Hi Caitlin, 2009/9/24 Caitlin Shaw <rogueeve@xxxxxxxxxxxxx>: > >>>> a fundamental scheme like that is bound to break things. >>>> >>> >>> We could also identify such applications when running them >>> (runtime_loader stuff) and change the app_server behavior for them (like >>> activating a "no aa" flag). >>> > > How good of an ability does the runtime_loader actually have to identify an > "old BeOS app that was written before diagonal lines were anti-aliased?". > IOW, imagine this scenario; say I'm the author of an app which I've been > using for a while on Haiku and it works perfectly. One day I take the source > to my works-great app and recompile it verbatim on Haiku. Is it suddenly > going to be broken? It's probably just a 1-line change, but I was not > expecting that, may not know about it, it may take me some time to even > realize why it's not working anymore, etc. All this time I'm annoyed because > I was probably trying to do something totally different, and didn't actually > want to tear into the source of my 10-year old program which was previously > working fine. I can't imagine there's a heuristic smart enough to detect the > intent behind the source rather than just build platform, compiler used, > file modify date, etc. You're right, there is no heuristic smart enough. It's up to you to realize that your BeOS app now became a Haiku app. This incompatibility should eventually be listed in the Haiku book though, so you would eventually find out. N>