Re: [Bug-wget] Problem with headers getting put into the file

[Tim Ruehsen]

On Tuesday 06 January 2015 05:37:49 Nuzhna Pomoshch wrote:
> Hi,
> 
> I have been using wget for many years, and have just recently begun to
> encounter a strange problem.
> 
> I typically do "wget -S http:/path.to/some.file -O local.filename", which
> has always worked fine in the past.
> 
> On some sites now, the headers are getting put into the beginning of the
> output file.
> 
> A typical set of those headers (from the saved file) is:
> 
> HTTP/1.1 200 OK
> Server: nginx
> Date: Thui, 01 Jan 2015 00:00:00 GMT
> Content-Type: application/force-download
> Content-Length: 1073741824
> Last-Modified: Thu, 25 Dec 2014 00:00:00 GMT
> Connection: keep-alive
> Content-Disposition: attachment; filename="some.file"
> ETag: "00000000-00000000"
> Accept-Ranges: bytes
> 
> I am wondering if the mime type "application/force-download" isn't causing
> the problem.
> 
> This is unpleasant at best (although I can usually remove the headers with
> dd). The big problem comes when the download is interrupted, and I try to
> resume it. When that happens, the partial range requested doesn't match
> what is on the disk (it is off by the size of the headers at the beginning
> of the file), and the file (most often a large media file) gets corrupted
> (bytes are missing from the middle).
> 
> Has anyone encountered this before and does anyone have any thoughts on how
> to resolve this?

It looks like the web server is misconfigured and sends an additional HTTP 
header within the response body.

BTW, application/force-download seems to be a hack (that some browsers 
support). see 
http://stackoverflow.com/questions/10615797/utility-of-http-header-content-type-application-force-download-for-mobile

If you give me your real URL and I can investigate to patch Wget to support 
this hacky mime-type.

Tim

signature.asc
Description: This is a digitally signed message part.