", derive(derive_more::Error))]
#[non_exhaustive]
pub enum BlockTypeError {
    #[display(
        fmt = "Invalid Blocktype number. Is: {num} Should be one of: 0, 1, 2, 3 (3 is reserved though"
    )]
    InvalidBlocktypeNumber { num: u8 },
}

#[derive(Debug, derive_more::Display)]
#[cfg_attr(feature = "std", derive(derive_more::Error))]
#[non_exhaustive]
pub enum BlockSizeError {
    #[display(
        fmt = "Blocksize was bigger than the absolute maximum {ABSOLUTE_MAXIMUM_BLOCK_SIZE} (128kb). Is: {size}"
    )]
    BlockSizeTooLarge { size: u32 },
}

#[derive(Debug, derive_more::Display, derive_more::From)]
#[cfg_attr(feature = "std", derive(derive_more::Error))]
#[non_exhaustive]
pub enum DecompressBlockError {
    #[display(fmt = "Error while reading the block content: {_0}")]
    #[from]
    BlockContentReadError(io::Error),
    #[display(
        fmt = "Malformed section header. Says literals would be this long: {expected_len} but there are only {remaining_bytes} bytes left"
    )]
    MalformedSectionHeader {
        expected_len: usize,
        remaining_bytes: usize,
    },
    #[display(fmt = "{_0:?}")]
    #[from]
    DecompressLiteralsError(DecompressLiteralsError),
    #[display(fmt = "{_0:?}")]
    #[from]
    LiteralsSectionParseError(LiteralsSectionParseError),
    #[display(fmt = "{_0:?}")]
    #[from]
    SequencesHeaderParseError(SequencesHeaderParseError),
    #[display(fmt = "{_0:?}")]
    #[from]
    DecodeSequenceError(DecodeSequenceError),
    #[display(fmt = "{_0:?}")]
    #[from]
    ExecuteSequencesError(ExecuteSequencesError),
}

#[derive(Debug, derive_more::Display, derive_more::From)]
#[cfg_attr(feature = "std", derive(derive_more::Error))]
#[non_exhaustive]
pub enum DecodeBlockContentError {
    #[display(fmt = "Can't decode next block if failed along the way. Results will be nonsense")]
    DecoderStateIsFailed,
    #[display(
        fmt = "Cant decode next block body, while expecting to decode the header of the previous block. Results will be nonsense"
    )]
    ExpectedHeaderOfPreviousBlock,
    #[display(fmt = "Error while reading bytes for {step}: {source}")]
    ReadError { step: BlockType, source: io::Error },
    #[display(fmt = "{_0:?}")]
    #[from]
    DecompressBlockError(DecompressBlockError),
}

pub fn new() -> BlockDecoder {
    BlockDecoder {
        internal_state: DecoderState::ReadyToDecodeNextHeader,
        header_buffer: [0u8; 3],
    }
}

const ABSOLUTE_MAXIMUM_BLOCK_SIZE: u32 = 128 * 1024;

impl BlockDecoder {
    pub fn decode_block_content(
        &mut self,
        header: &BlockHeader,
        workspace: &mut DecoderScratch, //reuse this as often as possible. Not only if the trees are reused but also reuse the allocations when building new trees
        mut source: impl Read,
    ) -> Result<u64, DecodeBlockContentError> {
        match self.internal_state {
            DecoderState::ReadyToDecodeNextBody => { /* Happy :) */ }
            DecoderState::Failed => return Err(DecodeBlockContentError::DecoderStateIsFailed),
            DecoderState::ReadyToDecodeNextHeader => {
                return Err(DecodeBlockContentError::ExpectedHeaderOfPreviousBlock)
            }
        }

        let block_type = header.block_type;
        match block_type {
            BlockType::RLE => {
                const BATCH_SIZE: usize = 512;
                let mut buf = [0u8; BATCH_SIZE];
                let full_reads = header.decompressed_size / BATCH_SIZE as u32;
                let single_read_size = header.decompressed_size % BATCH_SIZE as u32;

                source.read_exact(&mut buf[0..1]).map_err(|err| {
                    DecodeBlockContentError::ReadError {
                        step: block_type,
                        source: err,
                    }
                })?;
                self.internal_state = DecoderState::ReadyToDecodeNextHeader;

                for i in 1..BATCH_SIZE {
                    buf[i] = buf[0];
                }

                for _ in 0..full_reads {
                    workspace.buffer.push(&buf[..]);
                }
                let smaller = &mut buf[..single_read_size as usize];
                workspace.buffer.push(smaller);

