Decoder: remove test make wrong assumption

The byte count for strips is the compressed length. The compressed
length may be larger than the uncompressed length, especially with poor
compression choices (multi-byte samples with packbits) or small images.

src/decoder/mod.rs

@@ -744,7 +744,7 @@ impl<R: Read + Seek> Decoder<R> {
 
         let buffer_size = usize::try_from(self.width)? * strip_height * self.bits_per_sample.len();
 
-        if buffer.len() < buffer_size || usize::try_from(byte_count)? / buffer.byte_len() > buffer_size {
+        if buffer.len() < buffer_size {
             return Err(TiffError::FormatError(
                 TiffFormatError::InconsistentSizesEncountered,
             ));