On Tue, 15 Apr 2008, Prasad J Pandit wrote:

> I tried to post the same data(which is > 1500 bytes), as my program, using
> curl(1) command, and this is what I receieved:
>
> ==
> curl: (18) transfer closed with outstanding read data remaining

That means libcurl didn't get a complete chunked encoding stream.

> Does libcurl handles `chuncked' reply properly?

Yes, AFAIK. We have numerous test cases for it and people use it extensively.

> Second thing, when I used $http_proxy="http://127.0.0.1:3128/", curl(1)
> could/did not read the data file. All the following commands would just
> not exit and go on forever, and the server did not receive the POST data.

[...]

> I had to do `$ export -n http_proxy', and only then I got the above (18)
> reply.

Uh, what exactly is the difference here? What's the first method really about?

> Strange may it sound, but the folllowing command worked perfectly fine, and
> gave me the proper reply of [true], that too without fiddling with the
> $http_proxy
>
> $ wget --post-file=/tmp/post-data http://172.16.1.1:5666/test

It automatically found your proxy without you telling it about it? I wasn't
aware wget had such abilities.

> Again, don't tell me to go use something else. I want/expect the proper
> reason, for all this.

You can "expect" answers the day you start paying for them. Until then you
should remain friendly and polite and ask for help, not demand it.

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