Closed kmuseth closed 2 months ago
@apradhana - so other than the fact that the CI is broken you're fine if I merge this into master?
Hi, @kmuseth, yes, you can merge it to master. These are my reasons:
nanovdb-lite
and Windows-dynamic
CI were passing, (3) I'm working on top of the CI fixes that @Idclip has been working on (PR-1837) to fix the CI issues [see this PR]. As far as I can tell, the problems are not in your PR, but either in the CI runner or in CentOS 7.updateFiles.sh
and updateFiles.py
to aid developers make the necessary changes. That should cover all use cases. This also allows fVDB
to use the NanoVDB
version in the main repository, which should help us to discover any bug we may have missed.One thing to note after this change is that we need to keep in mind that Util.h
should always be at the bottom of the NanoVDB
dependency tree, followed by Math.h
, followed by NanoVDB.h
.
For completeness, I tested the NanoVDB version that comes from this PR along with the CI work that is currently in progress here and it shows that the failing CIs are the windows-static builds.
Maybe fix the CI first if you're doing that anyway
Introducing NanoVDB v32.7.0, which is a major refactoring of v32.6, primarily related to the introduction of new namespaces and an accompanying restructuring of the directory layout of header files. This will likely require changes to client code! To this end I have included a shell script (
cmd/updateFiles.sh
) that will scan files and make (most) of the required changes.The motivation for there these changes is twofold: First we want to use the same practice in terms of name spaces as OpenVDB and second the old version (32.6) was getting cluttered and messy. Specifically the main changes are:
nanovdb
as a namespacemath
,tools
,cuda
,io
Here are some examples of changes: