Re: [xml] NanoHTTP and HTTP User-Agent field

[Daniel Veillard <veillard redhat com>]

On Wed, Feb 28, 2007 at 11:19:48AM -0600, R. Steven Rainwater wrote:
> I'm using NanoHTTP as the basis of Advogato.org's blog aggregator with
> quite a bit of success but recently ran into an interesting problem.
> I've come across a website that blocks HTTP GET requests from agents
> that don't provide a User-Agent request header field. 
> 
> I checked RFC-2616 and the User-Agent field is listed as a field that
> SHOULD be implemented but is not strictly required. On the other hand,
> as the maintainer of several website that are frequently abused by bad
> user agents, I can see why someone might want to block agents that don't
> identify themselves. So, I think NanoHTTP should provide some kind of
> identification.
> 
> I don't see any obvious way of adding a User-Agent field in the
> documentation for NanoHTTP. I'm currently calling xmlNanoHTTPInit(),
> followed by a number of calls to xmlNanoHTTPFetch(), and finishing with
> a call to xmlNanoHTTPCleanup(). It seems like the init call would be the
> logical place to add an optional User-Agent value.
> 
> Does anyone have any suggestions (or examples) of how a User-Agent could
> be added without modifying the libxml2 code? 
> 
> If there's no easy way to implement a User-Agent with the current
> codebase, there seem to be two possible solutions:
> 
> 1. Patch libxml2 to add a default user agent, perhaps something like:
> 
>  NanoHTTP libxml2 version 2.6.27

  Libxml2 used to do that, people objected and I removed it, I would avoid
this.

> 2. Patch libxml2 to allow a user-configurable agent string. This might
> be done by adding a new argument to xmlNanoHTTPInit() or a new function
> like xmlNanoHTTPSetUserAgent() (or maybe something more generic like
> xmlNanoHTTPSetRequestHeader()?).

  xmlNanoHTTPSetRequestHeader() sounds better to me especially if it
is local to a xmlNanoHTTPCtxtPtr context, and that a new call override 
previous setting if any. Then NULL argument would reset to no header at all.
Just a quick idea but it sounds to me a simple patch adding this function
should be fine,

Daniel

-- 
Red Hat Virtualization group http://redhat.com/virtualization/
Daniel Veillard      | virtualization library  http://libvirt.org/
veillard redhat com  | libxml GNOME XML XSLT toolkit  http://xmlsoft.org/
http://veillard.com/ | Rpmfind RPM search engine  http://rpmfind.net/