Repository navigation
Releases
v3.5.4
Patch release that addresses several build and security issues, and also fixes a regression introduced in v3.3.0 that leads to excessive memory usage during write, especially with DWA compression (Issue #2791).
- Fix a configuration issue with OpenJPH 0.33.0 that caused it to silently and incorrectly fall back to the vendored version (PR #2813)
- The vendored OpenJPH is now built in a custom namespace to avoid symbol collision with other versions when built statically (2811) built statically,
- The vendored OpenJPH version is now 0.33.0.
- Fix
exrmultipart -converton tiled and deep single-part files OPENEXR_USE_INTERNAL_IMATH/OPENJPHare are now included inOpenEXRConfig.halongsideOPENEXR_USE_INTERNAL_DEFLATE/ZSTD.- 32-bit glibc builds now guarantee support for files of size 2GiB or more (2812)
This release addresses the following security vulnerabilities:
- CVE Pending (GHSA-PP92-4RV3-WFQ6) OpenEXR: a signed-int data-window-origin overflow in the OpenEXRUtil Image API turns
Imf::loadImageinto an attacker-steered out-of-bounds write - CVE Pending (GHSA-2V77-PVHX-25C3) Integer overflow in legacy OpenEXR.OutputFile.writePixels() causes heap over-read
See CHANGES.md for more details.
Contributors
v3.4.18
Patch release that addresses a regression introduced in v3.3.0 that leads to excessive memory usage during write, especially with DWA compression (Issue #2791).
This release also addresses the following security vulnerabilities:
- CVE Pending (GHSA-PP92-4RV3-WFQ6) OpenEXR: a signed-int data-window-origin overflow in the OpenEXRUtil Image API turns
Imf::loadImageinto an attacker-steered out-of-bounds write - CVE Pending (GHSA-2V77-PVHX-25C3) Integer overflow in legacy
OpenEXR.OutputFile.writePixels()causes heap over-read
v3.3.16
Patch release that addresses a regression introduced in v3.3.0 that leads to excessive memory usage during write, especially with DWA compression (Issue #2791).
This release also addresses the following security vulnerabilities:
- CVE Pending (GHSA-PP92-4RV3-WFQ6) OpenEXR: a signed-int data-window-origin overflow in the OpenEXRUtil Image API turns
Imf::loadImageinto an attacker-steered out-of-bounds write - CVE Pending (GHSA-2V77-PVHX-25C3) Integer overflow in legacy
OpenEXR.OutputFile.writePixels()causes heap over-read
v3.5.3
Patch release that addresses 15 security vulnerabilities, 10 with pending CVE IDs, primarily heap buffer overflows and out-of-bounds writes triggered by malformed or malicious EXR files â several capable of remote code execution. See the Security section below for details.
This release includes no new features or changes in behavior other than improved error detection and rejection of invalid input, except for a fix for a regression introduced in v3.4.14 in reading deep images on big-endian hosts, due to a byte-order bug that returned an incorrect deep sample count table. Little-endian hosts are unaffected.
As part of this effort, the full test suite now passes cleanly under the address, undefined-behavior, and thread sanitizers.
Also in this release:
- Add support for Python 3.15
- Several diagnostic/correctness fixes across the utility tools (exrmetrics, exrmultipart, exrcheck) and core decompression paths (courtesy of @ssoj13/Alex Khal)
- Address several long-standing build errors and test failures on i686 (32-bit linux) and s390x (big-endian) hosts.
- Fix build issues on macOS < 10.15
- Fix build issues on powerpc
- Fix for absolute paths for
libmin OpenEXRTargets.cmake - Fix for running on Windows 7 involving
CreateFile2
Security
This release addresses the following security vulnerabilities:
- CVE Pending (GHSA-WHF5-29MJ-GP53) Heap Out-of-Bounds Write via Signed Integer Overflow in DWAA/DWAB LossyDctDecoder
- CVE Pending (GHSA-VP77-HHRC-M2C3) Subsampled chunk-height miscount: heap overflow in B44 encoder, NULL dereference in DWA decoder/encoder
- CVE Pending (GHSA-RXWG-X3WX-PGFH) Heap out-of-bounds write via 32-bit tile-area integer overflow in DeepTiledOutputFile (deep-image re-save path) â possible incomplete fix of PR #939
- CVE Pending (GHSA-RJG8-982X-JHC2) Uninitialized Heap Memory Disclosure in DWA Decompression via Short zlib Stream Mapped to Success
- CVE Pending (GHSA-J76P-H7GW-VMPF) Heap buffer overflow in the OpenEXR Python bindings when two RGB channel groups coalesce to the same key
- CVE Pending (GHSA-G248-596F-9995) Unvalidated extra
chlistheader attributes merged intoHeader::channels()cause a heap overflow inrun_fill - CVE Pending (GHSA-FRX8-5P4H-J8M7) Heap OOB Write via Integer Overflow in B44 Chunk Unpacked-Size Computation â Remote Code Execution in OpenEXRCore
- CVE Pending (GHSA-98RX-QPFV-QCWV) Missing grid-total bound in ZSTD deep decode causes heap out-of-bounds write in OpenEXRCore
- CVE Pending (GHSA-638X-X7MR-89XC) Out-of-bounds write at an attacker-chosen displacement in legacy Python InputFile.channel()
- CVE Pending (GHSA-6358-7H9P-8G5C) Heap buffer overflow in OpenEXRCore ZSTD deep decode via unvalidated signed sample counts
- CVE Pending (GHSA-3CP4-779M-H837) Use-after-free in OpenEXRCore exr_copy_unset_attributes() when a channel name exceeds 31 bytes
- (GHSA-RJG8-982X-JHC2) Uninitialized Heap Memory Disclosure in DWA Decompression via Short zlib Stream Mapped to Success
- (GHSA-X92J-9982-GJG7) DWA Encoder
tmpHalfBufferInteger Overflow â Undersized Allocation - (GHSA-HRP4-XJP5-CQ7J) Misaligned pointer store in
ht_undo_impl()(HTJ2K decode) - (GHSA-FP76-2FH4-4XWW) Uncontrolled Resource Consumption and O(N^2) Processing Overhead via Unbounded Multipart Header Parsing
- (GHSA-HCP5-MX7Q-9642) Null pointer read in the DWA decoder:
compute_sampled_heightcounts one row instead of zero when y sampling exceeds the chunk height
See CHANGES.md for more details.
