On Fri, 16 Dec 2005, Dan Fandrich wrote:

> I don't see any point in bloating the code with theoretical cases like this.
> It just serves to complicate people's understanding and introduces a support
> burdon. Where do you draw the line about this kind of thing "if it
> happens"? If there actually *were* a platform that needed this, that would
> be a different story.

I'm with Dan on this.

Doing checks for "what if there are systems like this" don't make any sense
since we can easily add the checks once one such a system appears, and if we
do the tests before we know of such a system, there's always a big risk we did
the tests wrong anyway since we haven't been able to verify them properly.

Thus we get more complicated tests with no additional benefit.

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