Attendance:

Discussion:

• TAC project update:

•  

  • Mention Imath

  • Example images

  •  

    • Larry: We don’t have a modern production image with the metadata you’d expect.

    • Larry: I come across EXR images with godawful numbers of channels or parts

    • Joseph: I have some Alexa image that was crafted to use all 17 stops

    • Kimball: CG torture test 17 depth channels and 43 AOV’s

    • Peter:  should update the deep stuff. A production-level complexity deep file

    • Can serve as a benchmark

    • Peter: some standard images around useful to talk to vendors

    • Joseph: I sent out a communication to marketing people about some sample images

    • Joseph: Need an updated photo of Rod’s office.

  • Cary: should mention outreach to maintainers and alternate bindings projects

  •  

    • Joseph: frustration that macports people don’t even respond

    • Peter: redhat/suse people reached out about a problem after a release

    • Kimball: some kind of beta with them

    • Larry: ideally they see our release-staging issues.

    • Kimball: OpenEXR wrappings, or just OIIO?

    • Larry: not having a python binding is not as big a problem as that there’s a lame one out there.

    • Larry: It’s the decoupling of the efforts that causes the friction

    • Nick: Invite them to come by and talk 

    • Larry: Before, there wasn’t the option

    • PyOpenEXR

• Threading crash on Windows shutdown: 

•  

  • Kimball: I’m solving the Rubix cube blind, hoping to get the fix for random crash

  • Give it another week or so.

• Larry’s proposed API extension PR:

•  

  • Larry: it deserves much more of a discussion about what we want a stateless API to be. Needs to consider the deep, multipart.

• Next release: 

•  

  • Kimball: the C++ library doesn’t link against the core.

  • Peter: not a drop-in replacement

  • Cary: Plan for a 3.2 or 4.0

  • Nuke does link against the core