Jeremy Maitin-Shepard

I think I could take care of the merging (especially for search) but do you think you could take care of implementing the new features that require separate sphinx/python integration,...

The new features being implemented wouldn't need to block the merge but it would be nice not to lose track of them.

To be clear, we are observing about 50% of the CPU time spent in calls to recv in profiles. Since ultimately this is just doing a memcpy from some kernel...

See also https://github.com/google/tensorstore/issues/77 We would like to have them, but it is challenging because we currently build on Github-hosted runners which do not include a linux arm64 runner, and building...

Yes we were planning to move to manylinux_2_28 anyway for C++20 support. But is there a manylinux_2_28 cross-compilation image with gcc 12?

Looks good -- I'm okay with this as is, but I haven't see the term "showcase" used for this before. Not sure I have a better term though.

We may need to investigate exactly how the previous state is reloaded when sphinx does an incremental build. cpp apigen stuffs some extra attributes into `app.env`, and then define a...

> Well I was able to use the file content (in bytes) as the hash seed, and it did prevent writing new files when the content hasn't changed (for both...

@AA-Turner @jakobandersen Since there seems to be renewed activity on the C++ domain, if I rebase this to HEAD would you have time to look at it soon?

@jakobandersen > The idea is good, I'll try to review it as soon as possible. Okay, I rebased it.