Add checkpoint, storing an IFD offset of a tiff

CHANGES.md

@@ -1,20 +1,3 @@
-# Version 0.6.1
-
-New features:
-* Support for reading `u16` and ascii string tags.
-* Added `Limits::unlimited` for disabling all limits.
-* Added `ImageEncoder::rows_per_strip` to overwrite the default.
-
-Changes:
-* The default strip size for chunked encoding is now 1MB, up from 8KB. This
-  should lead to more efficient decoding and compression.
-
-Fixes:
-* Fixed a bug where LZW compressed strips could not be decoded, instead
-  returning an error `Inconsistent sizes encountered`.
-* Reading a tag with a complex type and a single value returns the proper Value
-  variant, instead of a vector with one entry.
-
 # Version 0.6.0
 
 New features:

Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "tiff"
-version = "0.6.1"
+version = "0.6.0"
 authors = [
     "ccgn",
     "bvssvni <bvssvni@gmail.com>",
@@ -15,7 +15,6 @@ license = "MIT"
 description = "TIFF decoding and encoding library in pure Rust"
 repository = "https://github.com/image-rs/image-tiff"
 exclude = ["tests/images/*", "tests/fuzz_images/*"]
-edition = "2018"
 
 [dependencies]
 weezl = "0.1.0"

src/decoder/ifd.rs

@@ -6,19 +6,18 @@ use std::io::{self, Read, Seek};
 use std::mem;
 
 use super::stream::{ByteOrder, EndianReader, SmartReader};
-use crate::tags::{Tag, Type};
-use crate::{TiffError, TiffFormatError, TiffResult};
+use tags::{Tag, Type};
+use {TiffError, TiffFormatError, TiffResult, TiffUnsupportedError};
 
 use self::Value::{
-    Ascii, Byte, Double, Float, List, Rational, RationalBig, SRational, SRationalBig, Short,
-    Signed, SignedBig, Unsigned, UnsignedBig,
+    Ascii, Byte, Double, Float, List, Rational, RationalBig, SRational, SRationalBig, Signed,
+    SignedBig, Unsigned, UnsignedBig,
 };
 
 #[allow(unused_qualifications)]
 #[derive(Debug, Clone, PartialEq)]
 pub enum Value {
     Byte(u8),
-    Short(u16),
     Signed(i32),
     SignedBig(i64),
     Unsigned(u32),
@@ -43,20 +42,8 @@ impl Value {
         }
     }
 
-    pub fn into_u16(self) -> TiffResult<u16> {
-        match self {
-            Short(val) => Ok(val),
-            Unsigned(val) => Ok(u16::try_from(val)?),
-            UnsignedBig(val) => Ok(u16::try_from(val)?),
-            val => Err(TiffError::FormatError(
-                TiffFormatError::UnsignedIntegerExpected(val),
-            )),
-        }
-    }
-
     pub fn into_u32(self) -> TiffResult<u32> {
         match self {
-            Short(val) => Ok(val.into()),
             Unsigned(val) => Ok(val),
             UnsignedBig(val) => Ok(u32::try_from(val)?),
             val => Err(TiffError::FormatError(
@@ -77,7 +64,6 @@ impl Value {
 
     pub fn into_u64(self) -> TiffResult<u64> {
         match self {
-            Short(val) => Ok(val.into()),
             Unsigned(val) => Ok(val.into()),
             UnsignedBig(val) => Ok(val),
             val => Err(TiffError::FormatError(
@@ -114,15 +100,6 @@ impl Value {
         }
     }
 
-    pub fn into_string(self) -> TiffResult<String> {
-        match self {
-            Ascii(val) => Ok(val),
-            val => Err(TiffError::FormatError(
-                TiffFormatError::SignedIntegerExpected(val),
-            )),
-        }
-    }
-
     pub fn into_u32_vec(self) -> TiffResult<Vec<u32>> {
         match self {
             List(vec) => {
@@ -162,22 +139,6 @@ impl Value {
         }
     }
 
-    pub fn into_u16_vec(self) -> TiffResult<Vec<u16>> {
-        match self {
-            List(vec) => {
-                let mut new_vec = Vec::with_capacity(vec.len());
-                for v in vec {
-                    new_vec.push(v.into_u16()?)
-                }
-                Ok(new_vec)
-            }
-            Short(val) => Ok(vec![val]),
-            val => Err(TiffError::FormatError(
-                TiffFormatError::UnsignedIntegerExpected(val),
-            )),
-        }
-    }
-
     pub fn into_i32_vec(self) -> TiffResult<Vec<i32>> {
         match self {
             List(vec) => {
@@ -332,207 +293,172 @@ impl Entry {
         limits: &super::Limits,
         decoder: &mut super::Decoder<R>,
     ) -> TiffResult<Value> {
-        // Case 1: there are no values so we can return immediately.
-        if self.count == 0 {
-            return Ok(List(Vec::new()));
-        }
-
         let bo = decoder.byte_order();
-        let value_bytes = self.count
-            * match self.type_ {
-                Type::BYTE | Type::SBYTE | Type::ASCII | Type::UNDEFINED => 1,
-                Type::SHORT | Type::SSHORT => 2,
-                Type::LONG | Type::SLONG | Type::FLOAT => 4,
-                Type::LONG8 | Type::DOUBLE | Type::RATIONAL | Type::SRATIONAL => 8,
-                Type::__NonExhaustive => unreachable!(),
-            };
-
-        // Case 2: there is one value.
-        if self.count == 1 {
-            // 2a: the value is 5-8 bytes and we're in BigTiff mode.
-            if decoder.bigtiff && value_bytes > 4 && value_bytes <= 8 {
-                return Ok(match self.type_ {
-                    Type::LONG8 => UnsignedBig(self.r(bo).read_u64()?),
-                    Type::DOUBLE => Double(self.r(bo).read_f64()?),
-                    Type::RATIONAL => {
-                        let mut r = self.r(bo);
-                        Rational(r.read_u32()?, r.read_u32()?)
-                    }
-                    Type::SRATIONAL => {
-                        let mut r = self.r(bo);
-                        SRational(r.read_i32()?, r.read_i32()?)
-                    }
-                    Type::BYTE
-                    | Type::SBYTE
-                    | Type::ASCII
-                    | Type::UNDEFINED
-                    | Type::SHORT
-                    | Type::SSHORT
-                    | Type::LONG
-                    | Type::SLONG
-                    | Type::FLOAT => unreachable!(),
-                    Type::__NonExhaustive => unreachable!(),
-                });
+        let bigtiff_value = if decoder.bigtiff {
+            match (self.type_, self.count) {
+                (Type::BYTE, 5) => Some(offset_to_bytes(5, self)?),
+                (Type::BYTE, 6) => Some(offset_to_bytes(6, self)?),
+                (Type::BYTE, 7) => Some(offset_to_bytes(7, self)?),
+                (Type::BYTE, 8) => Some(offset_to_bytes(8, self)?),
+                (Type::SBYTE, 5) => Some(offset_to_sbytes(5, self)?),
+                (Type::SBYTE, 6) => Some(offset_to_sbytes(6, self)?),
+                (Type::SBYTE, 7) => Some(offset_to_sbytes(7, self)?),
+                (Type::SBYTE, 8) => Some(offset_to_sbytes(8, self)?),
+                (Type::SHORT, 3) => {
+                    let mut r = self.r(bo);
+                    Some(List(vec![
+                        Unsigned(u32::from(r.read_u16()?)),
+                        Unsigned(u32::from(r.read_u16()?)),
+                        Unsigned(u32::from(r.read_u16()?)),
+                    ]))
+                }
+                (Type::SSHORT, 3) => {
+                    let mut r = self.r(bo);
+                    Some(List(vec![
+                        Signed(i32::from(r.read_i16()?)),
+                        Signed(i32::from(r.read_i16()?)),
+                        Signed(i32::from(r.read_i16()?)),
+                    ]))
+                }
+                (Type::SHORT, 4) => {
+                    let mut r = self.r(bo);
+                    Some(List(vec![
+                        Unsigned(u32::from(r.read_u16()?)),
+                        Unsigned(u32::from(r.read_u16()?)),
+                        Unsigned(u32::from(r.read_u16()?)),
+                        Unsigned(u32::from(r.read_u16()?)),
+                    ]))
+                }
+                (Type::SSHORT, 4) => {
+                    let mut r = self.r(bo);
+                    Some(List(vec![
+                        Signed(i32::from(r.read_i16()?)),
+                        Signed(i32::from(r.read_i16()?)),
+                        Signed(i32::from(r.read_i16()?)),
+                        Signed(i32::from(r.read_i16()?)),
+                    ]))
+                }
+                (Type::LONG, 2) => {
+                    let mut r = self.r(bo);
+                    Some(List(vec![Unsigned(r.read_u32()?), Unsigned(r.read_u32()?)]))
+                }
+                (Type::SLONG, 2) => {
+                    let mut r = self.r(bo);
+                    Some(List(vec![Signed(r.read_i32()?), Signed(r.read_i32()?)]))
+                }
+                (Type::RATIONAL, 1) => {
+                    let mut r = self.r(bo);
+                    let numerator = r.read_u32()?;
+                    let denominator = r.read_u32()?;
+                    Some(Rational(numerator, denominator))
+                }
+                (Type::SRATIONAL, 1) => {
+                    let mut r = self.r(bo);
+                    let numerator = r.read_i32()?;
+                    let denominator = r.read_i32()?;
+                    Some(SRational(numerator, denominator))
+                }
+                _ => None,
             }
+        } else {
+            None
+        };
 
