[Glenn9999]

(I'm not sure where the best place would be to ask this, and this seems best) Are there any generally accepted rules for HTTP requests that are consistent with most sites? I ask because I've been trying to work (code-wise) on downloading a file pointed to by a valid fully qualified URL. Where I have it now is trying to handle "strange" URLs where valid files aren't necessarily specified. But I get stopped connection-wise if I try a main site link (like http://www.msfn.org/). So this got me wondering, and I'm not finding any good pages on it.

What's the proper HTTP headers to be sending besides the obvious GET to download the file, in terms of what most sites will accept?

[CoffeeFiend]

Usually it's pretty straightforward. For msfn it surely is. You can even try using plain old telnet:

telnet 67.19.16.68 80
GET /board/ HTTP/1.1
Host: www.msfn.org
(followed by enter twice)

Works fine (you can see the markup fly by). Some sites require a few extras (e.g. "User-Agent: cloned-user-agent-string-here") but most of the time this is all you need. Obviously, some sites require authentication to download files.

[Glenn9999]

CoffeeFiend, on 21 March 2010 - 03:04 AM, said:

Usually it's pretty straightforward. For msfn it surely is.

Thanks. It does seem so. The problem usually is to figure out what to expect. My problem that prompted this turned out to be that I was expecting a content length and that isn't valid for HTML. But still a good question, since I want what I'm working on to not be too "strange".

[Glenn9999]

Glenn9999, on 21 March 2010 - 03:37 AM, said:

Thanks. It does seem so. The problem usually is to figure out what to expect.

Moreso again. I keep hitting this downloader code I have (I got it to resume now), and found a site that sends HTML *as* the HTTP headers. So I'm still wondering if there's some kind of universal standard (there has to be I would think or the browsers won't work right).

Any ideas on this one?