Failed to load RLE encoded file about tinyexr HOT 6 CLOSED

I have investigated your EXR file and offset tables looks corrupted.

http://www.openexr.com/openexrfilelayout.pdf (See page 10 for Offset tables)

First element of offset tables contains the byte offset to 1023th scanline correctly, but other elements are filled with zeros, so tinyexr reads pixel data from invalid location, resulting an assertion.

Also, exrheader program from OpenEXR displays your EXR is incomplete.

$ exrheader sample-tinyexr-bug.exr 

file sample-tinyexr-bug.exr (incomplete):

file format version: 2, flags 0x0
channels (type chlist):
    B, 16-bit floating-point, sampling 1 1
    G, 16-bit floating-point, sampling 1 1
    R, 16-bit floating-point, sampling 1 1
compression (type compression): run-length encoding
dataWindow (type box2i): (0 0) - (1023 1023)
displayWindow (type box2i): (0 0) - (1023 1023)
lineOrder (type lineOrder): increasing y
pixelAspectRatio (type float): 1
screenWindowCenter (type v2f): (0 0)
screenWindowWidth (type float): 1
type (type string): "scanlineimage"

How do you create your EXR file?

(For some reason, it looks OpenEXR can handle such an incomplete data, whereas tinyexr is not)

from tinyexr.

BTW I have found another issue related RLE, filed as #35

(FYI, offset tables in comp_rle.exr are correct, thus issue #35 looks purely the problem of RLE decoding)

from tinyexr.

The file was produced by a legacy toolchain, so I don't exactly have control over it. Anyway, I will see whether I can hack around the issue.

from tinyexr.

I found OpenEXR try to reconstruct offset information for incomplete EXR file.
(i.e. offset is zero)

https://github.com/openexr/openexr/blob/develop/OpenEXR/IlmImf/ImfScanLineInputFile.cpp#L338

And adding similar code to tinyexr, now your file can be read correctly.

BTW, Is it ok to add your EXR file to unit tests? @zeroeffects

from tinyexr.

It seems that the latest version is crashing with null pointer dereference in rleDecompress.

Also, you can use the image. It is a rectified image of alcantara material measured in camera dome under specific lighting and viewing conditions.

from tinyexr.

Ah, I found memory bug and fixed it. Now recent master should work.

Also, you can use the image.

Thank you. But I found the image has 5.8MB which is too large to add into tinyexr repo. Do you have much more smaller one? (e.g. 0.5 MB or less)

from tinyexr.