René's URL Explorer Experiment

{"@context":"https://schema.org","@type":"DiscussionForumPosting","headline":"[FEA]: `runtime.getLocalRuntimeVersion()` → `path_finder`","articleBody":"### Is this a duplicate?\n\n- [x] I confirmed there appear to be no [duplicate issues](https://github.com/NVIDIA/cuda-python/issues) for this request and that I agree to the [Code of Conduct](CODE_OF_CONDUCT.md)\n\n### Area\n\ncuda.bindings\n\n### Is your feature request related to a problem? Please describe.\n\nQuoting from a chat:\n> Someone, somewhere, will want to know the \"real\" cudart version and not what was used to build cuda-python.\n\nCurrently `runtime.getLocalRuntimeVersion()` is implemented using a direct `dlopen()` call:\n\n* https://github.com/NVIDIA/cuda-python/blob/1cb8a6fa79b4c6be95634e852d303fe11883ec5d/cuda_bindings/cuda/bindings/cyruntime.pyx.in#L1887\n\n\n### Describe the solution you'd like\n\nThe direct `dlopen()` call should be replaced with a call to `path_finder`, similar to existing calls, e.g. here:\n\n* https://github.com/NVIDIA/cuda-python/blob/314060c0f75aaf9ae0b3742a7a7485021f46b877/cuda_bindings/cuda/bindings/_bindings/cynvrtc.pyx.in#L52-L55\n\n### Describe alternatives you've considered\n\n_No response_\n\n### Additional context\n\n_No response_","author":{"url":"https://github.com/rwgk","@type":"Person","name":"rwgk"},"datePublished":"2025-06-10T21:42:29Z","url":"https://github.com/NVIDIA/cuda-python/issues/702"}