I do think that they are missing a fairly simple solution
That doesn't sound very simple at all. For one thing, it means the python runtime loses control over the http implementation it ships. It's also not obvious that libcurl provides an API that can do everything something like http.client can do. I think this is somewhere between 'not simple' and 'non-starter'.
That doesn't sound very simple at all. For one thing, it means the python runtime loses control over the http implementation it ships. It's also not obvious that libcurl provides an API that can do everything something like http.client can do. I think this is somewhere between 'not simple' and 'non-starter'.