Avoid cloning jpeg_tables and instead pass Arc<Vec<u8>>

src/decoder/image.rs

@@ -1,4 +1,5 @@
 use super::ifd::{Directory, Value};
+use super::stream::{add_app14segment, JpegTagApp14Transform};
 use super::tag_reader::TagReader;
 use super::Limits;
 use super::{stream::SmartReader, ChunkType};
@@ -6,6 +7,7 @@ use crate::tags::{CompressionMethod, PhotometricInterpretation, Predictor, Sampl
 use crate::{TiffError, TiffFormatError, TiffResult, TiffUnsupportedError};
 use std::convert::{TryFrom, TryInto};
 use std::io::{Read, Seek};
+use std::sync::Arc;
 
 #[derive(Debug)]
 pub(crate) struct StripDecodeState {
@@ -87,7 +89,7 @@ pub(crate) struct Image {
     pub photometric_interpretation: PhotometricInterpretation,
     pub compression_method: CompressionMethod,
     pub predictor: Predictor,
-    pub jpeg_tables: Option<Vec<u8>>,
+    pub jpeg_tables: Option<Arc<Vec<u8>>>,
     pub chunk_type: ChunkType,
     pub strip_decoder: Option<StripDecodeState>,
     pub tile_attributes: Option<TileAttributes>,
@@ -130,7 +132,7 @@ impl Image {
         let jpeg_tables = if compression_method == CompressionMethod::ModernJPEG
             && ifd.contains_key(&Tag::JPEGTables)
         {
-            let vec = tag_reader
+            let mut vec = tag_reader
                 .find_tag(Tag::JPEGTables)?
                 .unwrap()
                 .into_u8_vec()?;
@@ -139,7 +141,13 @@ impl Image {
                     TiffFormatError::InvalidTagValueType(Tag::JPEGTables),
                 ));
             }
-            Some(vec)
+
+            // TODO: Can we avoid this somehow?
+            if photometric_interpretation == PhotometricInterpretation::RGB {
+                add_app14segment(&mut vec, JpegTagApp14Transform::App14TransformUnknown)
+            }
+
+            Some(Arc::new(vec))
         } else {
             None
         };

src/decoder/mod.rs

@@ -1,6 +1,7 @@
 use std::collections::{HashMap, HashSet};
 use std::convert::TryFrom;
 use std::io::{self, Cursor, Read, Seek};
+use std::sync::Arc;
 use std::{cmp, ops::Range};
 
 use crate::{
@@ -1151,7 +1152,7 @@ impl<R: Read + Seek> Decoder<R> {
         photometric_interpretation: PhotometricInterpretation,
         compression_method: CompressionMethod,
         compressed_length: u64,
-        jpeg_tables: Option<Vec<u8>>,
+        jpeg_tables: Option<Arc<Vec<u8>>>,
     ) -> TiffResult<Box<dyn Read + 'r>> {
         Ok(match compression_method {
             CompressionMethod::None => Box::new(reader),
@@ -1228,70 +1229,9 @@ impl<R: Read + Seek> Decoder<R> {
         Ok(u32::try_from(self.image().chunk_offsets.len())?)
     }
 
+    #[deprecated = "Use read_image instead"]
     pub fn read_jpeg(&mut self) -> TiffResult<DecodingResult> {
-        self.check_chunk_type(ChunkType::Strip)?;
-
-        if self.image().chunk_offsets.len() == 0 {
-            return Err(TiffError::FormatError(TiffFormatError::RequiredTagEmpty(
-                Tag::StripOffsets,
-            )));
-        }
-
-        if self.image().chunk_offsets[0] as usize > self.limits.intermediate_buffer_size
-            || self
-                .image()
-                .chunk_bytes
-                .iter()
-                .any(|&x| x as usize > self.limits.intermediate_buffer_size)
-        {
-            return Err(TiffError::LimitsExceeded);
-        }
-
-        let mut res_img = Vec::with_capacity(self.image().chunk_offsets[0] as usize);
-
-        for idx in 0..self.image.chunk_offsets.len() {
-            let offset = self.image.chunk_offsets[idx];
-            let length = self.image.chunk_bytes[idx];
-
-            self.goto_offset_u64(offset)?;
-
-            if self.image().jpeg_tables.is_some() && length < 2 {
-                return Err(TiffError::FormatError(
-                    TiffFormatError::InvalidTagValueType(Tag::JPEGTables),
-                ));
-            }
-
-            let jpeg_tables = self.image().jpeg_tables.clone();
-            let photometric_interpretation = self.image().photometric_interpretation;
-
-            let jpeg_reader = JpegReader::new(
-                &mut self.reader,
-                length,
-                jpeg_tables,
-                &photometric_interpretation,
-            )?;
-            let mut decoder = jpeg::Decoder::new(jpeg_reader);
-
-            match decoder.decode() {
-                Ok(mut val) => res_img.append(&mut val),
-                Err(e) => {
-                    return match e {
-                        jpeg::Error::Io(io_err) => Err(TiffError::IoError(io_err)),
-                        jpeg::Error::Format(fmt_err) => {
-                            Err(TiffError::FormatError(TiffFormatError::Format(fmt_err)))
-                        }
-                        jpeg::Error::Unsupported(_) => Err(TiffError::UnsupportedError(
-                            TiffUnsupportedError::UnknownInterpretation,
-                        )),
-                        jpeg::Error::Internal(_) => Err(TiffError::UnsupportedError(
-                            TiffUnsupportedError::UnknownInterpretation,
-                        )),
-                    }
-                }
-            }
-        }
-
-        Ok(DecodingResult::U8(res_img))
+        self.read_image()
     }
 
     pub fn read_strip_to_buffer(&mut self, mut buffer: DecodingBuffer) -> TiffResult<()> {
@@ -1546,11 +1486,9 @@ impl<R: Read + Seek> Decoder<R> {
 