-            // 2b: the value is at most 4 bytes or doesn't fit in the offset field.
-            return Ok(match self.type_ {
-                Type::BYTE => Unsigned(u32::from(self.offset[0])),
-                Type::SBYTE => Signed(i32::from(self.offset[0] as i8)),
-                Type::UNDEFINED => Byte(self.offset[0]),
-                Type::SHORT => Unsigned(u32::from(self.r(bo).read_u16()?)),
-                Type::SSHORT => Signed(i32::from(self.r(bo).read_i16()?)),
-                Type::LONG => Unsigned(self.r(bo).read_u32()?),
-                Type::SLONG => Signed(self.r(bo).read_i32()?),
-                Type::FLOAT => Float(self.r(bo).read_f32()?),
-                Type::ASCII => {
-                    if self.offset[0] == 0 {
-                        Ascii("".to_string())
+        if let Some(v) = bigtiff_value {
+            Ok(v)
+        } else {
+            match (self.type_, self.count) {
+                // TODO check if this could give wrong results
+                // at a different endianess of file/computer.
+                (Type::BYTE, 1) => Ok(Unsigned(u32::from(self.offset[0]))),
+                (Type::BYTE, 2) => offset_to_bytes(2, self),
+                (Type::BYTE, 3) => offset_to_bytes(3, self),
+                (Type::BYTE, 4) => offset_to_bytes(4, self),
+                (Type::BYTE, n) => self.decode_offset(n, bo, limits, decoder, |decoder| {
+                    Ok(UnsignedBig(u64::from(decoder.read_byte()?)))
+                }),
+                (Type::SBYTE, 1) => Ok(Signed(i32::from(self.offset[0] as i8))),
+                (Type::SBYTE, 2) => offset_to_sbytes(2, self),
+                (Type::SBYTE, 3) => offset_to_sbytes(3, self),
+                (Type::SBYTE, 4) => offset_to_sbytes(4, self),
+                (Type::SBYTE, n) => self.decode_offset(n, bo, limits, decoder, |decoder| {
+                    Ok(SignedBig(i64::from(decoder.read_byte()? as i8)))
+                }),
+                (Type::SHORT, 1) => Ok(Unsigned(u32::from(self.r(bo).read_u16()?))),
+                (Type::SSHORT, 1) => Ok(Signed(i32::from(self.r(bo).read_i16()?))),
+                (Type::SHORT, 2) => {
+                    let mut r = self.r(bo);
+                    Ok(List(vec![
+                        Unsigned(u32::from(r.read_u16()?)),
+                        Unsigned(u32::from(r.read_u16()?)),
+                    ]))
+                }
+                (Type::SSHORT, 2) => {
+                    let mut r = self.r(bo);
+                    Ok(List(vec![
+                        Signed(i32::from(r.read_i16()?)),
+                        Signed(i32::from(r.read_i16()?)),
+                    ]))
+                }
+                (Type::SHORT, n) => self.decode_offset(n, bo, limits, decoder, |decoder| {
+                    Ok(UnsignedBig(u64::from(decoder.read_short()?)))
+                }),
+                (Type::SSHORT, n) => self.decode_offset(n, bo, limits, decoder, |decoder| {
+                    Ok(SignedBig(i64::from(decoder.read_sshort()?)))
+                }),
+                (Type::LONG, 1) => Ok(Unsigned(self.r(bo).read_u32()?)),
+                (Type::SLONG, 1) => Ok(Signed(self.r(bo).read_i32()?)),
+                (Type::LONG, n) => self.decode_offset(n, bo, limits, decoder, |decoder| {
+                    Ok(Unsigned(decoder.read_long()?))
+                }),
+                (Type::SLONG, n) => self.decode_offset(n, bo, limits, decoder, |decoder| {
+                    Ok(Signed(decoder.read_slong()?))
+                }),
+                (Type::FLOAT, n) => self.decode_offset(n, bo, limits, decoder, |decoder| {
+                    Ok(Float(decoder.read_float()?))
+                }),
+                (Type::DOUBLE, n) => self.decode_offset(n, bo, limits, decoder, |decoder| {
+                    Ok(Double(decoder.read_double()?))
+                }),
+                (Type::RATIONAL, n) => self.decode_offset(n, bo, limits, decoder, |decoder| {
+                    Ok(Rational(decoder.read_long()?, decoder.read_long()?))
+                }),
+                (Type::SRATIONAL, n) => self.decode_offset(n, bo, limits, decoder, |decoder| {
+                    Ok(SRational(decoder.read_slong()?, decoder.read_slong()?))
+                }),
+                (Type::LONG8, n) => self.decode_offset(n, bo, limits, decoder, |decoder| {
+                    Ok(UnsignedBig(decoder.read_long8()?))
+                }),
+                (Type::ASCII, n) => {
+                    let n = usize::try_from(n)?;
+                    if n > limits.decoding_buffer_size {
+                        return Err(TiffError::LimitsExceeded);
+                    }
+
+                    if (n <= 4 && !decoder.bigtiff) || (n <= 8 && decoder.bigtiff) {
+                        let mut buf = vec![0; n];
+                        self.r(bo).read_exact(&mut buf)?;
+                        let v = String::from_utf8(buf)?;
+                        let v = v.trim_matches(char::from(0));
+                        Ok(Ascii(v.into()))
                     } else {
-                        return Err(TiffError::FormatError(TiffFormatError::InvalidTag));
+                        if decoder.bigtiff {
+                            decoder.goto_offset_u64(self.r(bo).read_u64()?)?
+                        } else {
+                            decoder.goto_offset(self.r(bo).read_u32()?)?
+                        }
+                        let string = decoder.read_string(n)?;
+                        Ok(Ascii(string))
                     }
                 }
-                Type::LONG8 => {
-                    decoder.goto_offset(self.r(bo).read_u32()?)?;
-                    UnsignedBig(decoder.read_long8()?)
-                }
-                Type::DOUBLE => {
-                    decoder.goto_offset(self.r(bo).read_u32()?)?;
-                    Double(decoder.read_double()?)
-                }
-                Type::RATIONAL => {
-                    decoder.goto_offset(self.r(bo).read_u32()?)?;
-                    Rational(decoder.read_long()?, decoder.read_long()?)
-                }
-                Type::SRATIONAL => {
-                    decoder.goto_offset(self.r(bo).read_u32()?)?;
