 I still don't like Accept headers I just found out if you hit https://pypi.org/simple/pydantic/ with "Accept: application/vnd.pypi.simple.v1+json" you get back super-useful JSON about that package... but I can't link to a demo because I can't include the Accept header in a link! Here's a Gist instead: https://gist.github.com/simonw/8cf8a850739e2865cf3b9a74e6461b28 15 comments
@helge the only place I like them is for media files (audio, video, image) where you might want to serve back e.g. a smaller .webp file but only to browsers that support it For APIs they drive me up the wall, because I'm often doing API research from devices (like my phone) that don't have a convenient terminal to hand @helge I get the theory, but in twenty years of mucking around on the Web I've never actually seen real concrete benefit from APIs that do that - I've just found it frustrating @simon There's also a JSON API that doesn't require an accept header documented at https://warehouse.pypa.io/api-reference/json.html#get--pypi--project_name--json (e.g. https://pypi.org/pypi/pydantic/json) if you haven't seen that. (Not sure if this has the info you're looking for in the response.) @jonafato that's not the same information, sadly - I'm looking for the new "provenance" key @simon imo it's good practice for servers to have some kind of URL pattern that overrides the header and triggers equivalent behavior, for times when the URL author wants to do something particular. But in general, the whole purpose of 'em is for the URL author to get out of the way and let the client supply client-specific info (which the URL author *can't* know) to the server. @simon to me it all comes down to documentation: whether it’s the separate URLs to get non-HTML or the alternate media types Accept’d, docs are necessary. https://docs.pypi.org/api/integrity/#get-provenance-for-file @simon cc @jed about explicitly passing a format in the URL vs relying on a header cc @calpaterson too who relies on it in csvbase  @simon http content-negotiation sucks. It always gets in the way. Many thanks to the github people who showed how to do good url design. |
Gregory's Server
@simon I love accept headers, they get you the most suitable response the client can process. If you want to spoof them, you can easily do that using a proxy?