On Sun, 3 Dec 2006, Edda T. Littlefield wrote:

> I've been trying to set up curl to download a web pages, and files off an
> FTP server. I have spent a day trying various command line options but can
> only get the results below:
>
> == Info: About to connect() to ftp.bom.gov.au:21
> == Info: Connected to ftp.bom.gov.au (134.178.63.130) port 21
> <= Recv header, 55 bytes (0x37)
> 0000: 220-Welcome to the Bureau of Meteorology FTP service.
> <= Recv header, 6 bytes (0x6)
> 0000: 220
> => Send header, 16 bytes (0x10)
> 0000: USER anonymous
> == Info: Connection aborted
> == Info: Closing connection #0

Wow, that's... very weird behaviour!

> 0010: User-Agent: curl/7.10.5 (i386-pc-msdosdjgpp) libcurl/7.10.5 Open
> 0050: SSL/0.9.7c zlib/1.1.4
> 0067: Host: www.google.com.au
> 0080: Pragma: no-cache
> 0092: Accept: image/gif, image/x-xbitmap, image/jpeg, image/pjpeg, */*
> 00d4:
> == Info: Empty reply from server
> == Info: Connection #0 left intact

This too seems to indicate something very strange going on.

Both these sites work fine with a recent curl on Linux.

> As you can see the version is 7.10.5 DOS port. I am running it on an old
> '486 machine with a NE2000 compatible NIC connected to a broadband
> modem/router. On the same machine I have Lynx and Arachne brousers working
> perfectly. My plan is to automate the downloading of data while I'm absent
> for an extended period. Any suggestions greatly appreciated.

I have no idea. I always recommend using an as new version as possible and it
that doesn't work either, use a debugger to figure out the problems and adjust
accordingly. The 7.10.5 version is very old.

I think the number of DOS users running curl are extremely few, so don't
expect to get a lot of advice from fellow users...

-- 
  Commercial curl and libcurl Technical Support: http://haxx.se/curl.html