Re: curl: (18) transfer closed with 1 bytes remaining to read
Date: Fri, 4 Jun 2004 11:13:55 +0200 (CEST)
On Thu, 3 Jun 2004, Christian Bevacqua wrote:
The URL doesn't work for me atm.
> curl: (18) transfer closed with 1 bytes remaining to read
> This exact problem seems to occur with most attempts to download a file from
> archive.org, so i suspect their web server is configured to act in this way.
> How cann i get around it?
Try asking for it with HTTP 1.0, it might act differently then.
> This may be helpful: I first tried the download via the Mozilla browser,
> and received all but the last byte (same as using curl). Same for lynx.
> However, when i attempt the download via my Safari browser, the whole file
> is downloaded successfully! So that leads me to believe there is some
> config thing i can send to make it work via curl.
It is very hard to guess what the reason for this might be, based on this
little amount of actual facts.
If you fire up your ethereal (or similar) and snoop on the Safari
requests/transfer it might offer some clues.
Also, try using curl's --trace-ascii option to get a full and detailed trace
of what the server responds to fully understand why this happens.
HTTP is a standard, but the world is full of servers and clients that violate
the standard. curl also does its share of standard deviations in its attempts
to mimic what browsers do. I have no idea how conformant, strict or liberal
Safari is in these aspects..
-- Daniel Stenberg -- http://curl.haxx.se -- http://daniel.haxx.se Dedicated custom curl help for hire: http://haxx.se/curl.htmlReceived on 2004-06-04