Brendan

> in that every entity is documented on a separate page. I have been browsing the tensorstore docs, and I found it confusing when I click on a cross-reference and...

Your example does make sense for using a separate page. To be clear, I was confused at the structure of the TOC. It looks like each TOC entry is specific...

I've been looking into the [libclang python wrapper](https://pypi.org/project/libclang). There is a [types-clang](https://github.com/tgockel/types-clang) stub lib for it also. I think the stubs are maintained by the LLVM team, but not all...

I have written a test script that strips out all comment like syntax: ```py def strip_comment(cmt: str = None) -> Optional[str]: """ :param cmt: the `Cursor.raw_comment` property. :returns: A list...

I found a similar implementation for autodoc-ing C code using libclang... It monkey patches the libclang python binding to expose methods that allow using clang to parse the comments and...

I'm working in a separate repo (local only - nothing's been pushed). I'm trying to write a replacement for breathe, but there are things that doxygen does and clang doesn't;...

I started exploring the XML approach because the doxygen commands can have nested paragraphs (using `@par` and `@parblock ... @endparblock`) within a command that is designed to only accept a...

Doxygen's grouping is probably one of the worst implemented features it has. I quick tested this doctring ```cpp /** * @ingroup group2 * @brief class C3 in group 2 */...

I don't intend to support `@page` (and the like) commands because that is specific to Doxygen's ill-conceived static site generator (🤮).

> To be clear --- is that XML output from clang or from doxygen? I'm not running Doxygen at all. Its all clang output (using the monkey patch I referenced)...