-                    SRational(decoder.read_slong()?, decoder.read_slong()?)
-                }
-                Type::__NonExhaustive => unreachable!(),
-            });
-        }
-
-        // Case 3: There is more than one value, but it fits in the offset field.
-        if value_bytes <= 4 || decoder.bigtiff && value_bytes <= 8 {
-            match self.type_ {
-                Type::BYTE => return Ok(offset_to_bytes(self.count as usize, self)?),
-                Type::SBYTE => return Ok(offset_to_sbytes(self.count as usize, self)?),
-                Type::ASCII => {
-                    let mut buf = vec![0; self.count as usize];
-                    self.r(bo).read_exact(&mut buf)?;
-                    let v = String::from_utf8(buf)?;
-                    let v = v.trim_matches(char::from(0));
-                    return Ok(Ascii(v.into()));
-                }
-                Type::UNDEFINED => {
-                    return Ok(List(
-                        self.offset[0..self.count as usize]
-                            .iter()
-                            .map(|&b| Byte(b))
-                            .collect(),
-                    ));
-                }
-                Type::SHORT => {
-                    let mut r = self.r(bo);
-                    let mut v = Vec::new();
-                    for _ in 0..self.count {
-                        v.push(Short(r.read_u16()?));
-                    }
-                    return Ok(List(v));
-                }
-                Type::SSHORT => {
-                    let mut r = self.r(bo);
-                    let mut v = Vec::new();
-                    for _ in 0..self.count {
-                        v.push(Signed(i32::from(r.read_i16()?)));
-                    }
-                    return Ok(List(v));
-                }
-                Type::LONG => {
-                    let mut r = self.r(bo);
-                    let mut v = Vec::new();
-                    for _ in 0..self.count {
-                        v.push(Unsigned(r.read_u32()?));
-                    }
-                    return Ok(List(v));
-                }
-                Type::SLONG => {
-                    let mut r = self.r(bo);
-                    let mut v = Vec::new();
-                    for _ in 0..self.count {
-                        v.push(Signed(r.read_i32()?));
-                    }
-                    return Ok(List(v));
-                }
-                Type::FLOAT => {
-                    let mut r = self.r(bo);
-                    let mut v = Vec::new();
-                    for _ in 0..self.count {
-                        v.push(Float(r.read_f32()?));
-                    }
-                    return Ok(List(v));
-                }
-                Type::LONG8 | Type::RATIONAL | Type::SRATIONAL | Type::DOUBLE => unreachable!(),
-                Type::__NonExhaustive => unreachable!(),
+                (Type::UNDEFINED, n) => self.decode_offset(n, bo, limits, decoder, |decoder| {
+                    Ok(Byte(u8::from(decoder.read_byte()?)))
+                }),
+                _ => Err(TiffError::UnsupportedError(
+                    TiffUnsupportedError::UnsupportedDataType,
+                )),
             }
         }
-
-        // Case 4: there is more than one value, and it doesn't fit in the offset field.
-        match self.type_ {
-            // TODO check if this could give wrong results
-            // at a different endianess of file/computer.
-            Type::BYTE => self.decode_offset(self.count, bo, limits, decoder, |decoder| {
-                Ok(UnsignedBig(u64::from(decoder.read_byte()?)))
-            }),
-            Type::SBYTE => self.decode_offset(self.count, bo, limits, decoder, |decoder| {
-                Ok(SignedBig(i64::from(decoder.read_byte()? as i8)))
-            }),
-            Type::SHORT => self.decode_offset(self.count, bo, limits, decoder, |decoder| {
-                Ok(UnsignedBig(u64::from(decoder.read_short()?)))
-            }),
-            Type::SSHORT => self.decode_offset(self.count, bo, limits, decoder, |decoder| {
-                Ok(SignedBig(i64::from(decoder.read_sshort()?)))
-            }),
-            Type::LONG => self.decode_offset(self.count, bo, limits, decoder, |decoder| {
-                Ok(Unsigned(decoder.read_long()?))
-            }),
-            Type::SLONG => self.decode_offset(self.count, bo, limits, decoder, |decoder| {
-                Ok(Signed(decoder.read_slong()?))
-            }),
-            Type::FLOAT => self.decode_offset(self.count, bo, limits, decoder, |decoder| {
-                Ok(Float(decoder.read_float()?))
-            }),
-            Type::DOUBLE => self.decode_offset(self.count, bo, limits, decoder, |decoder| {
-                Ok(Double(decoder.read_double()?))
-            }),
-            Type::RATIONAL => self.decode_offset(self.count, bo, limits, decoder, |decoder| {
-                Ok(Rational(decoder.read_long()?, decoder.read_long()?))
-            }),
-            Type::SRATIONAL => self.decode_offset(self.count, bo, limits, decoder, |decoder| {
-                Ok(SRational(decoder.read_slong()?, decoder.read_slong()?))
-            }),
-            Type::LONG8 => self.decode_offset(self.count, bo, limits, decoder, |decoder| {
-                Ok(UnsignedBig(decoder.read_long8()?))
-            }),
-            Type::UNDEFINED => self.decode_offset(self.count, bo, limits, decoder, |decoder| {
-                Ok(Byte(u8::from(decoder.read_byte()?)))
-            }),
-            Type::ASCII => {
-                let n = usize::try_from(self.count)?;
-                if n > limits.decoding_buffer_size {
-                    return Err(TiffError::LimitsExceeded);
-                }
-
-                if decoder.bigtiff {
-                    decoder.goto_offset_u64(self.r(bo).read_u64()?)?
-                } else {
-                    decoder.goto_offset(self.r(bo).read_u32()?)?
-                }
-                let string = decoder.read_string(n)?;
-                Ok(Ascii(string))
-            }
-            Type::__NonExhaustive => unreachable!(),
-        }
     }
 
