112405 - [RFE] Option for user-specified content-type

Status: RESOLVED WONTFIX

(Core :: Networking: HTTP, enhancement)

Description

[Chris Brien]

In bug 112361, someone asked for Mozilla to ignore content-type. However, this
would imply that Mozilla should know better. For example an html file could be
rendered as text/plain if the server said so.

But not all servers are configured correctly.

It could be impossible to view a page as HTML if the server says it is
text/plain. Worse, an image that was for example image/jpeg, could be rendered
as image/x-icon. The user would see a broken image, and not understand why.

It would be nice if, after loading a file, the content-type could be set by the
user. A text/plain file could be viewed as text/html, or an HTML file could be
viewed as text/plain (this would be a better view-source implementation, killing
bug 76816). Of course a correct implemetation would be dependent on bug 55583.

Comment 1

[Matthias Versen [:Matti]]

-> Networking:http
(i vote for a wontfix)

Comment 2

[Boris Zbarsky [:bzbarsky]]

> Worse, an image that was for example image/jpeg, could be rendered
> as image/x-icon.

No.  Imagelib already looks at the image data itself to determine the image
type.  Apparently the number of servers that send anything resembling correct
types for images is rather lower than one would wish...

ccing rpotts... this would require a way to load something with an override mime
type passed in that then gets set on the channel... echhh.

Comment 3

[Darin Fisher]

-> future

Comment 4

[timeless]

i'd love to be able to have a dropdown next to the urlbar which would let me 
see what the current live mime type is, and modify it to force something to 
happen.

Comment 5

[Darin Fisher]

-> default owner