On Wed, Feb 20, 2008 at 09:10:28AM +0100, Quentin Garnier wrote: > > It turns out that Paul Janzen rewrote monop's save/load handling for > > OpenBSD some years ago. I am planning to merge this. > > > > Does anyone care if I break compat with old save files? > > Well, it's always a bad thing, but then it's monop(6). If it puts > someone in a life-or-death situation, or simply makes the machine > un-bootable, it's probably the user's very own fault. Yeah. Plus, that game is really a pile - it has no UI and is full of rules bugs. I am not entirely convinced it's worth fixing, but then, we like engaging windmills here. > > (For those who are blissfully unaware, monop saves by dumping its > > entire heap to disk as a blob and reading it back and hoping that > > works. I am not entirely convinced it is working properly as it is, or > > that if it was I haven't already broken it in the last few days.) > > Wouldn't that break when the malloc(3) implementation changes anyway? > That would mean it's already broken in -current. No; it has its own private malloc because it broke some years ago the last time the malloc implementation changed. However, since I wrote the above I've confirmed that it isn't working properly anyway. It only saves the heap, not the contents of the data segment, and thus the save files are missing various essential bits of information. So I think I'm just going to go ahead and change it. -- David A. Holland dholland%netbsd.org@localhost