René's URL Explorer Experiment

{"@context":"https://schema.org","@type":"DiscussionForumPosting","headline":"[BUG] Inner lists in `cuda.bindings` API reference are dropped","articleBody":"Taking NVRTC's `--gpu-architecture` option as example:\n- This is the generated docs (by parsing Doxygen docstrings with an internal tool)\n\n\u003cimg width=\"750\" height=\"117\" alt=\"Image\" src=\"https://github.com/user-attachments/assets/c0051e72-cba5-4d87-a7cb-72d4e9820a7f\" />\n\n- This is the public docs (the list is too long, so I only show a portion of it)\n\n\u003cimg width=\"748\" height=\"401\" alt=\"Image\" src=\"https://github.com/user-attachments/assets/27c052ae-2daf-4017-9e0e-3dba15b5edc9\" />\n\n- This is the docstring from `nvrtc.h` (also truncated):\n```C\n * The valid compiler options are:\n *\n *   - Compilation targets\n *     - \\c --gpu-architecture=\\\u003carch\\> (\\c -arch)\n * \n *       Specify the name of the class of GPU architectures for which the\n *       input must be compiled.\\n\n *       - Valid \u003cc>\\\u003carch\\>\u003c/c>s:\n *         - \\c compute_50\n *         - \\c compute_52\n *         - \\c compute_53\n *         - \\c compute_60\n *         - \\c compute_61\n *         - \\c compute_62\n *         - \\c compute_70\n *         - \\c compute_72\n *         - \\c compute_75\n *         - \\c compute_80\n *         - \\c compute_86\n *         - \\c compute_87\n *         - \\c compute_89\n```\n\nWe can see the inner lists are not parsed. ","author":{"url":"https://github.com/leofang","@type":"Person","name":"leofang"},"datePublished":"2025-07-21T16:59:04Z","url":"https://github.com/NVIDIA/cuda-python/issues/770"}