Merge pull request #1882 from sideeffects/sendupstream_tscnotes_20240827

Meeting notes of August 27, 2024.

Author: jmlait

tsc/meetings/2024-08-27.md

@@ -0,0 +1,117 @@
+Minutes from OpenVDB TSC meeting, August 27th, 2024
+
+Attendees: *Jeff* L., *Greg* H., *Rich* J, *Ken* M., *Andre* P., *Dan* B.
+
+Regrets: *Nick* A., *Ken* M.
+
+Additional Attendees:
+Dhruv Govil (Apple), Jonathan Swartz (NVIDIA), Barry Dempsey
+
+Agenda:
+
+1) Confirm quorum
+2) Secretary
+3) VDB 12 Release Items
+4) License Changes
+5) Maintainer List
+6) CI Issues
+7) PRs to merge
+8) Next meeting
+
+------------
+
+1) Confirm quorum
+
+Quorum is present.
+
+2) Secretary
+
+Secretary is Jeff Lait.
+
+3) VDB 12 Release Items
+
+a) Half support for v12
+
+Should half grid be a REAL grid?  Ie, should the macros for REAL grid
+type include it?  Conclusion was they should not.  Instead a
+higher-level named macro should contain Half.
+
+With this decided it is a v12 candidate.  Proxy grid test is
+failing, but other than that seems good.  The problem is we can't
+instantiate a proxy tree at the moment.
+
+b) fVDB
+
+Will be a feature branch.   Can we have a cadence separate from main VDB?
+Can binaries be released on a different place, like Anaconda?  
+
+NanoVDB has a separate release system.  
+
+Tags can be created on branches to mark a release point.
+
+This would require the fVDB release to be based against a stable
+version of VDB.  
+
+Currently NanoVDB and fVDB are tightly coupled - improvements to
+NanoVDB are coming from fVDB.  This suggests NanoVDB needs to match
+fVDB.
+
+4) License Changes
+
+Still trying to set up the second repo to verify CLAs.
+
+5) Maintainer List
+
+Unanimous for changing Committer to Maintainer
+
+Unanimous that it is two maintainers to commit.
+
+Unanimous that all current TSC member are also maintainers.
+
+Unanimous that Jonathan Swartz to be added as a maintainer.
+
+Jeff will attempt to update the relevant policy docs.
+
+6) CI Issues
+
+No good tricks to speed up CI debugging. Can sometimes turn off
+unnecessary runners, but some runners just take a long time until the
+first possible error.
+
+The secret is for the Houdini download.
+
+cudann download seemed to be failing.  This is likely not from our
+runner side.
+
+Need notes for the PR for CI to explain why some of the unusual
+solutions are present.
+
+Why certain versions of clang?  This is likely to try to match the g++
+of the VFX platform, but we are not sure.  The docker images came from
+the ASWF.  We have clang at all to catch more issues than g++ alone.
+
+VDB12 will likely drop support for gcc9.
+
+7) PRs to merge
+
+a) 744 - remove boost any
+
+Someone must approve by next meeting or it will be approved then.
+
+b) 1723 - boost conversion traits
+
+Waiting on the Half PR?  We should take it out of the Half PR so this
+can get in, Half PR can rebase on top of the resulting change.
+
+c) 1789, 1775 - Makes adaptive grid easier
+
+1789 is needed so you can use the same Grid with different custom Trees.
+
+1775 explores all possible instantiations and ensures they are handled.
+
+These are both on schedule to be decided on by the next meeting.
+
+8) Next Meeting
+
+Next meeting is on September 10th, 2024. 2pm-3pm EDT (GMT-4).
+