Re: Non-blocking I/O

From: pinkfloydhomer_at_[hidden]
Date: 2001-03-06 10:37:00

--- In boost_at_y..., Lois Goldthwaite <loisg_at_o...> wrote:
> Not argument about the utility -- on the rare occasions when it's
> absolutely needed -- but is this really something a language can
> standardize? Isn't it dependent on a BIOS (or lower-level) facility 
that
> the operating system might not expose? 

I don't think so, but I see your point.
The same problem could occur with ifstream, cout etc.
It _could_ all be features that a certain O/S did not support, but 
most do. Just as most O/S'es facilitates non-blocking I/O on certain 
logical or physical devices. Just as there are some constants that 
are defined by the system (the size of an int, the value of 
CLOCKS_PER_SEC etc.) there could be some classes or functions that 
were dependant on what the O/S was able to implement. At the very 
least, the stream hierachy could have a class that supported non-
blocking fill of the buffer, that would return "as soon as possible", 
that is, on most systems, it would return immediately, if nothing was 
ready to be added to the buffer. cin could belong to this class, and 
sure, programs using this feature in the way that I want my program 
to, wouldn't work on a microwave oven, but it would work on the 
machines I wanted it to work on.