     #[inline]

src/decoder/mod.rs

@@ -3,12 +3,10 @@ use std::collections::HashMap;
 use std::convert::TryFrom;
 use std::io::{self, Read, Seek};
 
-use crate::{ColorType, TiffError, TiffFormatError, TiffResult, TiffUnsupportedError};
+use {ColorType, TiffError, TiffFormatError, TiffResult, TiffUnsupportedError};
 
 use self::ifd::Directory;
-use crate::tags::{
-    CompressionMethod, PhotometricInterpretation, Predictor, SampleFormat, Tag, Type,
-};
+use tags::{CompressionMethod, PhotometricInterpretation, Predictor, SampleFormat, Tag, Type};
 
 use self::stream::{
     ByteOrder, DeflateReader, EndianReader, JpegReader, LZWReader, PackBitsReader, SmartReader,
@@ -176,24 +174,6 @@ pub struct Limits {
     _non_exhaustive: (),
 }
 
-impl Limits {
-    /// A configuration that does not impose any limits.
-    ///
-    /// This is a good start if the caller only wants to impose selective limits, contrary to the
-    /// default limits which allows selectively disabling limits.
-    ///
-    /// Note that this configuration is likely to crash on excessively large images since,
-    /// naturally, the machine running the program does not have infinite memory.
-    pub fn unlimited() -> Limits {
-        Limits {
-            decoding_buffer_size: usize::max_value(),
-            ifd_value_size: usize::max_value(),
-            intermediate_buffer_size: usize::max_value(),
-            _non_exhaustive: (),
-        }
-    }
-}
-
 impl Default for Limits {
     fn default() -> Limits {
         Limits {
@@ -218,6 +198,7 @@ where
     bigtiff: bool,
     limits: Limits,
     next_ifd: Option<u64>,
+    current_ifd: Option<u64>,
     ifd: Option<Directory>,
     width: u32,
     height: u32,
@@ -229,6 +210,11 @@ where
     strip_decoder: Option<StripDecodeState>,
 }
 
+/// Stores state necessary to seek back (or forward) to particular part of the tiff.
+pub struct Checkpoint {
+    offset: u64,
+}
+
 trait Wrapping {
     fn wrapping_add(&self, other: Self) -> Self;
 }
@@ -333,6 +319,7 @@ impl<R: Read + Seek> Decoder<R> {
             bigtiff: false,
             limits: Default::default(),
             next_ifd: None,
+            current_ifd: None,
             ifd: None,
             width: 0,
             height: 0,
@@ -360,11 +347,6 @@ impl<R: Read + Seek> Decoder<R> {
             PhotometricInterpretation::RGB => match self.bits_per_sample[..] {
                 [r, g, b] if [r, r] == [g, b] => Ok(ColorType::RGB(r)),
                 [r, g, b, a] if [r, r, r] == [g, b, a] => Ok(ColorType::RGBA(r)),
-                // FIXME: We should _ignore_ other components. In particular:
-                // > Beware of extra components. Some TIFF files may have more components per pixel
-                // than you think. A Baseline TIFF reader must skip over them gracefully,using the
-                // values of the SamplesPerPixel and BitsPerSample fields.
-                // > -- TIFF 6.0 Specification, Section 7, Additional Baseline requirements.
                 _ => Err(TiffError::UnsupportedError(
                     TiffUnsupportedError::InterpretationWithBits(
                         self.photometric_interpretation,
@@ -495,7 +477,6 @@ impl<R: Read + Seek> Decoder<R> {
             }
             _ => return Err(TiffUnsupportedError::UnsupportedSampleDepth(self.samples).into()),
         }
-
         Ok(())
     }
 