     /// Decodes the entire image and return it as a Vector
     pub fn read_image(&mut self) -> TiffResult<DecodingResult> {
-        let result = match self.image().chunk_type {
-            ChunkType::Strip => self.read_stripped_image()?,
-            ChunkType::Tile => self.read_tiled_image()?,
-        };
-
-        Ok(result)
+        match self.image().chunk_type {
+            ChunkType::Strip => self.read_stripped_image(),
+            ChunkType::Tile => self.read_tiled_image(),
+        }
     }
 }

src/decoder/stream.rs

@@ -1,7 +1,8 @@
 //! All IO functionality needed for TIFF decoding
 
 use std::convert::TryFrom;
-use std::io::{self, BufRead, BufReader, Read, Seek, Take};
+use std::io::{self, BufRead, BufReader, Read, Seek, SeekFrom, Take};
+use std::sync::Arc;
 
 use crate::tags::PhotometricInterpretation;
 
@@ -189,7 +190,13 @@ impl<R: Read> Read for LZWReader<R> {
 ///
 
 pub(crate) struct JpegReader {
+    jpeg_tables: Option<Arc<Vec<u8>>>,
+
+    photometric_interpretation: Option<PhotometricInterpretation>,
+
     buffer: io::Cursor<Vec<u8>>,
+
+    offset: usize,
 }
 
 impl JpegReader {
@@ -208,7 +215,7 @@ impl JpegReader {
     pub fn new<R>(
         reader: &mut SmartReader<R>,
         length: u64,
-        jpeg_tables: Option<Vec<u8>>,
+        jpeg_tables: Option<Arc<Vec<u8>>>,
         photometric_interpretation: &PhotometricInterpretation,
     ) -> io::Result<JpegReader>
     where
@@ -220,11 +227,11 @@ impl JpegReader {
         reader.read_exact(&mut segment[..])?;
 
         match jpeg_tables {
-            Some(mut jpeg_data) => {
+            Some(jpeg_tables) => {
                 assert!(
-                    jpeg_data.len() >= 2,
+                    jpeg_tables.len() >= 2,
                     "jpeg_tables, if given, must be at least 2 bytes long. Got {:?}",
-                    jpeg_data
+                    jpeg_tables
                 );
 
                 assert!(
@@ -233,33 +240,43 @@ impl JpegReader {
                     length
                 );
 
-                if photometric_interpretation == &PhotometricInterpretation::RGB {
-                    add_app14segment(&mut jpeg_data, JpegTagApp14Transform::App14TransformUnknown)
-                }
+                // if photometric_interpretation == &PhotometricInterpretation::RGB {
+                //     add_app14segment(&mut jpeg_data, JpegTagApp14Transform::App14TransformUnknown)
+                // }
 
-                let truncated_length = jpeg_data.len() - 2;
-                jpeg_data.truncate(truncated_length);
-                jpeg_data.extend_from_slice(&segment[2..]);
+                // let truncated_length = jpeg_data.len() - 2;
+                // jpeg_data.truncate(truncated_length);
+                // jpeg_data.extend_from_slice(&segment[2..]);
+
+                let mut buffer = io::Cursor::new(segment);
+                // Skip the first two bytes
+                buffer.seek(SeekFrom::Start(2))?;
 
                 Ok(JpegReader {
-                    buffer: io::Cursor::new(jpeg_data),
+                    buffer,
+                    jpeg_tables: Some(jpeg_tables),
+                    photometric_interpretation: Some(*photometric_interpretation),
+                    offset: 0,
                 })
             }
             None => Ok(JpegReader {
                 buffer: io::Cursor::new(segment),
+                jpeg_tables: None,
+                photometric_interpretation: None,
+                offset: 0,
             }),
         }
     }
 }
 
 #[repr(u8)]
-enum JpegTagApp14Transform {
+pub(crate) enum JpegTagApp14Transform {
     // App14TransformYCCK = 2,
     // App14TransformYCbCr = 1,
     App14TransformUnknown = 0,
 }
 
-fn add_app14segment(jpeg_tables: &mut Vec<u8>, transform: JpegTagApp14Transform) {
+pub(crate) fn add_app14segment(jpeg_tables: &mut Vec<u8>, transform: JpegTagApp14Transform) {
     // Add JPEG Tag APP14 Adobe segment to jpeg_tables.
     // This segment stores image encoding information for DCT filters.
     // When `transform` value is 0 which is defined as Unknown, jpeg-decoder interpret the
@@ -307,7 +324,31 @@ fn add_app14segment(jpeg_tables: &mut Vec<u8>, transform: JpegTagApp14Transform)
 impl Read for JpegReader {
     // #[inline]
     fn read(&mut self, buf: &mut [u8]) -> io::Result<usize> {
-        self.buffer.read(buf)
+        let mut start = 0;
+
+        if let Some(jpeg_tables) = &self.jpeg_tables {
+            if jpeg_tables.len() - 2 > self.offset {
+                // Read (rest of) jpeg_tables to buf (without the last two bytes)
+                let size_remaining = jpeg_tables.len() - self.offset - 2;
+                let to_copy = size_remaining.min(buf.len());
+
+                buf[start..start + to_copy]
+                    .copy_from_slice(&jpeg_tables[self.offset..self.offset + to_copy]);
+
+                self.offset += to_copy;
+
+                if to_copy == buf.len() {
+                    return Ok(to_copy);
+                }
+
+                start += to_copy;
+            }
+        }
+
+        let read = self.buffer.read(&mut buf[start..])?;
+        self.offset += read;
+
+        Ok(read + start)
     }
 }