Which Header Lines Are Required?

[Gary Beene]

This minimal code (only the initial line in the header - no other header lines) works to download a small file, but I wondered if there were any other header lines in the request that I should add? Perhaps I was just lucky in having only the initial line?

I looked on the web but couldn't find a discussion on which lines are required. For all I know, it may be also considered impolite if I don't include certain headers line?

   Tcp Open "HTTP" AT "www.garybeene.com" AS #1 TIMEOUT 60000
   Tcp Print #1, "GET  /files/tcp_test.txt  HTTP/1.0"
   Tcp Print #1, ""
   Tcp Recv #1, 4096, Buffer$
   Tcp Close #1

If there's a reference to read, that would be nice to know as well.

[Stuart McLachlan]

You just require the one REQUEST line, but you can include other Header information if so desired.

For HTTP 1.0 which you are using , the
prime reference is RFC 1945
http://www.w3.org/Protocols/rfc1945/rfc1945

for HTTP 1.1, the reference is RFC 2616
http://www.w3.org/Protocols/rfc2616/rfc2616.html

See Sections 4 aand 5 of the reference for info on Headers.

[Gary Beene]

Stuart, thanks for the response. The answer was what I hoped it would be!

[John Thompson]

I believe it's important you also include a "Host: site.com" header after your first one. I think it is important for virtual hosts and anywhere a server is hosting more than one domain from one IP.

I've always included the request, host, and User-Agent (claiming to be IE) and have never had any issues.

You may find, when dealing with dynamically created web pages, that your inclusion or exclusion of a user agent (and what it claims to be) could be the reason you are or aren't getting the data you wanted.

[Jeff Blakeney]

I've been adding the Host header as well but only when using HTTP/1.1. I think it is required for HTTP/1.1 but HTTP/1.0 doesn't require any headers if I remember correctly.

[John Thompson]

I think it might just depend on if it's a shared server with virtual hosts.

If I've got Apache serving ten different domains...

   Tcp Open "HTTP" AT "www.garybeene.com" AS #1 TIMEOUT 60000
   Tcp Print #1, "GET  /files/tcp_test.txt  HTTP/1.0"
   Tcp Print #1, ""
   Tcp Recv #1, 4096, Buffer$
   Tcp Close #1

...doesn't tell me which domain you're requesting from.

I don't know how it would get around it (unless anything HTTP/1.0 could only serve one domain and no virtual hosts?), but if anything unusual happens maybe throw one in there.

-John