@@ -611,6 +592,44 @@ impl<R: Read + Seek> Decoder<R> {
         self.reader.seek(io::SeekFrom::Start(offset)).map(|_| ())
     }
 
+    /// Retrieve a checkpoint that allows seeking before the TIFF header.
+    ///
+    /// This is a special checkpoint where the caller should call `init` after restoring the
+    /// checkpoint and no ifd is being read.
+    pub fn checkpoint_init(&self) -> Checkpoint {
+        Checkpoint { offset: Checkpoint::INITIAL_OFFSET }
+    }
+
+    /// Retrieve a `Checkpoint` that can be used to seek to the current image directory.
+    pub fn checkpoint(&self) -> Checkpoint {
+        match self.current_ifd {
+            None => self.checkpoint_init(),
+            Some(offset) => Checkpoint { offset },
+        }
+    }
+
+    /// Get the next image directory as a checkpoint.
+    ///
+    /// This allows jumping forward to the next image directory later, without needing to pay the
+    /// additional cost of parsing it right away.
+    pub fn checkpoint_next(&self) -> Option<Checkpoint> {
+        match self.next_ifd {
+            None => None,
+            Some(offset) => Some(Checkpoint { offset })
+        }
+    }
+
+    /// Restore the state of a previous checkpoint.
+    pub fn restore(&mut self, checkpoint: &Checkpoint) -> TiffResult<()> {
+        if checkpoint.offset == Checkpoint::INITIAL_OFFSET {
+            self.goto_offset_u64(0)?;
+            Ok(())
+        } else {
+            self.next_ifd = Some(checkpoint.offset);
+            self.next_image()
+        }
+    }
+
     /// Reads a IFD entry.
     // An IFD entry has four fields:
     //
@@ -646,7 +665,10 @@ impl<R: Read + Seek> Decoder<R> {
                     TiffFormatError::ImageFileDirectoryNotFound,
                 ))
             }
-            Some(offset) => self.goto_offset_u64(offset)?,
+            Some(offset) => {
+                self.goto_offset_u64(offset)?;
+                self.current_ifd = Some(offset);
+            }
         }
         let num_tags = if self.bigtiff {
             self.read_long8()?
@@ -752,10 +774,6 @@ impl<R: Read + Seek> Decoder<R> {
     pub fn get_tag_u32_vec(&mut self, tag: Tag) -> TiffResult<Vec<u32>> {
         self.get_tag(tag)?.into_u32_vec()
     }
-
-    pub fn get_tag_u16_vec(&mut self, tag: Tag) -> TiffResult<Vec<u16>> {
-        self.get_tag(tag)?.into_u16_vec()
-    }
     pub fn get_tag_u64_vec(&mut self, tag: Tag) -> TiffResult<Vec<u64>> {
         self.get_tag(tag)?.into_u64_vec()
     }
@@ -775,11 +793,6 @@ impl<R: Read + Seek> Decoder<R> {
         self.get_tag(tag)?.into_u8_vec()
     }
 
-    /// Tries to retrieve a tag and convert it to a ascii vector.
-    pub fn get_tag_ascii_string(&mut self, tag: Tag) -> TiffResult<String> {
-        self.get_tag(tag)?.into_string()
-    }
-
     /// Decompresses the strip into the supplied buffer.
     /// Returns the number of bytes read.
     fn expand_strip<'a>(
@@ -787,7 +800,7 @@ impl<R: Read + Seek> Decoder<R> {
         buffer: DecodingBuffer<'a>,
         offset: u64,
         length: u64,
-        strip_sample_count: usize,
+        max_uncompressed_length: usize,
     ) -> TiffResult<usize> {
         let color_type = self.colortype()?;
         self.goto_offset_u64(offset)?;
@@ -803,7 +816,7 @@ impl<R: Read + Seek> Decoder<R> {
                 let (bytes, reader) = LZWReader::new(
                     &mut self.reader,
                     usize::try_from(length)?,
-                    strip_sample_count * buffer.byte_len(),
+                    max_uncompressed_length * buffer.byte_len(),
                 )?;
                 (bytes, Box::new(reader))
             }
@@ -814,7 +827,8 @@ impl<R: Read + Seek> Decoder<R> {
                 (bytes, Box::new(reader))
             }
             CompressionMethod::OldDeflate => {
-                let (bytes, reader) = DeflateReader::new(&mut self.reader, strip_sample_count)?;
+                let (bytes, reader) =
+                    DeflateReader::new(&mut self.reader, max_uncompressed_length)?;
                 (bytes, Box::new(reader))
             }
             method => {
@@ -824,14 +838,9 @@ impl<R: Read + Seek> Decoder<R> {
             }
         };
 
-        // FIXME: this might be suboptimal. We might default remaining bits to ´0`, which some
-        // other decoders might do.
-        if bytes / buffer.byte_len() > strip_sample_count {
+        if bytes / buffer.byte_len() > max_uncompressed_length {
             return Err(TiffError::FormatError(
-                TiffFormatError::UnexpectedCompressedData {
-                    actual_bytes: bytes,
-                    required_bytes: strip_sample_count * buffer.byte_len(),
-                },
+                TiffFormatError::InconsistentSizesEncountered,
             ));
         }
 