                Ok(1)
            }
            BlockType::Raw => {
                const BATCH_SIZE: usize = 128 * 1024;
                let mut buf = [0u8; BATCH_SIZE];
                let full_reads = header.decompressed_size / BATCH_SIZE as u32;
                let single_read_size = header.decompressed_size % BATCH_SIZE as u32;

                for _ in 0..full_reads {
                    source.read_exact(&mut buf[..]).map_err(|err| {
                        DecodeBlockContentError::ReadError {
                            step: block_type,
                            source: err,
                        }
                    })?;
                    workspace.buffer.push(&buf[..]);
                }

                let smaller = &mut buf[..single_read_size as usize];
                source
                    .read_exact(smaller)
                    .map_err(|err| DecodeBlockContentError::ReadError {
                        step: block_type,
                        source: err,
                    })?;
                workspace.buffer.push(smaller);

                self.internal_state = DecoderState::ReadyToDecodeNextHeader;
                Ok(u64::from(header.decompressed_size))
            }

            BlockType::Reserved => {
                panic!("How did you even get this. The decoder should error out if it detects a reserved-type block");
            }

            BlockType::Compressed => {
                self.decompress_block(header, workspace, source)?;

                self.internal_state = DecoderState::ReadyToDecodeNextHeader;
                Ok(u64::from(header.content_size))
            }
        }
    }

    fn decompress_block(
        &mut self,
        header: &BlockHeader,
        workspace: &mut DecoderScratch, //reuse this as often as possible. Not only if the trees are reused but also reuse the allocations when building new trees
        mut source: impl Read,
    ) -> Result<(), DecompressBlockError> {
        workspace
            .block_content_buffer
            .resize(header.content_size as usize, 0);

        source.read_exact(workspace.block_content_buffer.as_mut_slice())?;
        let raw = workspace.block_content_buffer.as_slice();

        let mut section = LiteralsSection::new();
        let bytes_in_literals_header = section.parse_from_header(raw)?;
        let raw = &raw[bytes_in_literals_header as usize..];
        vprintln!(
            "Found {} literalssection with regenerated size: {}, and compressed size: {:?}",
            section.ls_type,
            section.regenerated_size,
            section.compressed_size
        );

        let upper_limit_for_literals = match section.compressed_size {
            Some(x) => x as usize,
            None => match section.ls_type {
                LiteralsSectionType::RLE => 1,
                LiteralsSectionType::Raw => section.regenerated_size as usize,
                _ => panic!("Bug in this library"),
            },
        };

        if raw.len() < upper_limit_for_literals {
            return Err(DecompressBlockError::MalformedSectionHeader {
                expected_len: upper_limit_for_literals,
                remaining_bytes: raw.len(),
            });
        }

        let raw_literals = &raw[..upper_limit_for_literals];
        vprintln!("Slice for literals: {}", raw_literals.len());

        workspace.literals_buffer.clear(); //all literals of the previous block must have been used in the sequence execution anyways. just be defensive here
        let bytes_used_in_literals_section = decode_literals(
            &section,
            &mut workspace.huf,
            raw_literals,
            &mut workspace.literals_buffer,
        )?;
        assert!(
            section.regenerated_size == workspace.literals_buffer.len() as u32,
            "Wrong number of literals: {}, Should have been: {}",
            workspace.literals_buffer.len(),
            section.regenerated_size
        );
        assert!(bytes_used_in_literals_section == upper_limit_for_literals as u32);

        let raw = &raw[upper_limit_for_literals..];
        vprintln!("Slice for sequences with headers: {}", raw.len());

