← Back to context

Comment by hk__2

14 years ago

Chrome does the same, when you right-clic on a file and select "Save as…", the download begin in a temporary file, and then when you select the emplacement where you want to put the new file, it moves the temporary file in it.

7 comments

hk__2

Reply
masklinn  14 years ago

I don't know a browser which does not do that. They need to start the download to get the metadata (file name and type), and they'll do it if only because shitty servers may react very badly to the browser just stopping with a full socket buffer.

  • skeletonjelly  14 years ago

    They don't need to start it to get the metadata. All that's sent is the content-disposition header with the suggested file name.

    • masklinn  14 years ago

      > They don't need to start it to get the metadata. All that's sent is the content-disposition header with the suggested file name.

      That is the start, unless you perform a HEAD followed by a GET (which takes 2 round-trips instead of one and may not even be correctly supported). And because any link can lead to that, every single link you click would have to be implemented via 2 round-trips, that's complete madness.

      > All that's sent is the content-disposition header with the suggested file name.

      Nonsense, there is no situation in which "all that's sent is the content-disposition header", that header is part of a full header section and immediately precedes the content itself.

      1 reply →