[Rd] [R] HTTP User-Agent header

[Seth Falcon]

Prof Brian Ripley <ripley at stats.ox.ac.uk> writes:
> I also thought that there was no need for this level of complexity. 
> (BTW, some of the patch is changes Seth has made for other purposes, e.g. 
> that to memory.c, so please no one apply all of it.)

*blush* sorry about that.  I made the final diff from a src tree on
another machine and it was dirty.  memory.c should NOT have been
touched by my patch.

> I'd be happy for R to just identify itself as 'R', which seems allowed:
> (http://www.w3.org/Protocols/rfc2616/rfc2616-sec14.html).  But I am a bit 
> concerned that sites may not just require the field but also require a 
> particular format (even though W3C does not).

As long as it is going to identify itself, I think there is value in
having it provide version, platform, OS info.

Given the concern that some sites that currently work may stop working
(example?), perhaps making this a global option is a good compromise.
The option could be httpRequestHeader and the default value would be
as proposed in my patch.  A NULL value would result in the current
behavior, no extra header info in the request.