Hi, > I think we will need: > - one thread for plug/unplug of readers > - one thread for getting client connections > this "mother" thread forks for each new client so each client will > have a different thread when talking to a different reader. > > So each client talking to a different reader will be managed by a > different thread. And it is the jobs of the kernel scheduler to avoid > starvation among the threads. I think also something like that. Perhaps we can keep the actual code and just modify the code in winscard_svc.c I have not yet see the code for SCF and I don't know the impact on it.
> I don't know if we need to stabilise the 1.2.0 version as it is now and > include this new feature in a 1.3.0 version or if we include this work > now in 1.2.0 and will not have a stable official release before a long > time. > > I am happy with the version we have in CVS (only one, easy to include, > feature request on alioth tracker [1]). But I cannot test pcsc-lite on > all the platforms pcsc-lite is supposed to support. What do you (the > muscle community and pccs-lite users) think? I agree to stabilise the 1.2.0 version as it is now. I have just a little bug to correct before you stabilise it. I try to do this in the week. I may try PC/SC Lite on a Solaris but only with serial reader and perhaps on a NetBSD with USB and serial readers. I hope that even if we release the version 1.2.0, we will have quickly a version 1.3.0. If we release a PC/SC Lite version 1.2.0, I think that it will be good to update it on freshmeat [1] in order to give a visibility on our community. [1] http://freshmeat.net/projects/pcscresourcemanager/ Regards, Damien Sauveron ------------------------------------------------- This mail sent through IMP: http://horde.org/imp/ _______________________________________________ Muscle mailing list [EMAIL PROTECTED] http://lists.musclecard.com/mailman/listinfo/muscle
