Add TSC Meeting notes from 04-05-2022.

Signed-off-by: Andre Pradhana <[email protected]>

Author: apradhana

tsc/meetings/2022-04-05.md

@@ -0,0 +1,128 @@
+Minutes from 128-th OpenVDB TSC meeting, April 5th, 2022, (EDT)
+
+Attendees: *Ken* M., *Jeff* L., *Nick* A., *Rich* J., *Andre* P.
+
+Additional Attendees: Kyle Wardlow, Karl Marrett (UCLA), Bruce Cherniak (Intel)
+Sebastian Gaida, Greg Hurst (United Therapeutics), J. T. Nelson (Blender)
+
+Regrets: *Dan* B.
+
+Agenda:
+
+1) Confirm quorum
+2) Secretary
+3) Forum
+4) Half Implementation
+5) PR-1347: Meeting Notes
+6) PR-1343: Speed up Windows CI
+7) PR-1342: MinGW fix
+8) PR-1339: Meeting notes and consolidate pending-changes
+9) PR-1336: Exposed TypeLists of the default set of OpenVDB types
+10) PR-1333: Small CMake change
+11) PR-1008: Rasterization
+12) PR-1317: openvdb_tool
+13) OpenVDBLink: Link between Mathematica and OpenVDB
+14) USD
+15) Next Meeting
+
+--------------------
+
+1) Confirm quorum
+
+Quorum is present.
+
+2) Secretary
+
+Secretary is Andre Pradhana.
+
+3) Forum
+
+a) Windows static build
+
+Several Windows users have asked for a better CMake support, see e.g. Issue
+[1160](https://github.com/AcademySoftwareFoundation/openvdb/issues/1160).
+
+We need to improve the CMake support especially for a better use of OpenVDB
+downstream. The CMake workflow should allow a user to easily bring VDB
+library, its dependencies, and compiled-flags; which all depend on the
+platform. This is primarily a Windows problem because Windows requires more
+specific branching in CMake. We need a custom find-modules in our CMake and
+rewrite the findPackage logic, e.g. findOpenVDB.cmake should not exist. This
+can be replaced by an OpenVDBconfig.cmake file (config module) instead.
+
+Ken asks if distributing binaries actually solves the problem. Nick thinks that
+it can alleviate the problem, but not solve it.
+
+Ken will bring this issue to the TAC meeting to ask for from other ASWF
+projects on April 20th 2022.
+
+b) Post in the forum about building VDB 9.0 with embedded half.
+
+Jeff said that the user included both openvdb.lib and libopenvdb.lib, which
+causes the problem.
+
+4) Half Implementation
+
+Ken is going to ask Autodesk about their requirements for a half-grid
+implementation. Greg Hurst plans to spear-head the effort once he is done
+with Mathematica-OpenVDBLink PR.
+
+5) PR-1347: Meeting Notes
+
+Somebody will take a look.
+
+6) PR-1343: Speed up Windows CI
+
+Adding Caching support for vcpckg installs, which should speed up the Windows
+CI. Nick and Jeff are unsure about store password logic that the PR includes.
+It would be nice if Dan takes a look.
+
+7) PR-1342: MinGW fix
+
+Nick is happy with this PR. Jeff approves.
+
+8) PR-1339: Meeting notes and consolidate pending-changes
+
+Merged.
+
+9) PR-1336: Exposed TypeLists of the default set of OpenVDB types
+
+This allows a user to do static-dispatch using a pre-approved list of grids
+to be supported by an algorithm at compile time. At runtime, however, it still
+does a string comparison. We could add an enum to replace the string
+comparison logic.
+
+We are happy to remove the Native keyword from the type-lists.
+
+10) PR-1333: Small CMake change
+
+Modernize CMake: change find_package(PythonLibs) to find_package(Python).
+
+11) PR-1008: Rasterization
+
+Rasterization PR from Nick. We have talked about approving both rasterization
+PRs from Nick and Dan. Andre approves the PR.
+
+12) PR-1317: openvdb_tool
+
+Ken is pushing changes, including CMake. The PR is ready for some feedback.
+
+13) OpenVDBLink: Link between Mathematica and OpenVDB
+
+Greg will create a new branch containing OpenVDBLink, which provides a link
+between OpenVDB and Wolfram Mathematica, and will create a PR based on it.
+
+Greg ask about an advection example. Andre suggests to use Enright field to
+advect a density field. Nick suggests to advect points with a curl noise.
+Ken will help Greg.
+
+14) USD
+
+Bruce Cherniak raises awareness on an USD PR to [support OpenVDB without
+depending on OpenEXR](https://github.com/PixarAnimationStudios/USD/pull/1728).
+
+The PR is from NVIDIA Gameworks. Nick and Ken will take a look.
+
+15) Next Meeting
+
+Next meeting is in one week, April 12th, 2022. 1pm-2pm EST (GMT-5).