@@ -1064,10 +1073,7 @@ impl<R: Read + Seek> Decoder<R> {
 
         if units_read < buffer_size {
             return Err(TiffError::FormatError(
-                TiffFormatError::InconsistentStripSamples {
-                    actual_samples: units_read,
-                    required_samples: buffer_size,
-                },
+                TiffFormatError::InconsistentSizesEncountered,
             ));
         }
         if let Ok(predictor) = self.get_tag_unsigned(Tag::Predictor) {
@@ -1159,3 +1165,7 @@ impl<R: Read + Seek> Decoder<R> {
         Ok(result)
     }
 }
+
+impl Checkpoint {
+    const INITIAL_OFFSET: u64 = 0;
+}

src/decoder/stream.rs

@@ -267,11 +267,10 @@ impl LZWReader {
         let mut uncompressed = Vec::with_capacity(max_uncompressed_length);
         let mut decoder = weezl::decode::Decoder::with_tiff_size_switch(weezl::BitOrder::Msb, 8);
         let mut bytes_read = 0;
-
-        while uncompressed.len() < max_uncompressed_length {
+        while bytes_read < compressed_length && uncompressed.len() < max_uncompressed_length {
             let bytes_written = uncompressed.len();
             uncompressed.reserve(1 << 12);
-            let buffer_space = uncompressed.capacity().min(max_uncompressed_length);
+            let buffer_space = uncompressed.capacity();
             uncompressed.resize(buffer_space, 0u8);
 
             let result = decoder.decode_bytes(

src/encoder/colortype.rs

@@ -1,4 +1,4 @@
-use crate::tags::{PhotometricInterpretation, SampleFormat};
+use tags::{PhotometricInterpretation, SampleFormat};
 
 /// Trait for different colortypes that can be encoded.
 pub trait ColorType {

src/encoder/mod.rs

@@ -3,9 +3,9 @@ use std::convert::TryFrom;
 use std::io::{Seek, Write};
 use std::{cmp, io, mem};
 
-use crate::bytecast;
-use crate::error::{TiffError, TiffFormatError, TiffResult};
-use crate::tags::{self, ResolutionUnit, Tag, Type};
+use bytecast;
+use error::{TiffError, TiffFormatError, TiffResult};
+use tags::{self, ResolutionUnit, Tag, Type};
 
 pub mod colortype;
 mod writer;
@@ -589,9 +589,6 @@ impl<'a, W: Write + Seek> Drop for DirectoryEncoder<'a, W> {
 /// // You can encode tags here
 /// image.encoder().write_tag(Tag::Artist, "Image-tiff").unwrap();
 ///
-/// // Strip size can be configured before writing data
-/// image.rows_per_strip(2).unwrap();
-///
 /// let mut idx = 0;
 /// while image.next_strip_sample_count() > 0 {
 ///     let sample_count = image.next_strip_sample_count() as usize;
@@ -625,9 +622,8 @@ impl<'a, W: 'a + Write + Seek, T: ColorType> ImageEncoder<'a, W, T> {
         let row_samples = u64::from(width) * u64::try_from(<T>::BITS_PER_SAMPLE.len())?;
         let row_bytes = row_samples * u64::from(<T::Inner>::BYTE_LEN);
 
-        // Limit the strip size to prevent potential memory and security issues.
-        // Also keep the multiple strip handling 'oiled'
-        let rows_per_strip = (1_000_000 + row_bytes - 1) / row_bytes;
+        // As per tiff spec each strip should be about 8k long
+        let rows_per_strip = (8000 + row_bytes - 1) / row_bytes;
 
         let strip_count = (u64::from(height) + rows_per_strip - 1) / rows_per_strip;
 
@@ -759,28 +755,6 @@ impl<'a, W: 'a + Write + Seek, T: ColorType> ImageEncoder<'a, W, T> {
         self.encoder.write_tag(Tag::YResolution, value).unwrap();
     }
 
-    /// Set image number of lines per strip
-    ///
-    /// This function needs to be called before any calls to `write_data` or
-    /// `write_strip` and will return an error otherwise.
-    pub fn rows_per_strip(&mut self, value: u32) -> TiffResult<()> {
-        if self.strip_idx != 0 {
-            return Err(io::Error::new(
-                io::ErrorKind::InvalidInput,
-                "Cannot change strip size after data was written",
-            )
-            .into());
-        }
-        // Write tag as 32 bits
-        self.encoder.write_tag(Tag::RowsPerStrip, value)?;
-
-        let value: u64 = value as u64;
-        self.strip_count = (self.height as u64 + value - 1) / value;
-        self.rows_per_strip = value;
-
-        Ok(())
-    }
-
     fn finish_internal(&mut self) -> TiffResult<()> {
         self.encoder
             .write_tag(Tag::StripOffsets, &*self.strip_offsets)?;

src/encoder/writer.rs

@@ -1,4 +1,4 @@
-use crate::error::TiffResult;
+use error::TiffResult;
 use std::io::{self, Seek, SeekFrom, Write};
 
 pub fn write_tiff_header<W: Write>(writer: &mut TiffWriter<W>) -> TiffResult<()> {

src/error.rs

@@ -3,12 +3,10 @@ use std::fmt;
 use std::io;
 use std::string;
 
