On Thu, Jul 25, 2013 at 02:03:33PM +0200, Gerd Hoffmann wrote: > On 07/25/13 13:22, Michael S. Tsirkin wrote: > > On Thu, Jul 25, 2013 at 01:05:12PM +0200, Gerd Hoffmann wrote: > >> Hi, > >> > >>> I can change the implementation but I don't think it's > >>> a good idea to copy property names around: > >>> it's too fragile, compiler won't warn us if we > >>> change the name or value semantics, > >> > >> I'm not worried. Changing the strings will break the command line > >> interface too (qemu -device pvpanic,ioport=...), so that isn't going to > >> happen. > >> > >> cheers, > >> Gerd > > > > What will catch this breakage? > > There are 0 users actually tweaking the port > > number so I'm sure no one will notice this. > > > > In any case, catching errors at compile time > > is much better than at runtime. > > > > What exactly are advantages of duplicating > > property names in this way? I don't see any. > > You don't need access to pvpanic internals then and thus the code can be > moved over to the acpi generator. At least in this case where all info > needed is already available via properties. > > cheers, > Gerd
We'll have to disagree here. There's no access to internals with an API. I prefer using APIs, since they are compiler-checked. -- MST