Attendance:

Guests:

• Pierre-Anthony Lemieux, Sandflow Consulting

• Michael Smith, Intel / Wavelet Consulting

• Li Ji, Industrial Light & Magic

Discussion:

Quick Recap

The team discussed the upcoming release timeline and versioning decisions for the imath library, with Cary explaining the need for a patch release to fix a bug related to undefined behavior. They addressed documentation timing for a codec release, deciding to proceed with posting documentation while clearly indicating it's for version 3.4 and beyond. Pierre-Anthony confirmed there were no outstanding issues on his side and discussed his presentation for the Round Table, deciding to use a static screenshot instead of a slide deck.

Next Steps

• Cary to push out the release with the foundry fix tomorrow

• Cary to follow up with Kimball regarding further changes needed before the release

• Cary to make a new patch release for Imath with the fix to the index operator

• Pierre-Anthony to present at the Round Table meeting using his own slide deck

• Pierre-Anthony to submit updated wording for the documentation of the lossless compression feature

• Pierre-Anthony to create a draft PR for the documentation update and mark it as not yet in an official release

Summary

Foundry Fix and Release Planning:Cary discussed the upcoming release and mentioned that Kimball cannot make it. He talked about the foundry fix that was staged and planned to push it out the next day. Cary emphasized the importance of targeting the end of August for the release, short of resolving the issue with the Call our interop id. He also noted that a new release of Imath is needed to fix the index operator issue and allow the new open exr release to rely on it.

Patch Release for Union Bug Fix:Cary, Larry, and Peter discussed versioning issues in the imath library, focusing on whether a release should be a patch or major version update. They agreed that since the change was inline and did not require source code changes on the application side, it was acceptable to label it as a patch release. Cary explained that the bug fixed in the new version was related to undefined behavior involving union usage and memory indexing, which caused incorrect values in some compiler scenarios. Peter confirmed that the fix would only benefit applications that were previously affected by the bug, and Larry noted that the change was more of a build-time implementation rather than a syntax change.

Minor Release and Presentation Planning:Cary and Peter discussed a minor release and the need for recompilation, noting that users only need to recompile if they want the fix. They decided that a patch might be easier for people to adopt, given the potential pain of a major version change. Cary mentioned waiting for the Open Color IO folks to resolve a color interop ID comment. Pierre-Anthony confirmed there was nothing outstanding on his side and discussed his presentation for the Round Table, deciding to use a static screenshot instead of a slide deck. Cary confirmed the presentation would be posted on YouTube.

Codec Documentation Release Timing:The team discussed documentation timing for a codec release. Cary explained that merging the documentation PR (2052) should align with the release to avoid website updates before the codec is officially available. Pierre-Anthony agreed to submit updated documentation wording, with Peter suggesting to mark it as "upcoming 3.4" and remove once the release is official. The team decided to proceed with posting the documentation while clearly indicating it's for version 3.4 and beyond.