- wineserver: that's the tricky one. I'm not sure lazy initialization would do anything here, but in the future we may provide a dummy version of sorts where the Winelib user knows there is only one process to worry about. Maybe even start it automatically only if there are more than one process to worry about :)
Hey, I don't know how feasible this ideas are, but it's probably worth it to try to make Winelib more easily useable as a library rather than encourage people to hack some bits out of it. I think many of the changes are beneficial to Wine in any event. That's why I think tasking out a long time goal may be a good thing as some of the tasks may bubble out anyway (like the configuration work around winecfg).
I know some Linux distros (RHAT) set wineserver to allways run a a background process via xinetd or other rc.d scripts so doesnt this method take care of part of the problem of wineserver initialization? I mean if the system is going to need wine for software the user is going to run most or all the time then shouldnt wineserver always run?
Thanks Steveen