> > Top priority tasks list: > > - Unbreak the userland net_server version, so we can debug far more > > easily... > I will look into it this weekend. Did Axel ever finish the IPC stuff > required for the userland version to work? Of course he did :-)) The only thing that is not yet working (and that I am aware of) is select() support in the userland version, but that won't be hard to port over, after Philippe has finished it for the kernel. > > - Fix the R5 select() support (I'm working on it, with a new silly > > idea: I pass the select_wakeup_sem id to > > driver/socket event notification function to wakeup the select() > > call > > and the released "count" to report > > back the fd/event (read, write, exception) pair raised. Yep, it's > > silly and hacky: I'm desperate! ;-) ) > If you can get that to work, that will rock! if not, I propose we > still > implement select() suport but it would only work with the BONE > kernel. True. Anyway, I don't understand how you'll want to use the release count to report anything back, bug using a semaphore for select() seems to be the right way (the other would be a port). > > Next milestone: release a first alpha-testers targeted BeOS R5 > > (vanilla and BONE) net stack replacement, > > to boost bug reports rate, and hopefully the fixed bugs in the way. That would be really cool, and would boost the usefulness of plain R5 : -) Adios... Axel.