Jeremy Maitin-Shepard

Thanks for investigating this so thoroughly! We can probably ensure that tensorstore includes the uncompressed size in the header in this case, but in general there could be multiple variable-output-size...

For zarr v3 the pledged size is specified but not for zarr v2. It might be fixed by a later refactor of the zarr v2 codec handling.

All `const` methods are thread-safe, non-`const` methods are not thread safe. All `const` uses of a `TensorStore` object (which is almost all uses) are thread-safe, for example. Most objects in...

I didn't think about this issue previously, but the way that it can be made to work is to first preprocess with `-E -C -dD` which will preserve macro definitions...

I intentionally excluded inline since I considered it an implementation detail that doesn't belong in the documentation: https://github.com/jbms/sphinx-immaterial/blob/3c8fe16a499407a9a9b71b7dd2133c559cdccf95/sphinx_immaterial/apidoc/cpp/api_parser.py#L1183 However, I suppose we could add an option to control that behavior.

It would be great to expose the MRU timestamps of each window over IPC --- then the actual switching logic can be implemented externally with greater flexibility, e.g. separate commands...

There is a related issue: functions created with cpp_function are not picklable. Note that `PyCFunction` does implement `__reduce__`, but then pickling of the `PyCapsule` object fails. This could be solved...

Note: The issue from https://github.com/sphinx-doc/sphinx/issues/10785 also applies to `py:type` directives --- the default Python type xref uses the `class` role which currently does not match `py:type` objects. The easiest solution...

@mmatous It would be great to move this forward --- are you still interested in working on this? I started my own monkey-patching implementation of this in the sphinx-immaterial theme...

> @jbms would you have time to open a rebased PR? > > A Yeah I'm actually working on rebasing and addressing other comments now.