-use crate::decoder::ifd::Value;
-use crate::tags::{
-    CompressionMethod, PhotometricInterpretation, PlanarConfiguration, SampleFormat, Tag,
-};
-use crate::ColorType;
+use decoder::ifd::Value;
 use miniz_oxide::inflate::TINFLStatus;
+use tags::{CompressionMethod, PhotometricInterpretation, PlanarConfiguration, SampleFormat, Tag};
+use ColorType;
 
 /// Tiff error kinds.
 #[derive(Debug)]
@@ -43,14 +41,6 @@ pub enum TiffFormatError {
     TiffSignatureInvalid,
     ImageFileDirectoryNotFound,
     InconsistentSizesEncountered,
-    UnexpectedCompressedData {
-        actual_bytes: usize,
-        required_bytes: usize,
-    },
-    InconsistentStripSamples {
-        actual_samples: usize,
-        required_samples: usize,
-    },
     InvalidTag,
     InvalidTagValueType(Tag),
     RequiredTagNotFound(Tag),
@@ -74,26 +64,6 @@ impl fmt::Display for TiffFormatError {
             TiffSignatureInvalid => write!(fmt, "TIFF signature invalid."),
             ImageFileDirectoryNotFound => write!(fmt, "Image file directory not found."),
             InconsistentSizesEncountered => write!(fmt, "Inconsistent sizes encountered."),
-            UnexpectedCompressedData {
-                actual_bytes,
-                required_bytes,
-            } => {
-                write!(
-                    fmt,
-                    "Decompression returned different amount of bytes than expected: got {}, expected {}.",
-                    actual_bytes, required_bytes
-                )
-            }
-            InconsistentStripSamples {
-                actual_samples,
-                required_samples,
-            } => {
-                write!(
-                    fmt,
-                    "Inconsistent elements in strip: got {}, expected {}.",
-                    actual_samples, required_samples
-                )
-            }
             InvalidTag => write!(fmt, "Image contains invalid tag."),
             InvalidTagValueType(ref tag) => {
                 write!(fmt, "Tag `{:?}` did not have the expected value type.", tag)

src/tags.rs

@@ -120,14 +120,6 @@ pub enum Tag(u16) unknown("A private or extension tag") {
     TileOffsets = 324,
     TileByteCounts = 325,
     SampleFormat = 339,
-    // GeoTIFF
-    ModelPixelScaleTag = 33550, // (SoftDesk)
-    ModelTransformationTag = 34264, // (JPL Carto Group)
-    ModelTiepointTag = 33922, // (Intergraph)
-    GeoKeyDirectoryTag = 34735, // (SPOT)
-    GeoDoubleParamsTag = 34736, // (SPOT)
-    GeoAsciiParamsTag = 34737, // (SPOT)
-    GdalNodata = 42113, // Contains areas with missing data
 }
 }
 

tests/encode_images.rs

@@ -360,42 +360,3 @@ fn test_multipage_image() {
         assert_eq!(img_decoder.dimensions().unwrap(), (3, 3));
     }
 }
-
-#[test]
-/// verify rows per strip setting
-fn test_rows_per_strip() {
-    let mut file = Cursor::new(Vec::new());
-    {
-        let mut img_encoder = TiffEncoder::new(&mut file).unwrap();
-
-        let mut image = img_encoder.new_image::<colortype::Gray8>(100, 100).unwrap();
-        assert_eq!(image.next_strip_sample_count(), 100 * 100);
-        image.rows_per_strip(2).unwrap();
-        assert_eq!(image.next_strip_sample_count(), 2 * 100);
-
-        let img2: Vec<u8> = vec![0; 2 * 100];
-        image.write_strip(&img2[..]).unwrap();
-        assert!(image.rows_per_strip(5).is_err());
-        for i in 1..50 {
-            let img2: Vec<u8> = vec![i; 2 * 100];
-            image.write_strip(&img2[..]).unwrap();
-        }
-        assert!(image.write_strip(&img2[..]).is_err());
-        image.finish().unwrap();
-    }
-
-    file.seek(SeekFrom::Start(0)).unwrap();
-    {
-        let mut decoder = Decoder::new(&mut file).unwrap();
-        assert_eq!(decoder.get_tag_u64(Tag::RowsPerStrip).unwrap(), 2);
-        assert_eq!(decoder.strip_count().unwrap(), 50);
-
-        for i in 0..50 {
-            let img2 = [i; 2 * 100];
-            match decoder.read_strip().unwrap() {
-                DecodingResult::U8(data) => assert_eq!(&img2[..], &data[..]),
-                other => panic!("Incorrect strip type {:?}", other),
-            }
-        }
-    }
-}