        let mut seq_section = SequencesHeader::new();
        let bytes_in_sequence_header = seq_section.parse_from_header(raw)?;
        let raw = &raw[bytes_in_sequence_header as usize..];
        vprintln!(
            "Found sequencessection with sequences: {} and size: {}",
            seq_section.num_sequences,
            raw.len()
        );

        assert!(
            u32::from(bytes_in_literals_header)
                + bytes_used_in_literals_section
                + u32::from(bytes_in_sequence_header)
                + raw.len() as u32
                == header.content_size
        );
        vprintln!("Slice for sequences: {}", raw.len());

        if seq_section.num_sequences != 0 {
            decode_sequences(
                &seq_section,
                raw,
                &mut workspace.fse,
                &mut workspace.sequences,
            )?;
            vprintln!("Executing sequences");
            execute_sequences(workspace)?;
        } else {
            workspace.buffer.push(&workspace.literals_buffer);
            workspace.sequences.clear();
        }

        Ok(())
    }

    pub fn read_block_header(
        &mut self,
        mut r: impl Read,
    ) -> Result<(BlockHeader, u8), BlockHeaderReadError> {
        //match self.internal_state {
        //    DecoderState::ReadyToDecodeNextHeader => {/* Happy :) */},
        //    DecoderState::Failed => return Err(format!("Cant decode next block if failed along the way. Results will be nonsense")),
        //    DecoderState::ReadyToDecodeNextBody => return Err(format!("Cant decode next block header, while expecting to decode the body of the previous block. Results will be nonsense")),
        //}

        r.read_exact(&mut self.header_buffer[0..3])?;

        let btype = self.block_type()?;
        if let BlockType::Reserved = btype {
            return Err(BlockHeaderReadError::FoundReservedBlock);
        }

        let block_size = self.block_content_size()?;
        let decompressed_size = match btype {
            BlockType::Raw => block_size,
            BlockType::RLE => block_size,
            BlockType::Reserved => 0, //should be catched above, this is an error state
            BlockType::Compressed => 0, //unknown but will be smaller than 128kb (or window_size if that is smaller than 128kb)
        };
        let content_size = match btype {
            BlockType::Raw => block_size,
            BlockType::Compressed => block_size,
            BlockType::RLE => 1,
            BlockType::Reserved => 0, //should be catched above, this is an error state
        };

        let last_block = self.is_last();

        self.reset_buffer();
        self.internal_state = DecoderState::ReadyToDecodeNextBody;

        //just return 3. Blockheaders always take 3 bytes
        Ok((
            BlockHeader {
                last_block,
                block_type: btype,
                decompressed_size,
                content_size,
            },
            3,
        ))
    }

    fn reset_buffer(&mut self) {
        self.header_buffer[0] = 0;
        self.header_buffer[1] = 0;
        self.header_buffer[2] = 0;
    }

    fn is_last(&self) -> bool {
        self.header_buffer[0] & 0x1 == 1
    }

    fn block_type(&self) -> Result<BlockType, BlockTypeError> {
        let t = (self.header_buffer[0] >> 1) & 0x3;
        match t {
            0 => Ok(BlockType::Raw),
            1 => Ok(BlockType::RLE),
            2 => Ok(BlockType::Compressed),
            3 => Ok(BlockType::Reserved),
            other => Err(BlockTypeError::InvalidBlocktypeNumber { num: other }),
        }
    }

    fn block_content_size(&self) -> Result<u32, BlockSizeError> {
        let val = self.block_content_size_unchecked();
        if val > ABSOLUTE_MAXIMUM_BLOCK_SIZE {
            Err(BlockSizeError::BlockSizeTooLarge { size: val })
        } else {
            Ok(val)
        }
    }

    fn block_content_size_unchecked(&self) -> u32 {
        u32::from(self.header_buffer[0] >> 3) //push out type and last_block flags. Retain 5 bit
            | (u32::from(self.header_buffer[1]) << 5)
            | (u32::from(self.header_buffer[2]) << 13)
    }
}