Contributors
v3.4.17
Patch release that addresses the following security vulnerabilities:
- CVE Pending (GHSA-WHF5-29MJ-GP53) Heap Out-of-Bounds Write via Signed Integer Overflow in DWAA/DWAB
LossyDctDecoder - CVE Pending (GHSA-VP77-HHRC-M2C3) Subsampled chunk-height miscount: heap overflow in B44 encoder, NULL dereference in DWA decoder/encoder
- CVE Pending (GHSA-RXWG-X3WX-PGFH) Heap out-of-bounds write via 32-bit tile-area integer overflow in
DeepTiledOutputFile(deep-image re-save path) â possible incomplete fix of PR #939 - CVE Pending (GHSA-RJG8-982X-JHC2) Uninitialized Heap Memory Disclosure in DWA Decompression via Short zlib Stream Mapped to Success
- CVE Pending (GHSA-J76P-H7GW-VMPF) Heap buffer overflow in the OpenEXR Python bindings when two RGB channel groups coalesce to the same key
- CVE Pending (GHSA-G248-596F-9995) Unvalidated extra
chlistheader attributes merged intoHeader::channels()cause a heap overflow inrun_fill - CVE Pending (GHSA-FRX8-5P4H-J8M7) Heap OOB Write via Integer Overflow in B44 Chunk Unpacked-Size Computation â Remote Code Execution in OpenEXRCore
- CVE Pending (GHSA-638X-X7MR-89XC) Out-of-bounds write at an attacker-chosen displacement in legacy Python
InputFile.channel() - CVE Pending (GHSA-3CP4-779M-H837) Use-after-free in OpenEXRCore
exr_copy_unset_attributes()when a channel name exceeds 31 bytes
This version also addresses a regression introduced in v3.4.14 in reading deep images on big-endian hosts, due to a byte-order bug in the deep sample count table. Little-endian hosts are unaffected.
See CHANGES.md for more details.
Contributors
v3.3.15
Patch release that addresses the following security vulnerabilities:
- CVE Pending (GHSA-WHF5-29MJ-GP53) Heap Out-of-Bounds Write via Signed Integer Overflow in DWAA/DWAB
LossyDctDecoder - CVE Pending (GHSA-RXWG-X3WX-PGFH) Heap out-of-bounds write via 32-bit tile-area integer overflow in
DeepTiledOutputFile(deep-image re-save path) â possible incomplete fix of PR #939 - CVE Pending (GHSA-RJG8-982X-JHC2) Uninitialized Heap Memory Disclosure in DWA Decompression via Short zlib Stream Mapped to Success
- CVE Pending (GHSA-J76P-H7GW-VMPF) Heap buffer overflow in the OpenEXR Python bindings when two RGB channel groups coalesce to the same key
- CVE Pending (GHSA-G248-596F-9995) Unvalidated extra
chlistheader attributes merged intoHeader::channels()cause a heap overflow inrun_fill - CVE Pending (GHSA-FRX8-5P4H-J8M7) Heap OOB Write via Integer Overflow in B44 Chunk Unpacked-Size Computation â Remote Code Execution in OpenEXRCore
- CVE Pending (GHSA-638X-X7MR-89XC) Out-of-bounds write at an attacker-chosen displacement in legacy Python
InputFile.channel() - CVE Pending (GHSA-3CP4-779M-H837) Use-after-free in OpenEXRCore
exr_copy_unset_attributes()when a channel name exceeds 31 bytes
See CHANGES.md for more details.
Contributors
v3.5.2
Patch release that introduces the use of OpenJPH's built-in LUT to apply the decode/encode transfer function for lossy LJ2K compression.
