Attendance:

Discussion:

• Revisiting plans for 3.2 release:

•  

  • Email with Nick Cannon implied it’s no big deal, just send a message to the group, but expectations were tentative anyway.

  • Cary will send a message to VES committee, but Kimball will draft it first and circulate for review

  • Should hint that we’ll backport the faster reading operation in a ABI compatible way

• Larry: Somebody at Sony was using an obscure header and noticed some stray references to half.h w/no directory. Will encourage her to submit a PR.

• Peter: A fuzz issue is still open: fast half decode in Core library. Kimball: Oops, didn’t see that one.

• CII Silver/Gold requirements:

•  

  • Cary: Many items are already met. Potential thorn is test coverage.

  •  

    • Larry: I recently worked on improving OSL coverage. Started at 70%, just broke 80%, would be really hard to get to 90%. OIIO is struggling to get to 70%.

    • Larry: OpenEXR is no longer passing the SonarCloud gates:

    •  

      • https://sonarcloud.io/project/overview?id=AcademySoftwareFoundation_openexr

      • https://sonarcloud.io/summary/overall?id=AcademySoftwareFoundation_openexr

    • Kimball: I had to generate memory failures

    • Peter: We have several unreachable error handles

• Project Update for the TAC is coming November 30:

•  

  • Cary: We struggle with Windows support, and addressing test failures on machine setups we don’t have access to

  • Kimball: I have a Windows laptop for the first time in 20 years!

  • Cary: We need to do a better job of grooming the next generation of developers/maintainers

  •  

    • Larry: The best way to attract new people is to do exciting things. Nvidia expanding direct loading of GPU.

    • Kimball: Support bfloat16

    • Larry: I’d love to have support for 16-bit ints

    • Kimball: We have to care how that’s encoded. Could allow only certain encodings.

    • Peter: could we get away with not needing it by clever compression