curl-users
Re: curl http downloads corrupt zip files
Date: Fri, 22 Aug 2003 16:07:21 +0100
----- Original Message -----
From: Daniel Stenberg <daniel_at_haxx.se>
Date: Friday, August 22, 2003 11:27 am
Subject: Re: curl http downloads corrupt zip files
> On Fri, 22 Aug 2003, Matthew Burgess wrote:
>
> > I'm attempting to use the curl-7.10.7 win32-nossl binary to
> download some
> > self extracting zip files from a website. The command line is
> roughly:>
> > curl -s -O -u username:password
> http://my.url.com/path/to/some.exe (with
> > HTTP_PROXY set to enable access through the corporate firewall).
> >
> > Using "unzip -t" to verify the file is complete reports several
> errors which
> > are unable to be fixed using "zip -F". This seems similar to
> what may
> > happen with binary files transferred in Ascii mode over FTP, but
> I'm not
> > aware of any similar issues over HTTP. Incidentally,
> downloading this file
> > manually using IE on WinNT 4 SP6 downloads the file correctly.
>
> I just now tried this operation on a Windows XP machine:
>
> 1. I downloaded and unpacked the curl-7.10.7 win32-nossl package
> using my
> browser
> 2. Then used the brand new curl executable to download the curl-
> 7.10.7 source
> package zip archive from the curl web site.
> 3. I tried winzip'ing the newly downloaded zip archive. Worked
> just fine.
>
> > Any pointers to any settings that may be affecting this file
> transfer would
> > be greatly appreciated,
>
> Can you please provide us with a better receipe on how to repeat
> this problem?
Apologies for the inaccurate report. It now looks as if the files (or the web server) are the cause of the problem. The files were downloaded successfully via IE on the 11th August. The files HTTP headers suggest no modifications have been made since July 30th but downloading in both IE and with curl results in a corrupted file. The strange thing is some files download fine, others don't and with 340 files to grab I'd rather they all came down correctly first time. I even used wget with the --ignore-length option to ensure that there wasn't something screwy going on with the content-length header but that resulted in a corrupted download as well. I can now only assume that the problem is on the server side and not on my (3) clients! Unfortunately the files are not publicly available for anyone else to verify it's not just my bad luck!
Thanks and apologies again for the erroneous report,
Matt.
-------------------------------------------------------
This SF.net email is sponsored by: VM Ware
With VMware you can run multiple operating systems on a single machine.
WITHOUT REBOOTING! Mix Linux / Windows / Novell virtual machines
at the same time. Free trial click here:http://www.vmware.com/wl/offer/358/0
Received on 2003-08-23