Also in this release:
- Prebuilt wheels for Python 3.14.
- Fix for a bug when statically linking and libdeflate, zstd, or OpenJPH is found via pkg-config rather than CMake config (#2673).
- Pin the
IMATH_TAGtov3.2.3; it was previously incorrectly set tomain. - Pin the JPEG 2000 progression order to RPCL and adds TLM markers to allow efficient random access to resolution layers for J2K compression (#2672)
- Fix for a failure case in
exrcheckinvolving large images (#2688)
See CHANGES.md for more details.
Contributors
v3.4.16
v3.4.16 fixes a bug which created unreadable DWAB files in certain circumstances.
There is also a fix to prevent 'undefined behavior' warnings with the BytesAttribute
v3.5.1
Patch release that addresses a security vulnerability in ZSTD:
- fix heap buffer overflow in ZSTD deep scanline decoding
Also, the v3.5.0 release incorrectly identified the OPENEXR_VERSION_RELEASE_TYPE and associated OPENEXR_PACKAGE_NAME as -dev.
See CHANGES.md for more details.
Contributors
v3.5.0
-
Support for lossless compression via Zstandard
- Introduces
ZSTD_COMPRESSION/EXR_COMPRESSION_ZSTDenum - Features a custom data prep stage that includes ByteShuffle, optional Delta Encoding, and custom data reordering.
- Universal lossless support: Works for all image types (including Deep) and supports all OpenEXR pixel/data types.
- Deep image gains: Extensive testing with Deep scanlines shows a 20â30% size reduction compared to ZIPS based on image channels (with the best gains on Deep Alpha and Deep ID).
- Flat image file size: Achieves comparable compressed sizes to ZIPS.
- Flat image speed: Delivers slightly faster compression and decompression speeds compared to ZIPS
- Introduces a new dependency on the zstd library; a vendored version (v1.5.7) is provided as a fallback if no installed version is present.
- Introduces
-
Support for lossy compression based on the High Throughput JPEG 2000 (HTJ2K) standard.
- The LJ2K compressor uses the same JPEG 2000 constraints and codec library (OpenJPH) as HTJ2K256, but with lossy coding, producing smaller files at the cost of distortion.
- LJ2K should generally offer reduced file sizes and higher throughput compared to DWA.
- Currently supports 16-bit HALF, 32-bit FLOAT and 32-bit UINT channels. Lossy coding is currently applied only to RGB channels; other channels are losslessly coded.
- The amount of distortion is controlled via a quality level parameter that ranges from 1 to 150.
- As with DWA, lossy RGB channels are first transformed to a non-linear perceptual domain before coding. The transformation is log-based when the sample's magnitude is greater than 1, and power-law otherwise.
-
Improved support for the
colorInteropIDstandard attribute as a mechanism to identify the color space of the RGB images:- Added support for
colorInteropIDto command-line toolexrstdattr. - Added API functions to validate color metadata in the header to both the C++ and core APIs.
- Extended the exrinfo command-line tool to print warnings regarding header color metadata.
- Added helper functions to convert between chromaticities and
colorInteropID. - Added Python bindings for all new functions.
- See An ID for Color Interop
and Identifying the Color Space of OpenEXR Files for details.
- Added support for
-
Support for a cache for a (de)compression context in the core decode pipeline.
- This is a backwards compatible change (i.e. older versions of code compiled against previous headers will still work), with the exception that it does add the requirement that application code actually use the provided initializer when defining
exr_decode_pipeline_tobjects.
- This is a backwards compatible change (i.e. older versions of code compiled against previous headers will still work), with the exception that it does add the requirement that application code actually use the provided initializer when defining
-
For the OpenEXR python module:
- Support for the
idManifestattribute and object - Support for multithreaded read and write
- Support for reading/writing from/to
BytesIOobject - Add
setMaxImageSize/setMaxTileSizefunctions
- Support for the
-
Other changes:
AcesInputFile/AcesOutputFileandexr2acesas marked as deprecated and will be removed in a future release.- Improved handling of many channels
- All Imath headers are now included internally via
#include <Imath/...> - Remove
iex_debugTrap()function (was not used) - Improved
ChannelListIterator API - Fixed a bug that with DWA compression that could silently produce corrupted files.
-
Also of note:
- The vendored version of deflate is 1.26
- The vendored version of OpenJPH is 0.32.0
- The minimum required version of OpenJPH is 0.32.0
Security
This release addresses the following security vulnerabilities:
- OSS-Fuzz 514487287 Floating-point-exception in
part_exceeds_memory_limits - OSS-Fuzz 514423826 Divide-by-zero in
part_exceeds_memory_limits - OSS-Fuzz 513282267 Timeout in
openexr_exrgaps_fuzzer - OSS-Fuzz 512988066 Out-of-memory in
openexr_exrgaps_fuzzer
See CHANGES.md for more details.




























