You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I propose that the resolution is to search for the provided mime_type within the map using mime_types.find. If it is not found, a warning should be logged, but the content should still be delivered using the following logic:
Does the provided type look like a MIME type already? I.e. is it one of the IANA-registered types (application, audio, example, font, image, model, text, video), followed by a slash, followed by text? If so, send that verbatim.
Otherwise send text/plain.
Although this approach isn't quite as lightweight as simply sending text/plain, there are some benefits:
I'd be able, for example, to send content as text/csv without modifying the mime types map.
It would also allow people to use the full mime-type format, including parameters (e.g. text/plain;charset=UTF-8)
Developers coming from other technologies that use full mime types won't be faced with an internal server error for following an old habit.
I fully agree, the current solution is way to limited. If you need to construct custom mime-types now, you can always set the Content-Type header yourself. But this feels like a workaround.
Jake, if you want, you can implement the fix youself and make your first contribution to Crow! 🔥
Activity
dranikpg commentedon Sep 10, 2022
I fully agree, the current solution is way to limited. If you need to construct custom mime-types now, you can always set the
Content-Typeheader yourself. But this feels like a workaround.Jake, if you want, you can implement the fix youself and make your first contribution to Crow! 🔥
Else, I'll handle this in the following week.
Address #534 - handling missing mime types gracefully (#536)