On Wed, Aug 02, 2006, EV wrote:
> I don't see what has to be changed from the API point of view.  
> As you say, lk_karma_connect() should be the only function to be
> called from applications.  An we already have it.  It's a matter
> of changing whatever necessary so that it makes all the work and
> returns -1 if fail, 0 if connected through Ethernet and 1 if USB.
> 
> What do you think?

I was under the mistaken impression that device and/or port numbers
were necessary, but on closer inspection I see that these are
auto-detected. The only missing piece is the number of SSDP retries
which are performed, but I see that this is hard-coded into riocp
anyway so there's no compelling reason not to hard-code it into the
library instead.

If this returns the device type then that removes the need for the
extra function that we were discussing in relation to lkarmafs.
So it all seems to work out nicely.

However, this *is* effectively a change in API since the call now has
totally different semantics than it did before. Anyone attempting
to build an old riocp or lkarmafs against this new library will get
unexpected results.

I think that perhaps it makes more sense to change the name of the
function call and deprecate/remove the old version.

Or maybe, as Bob says, we are a small enough community that it doesn't
matter. We can just plaster the READMEs with warnings about the
change and be done with it.

What does everyone else think?

Keith.

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
linux-karma-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-karma-devel

Reply via email to