hey, I'm sure the webkit folks would love it if you ported their code to 2.95...but only if it also compiles for 4.3 or whatever. Or, you could just maintain a downstream 2.95/haiku patch. Personally, I'd recommend the latter, because it means less hassle overall, and perhaps easier porting once R2 comes around. On 8/21/07, Ryan Leavengood <leavengood@xxxxxxxxx> wrote: > On 8/21/07, Philippe Houdoin <philippe.houdoin@xxxxxxx> wrote: > > > > Did you try to simplify this silly macro to this: > > > > #define PLATFORM(WTF_FEATURE) WTF_PLATFORM_##WTF_FEATURE > > > > According to #if preprocessing directive specification, any undefined > > identifier > > will be evaluated as 0. > > > > Which means "#if PLATFORM(DARWIN)" will give "#if WTF_PLATFORM_DARWIN" > > which, > > logically, should be evaluated to zero if such identifier is undefined. > > > > I failed to see why someone defining a macro as an expression will ever > > needs to > > check if an identifier is defined before evaluating its value, as 0 is > > specified > > as the default value for any undefined identifier... > > > > Anyway, maybe it worth a try. > > You are right, based on what I've been reading. I agree it is kind of > silly and will investigate making this change tonight. Maybe all this > discussion about GCC 2.95.3 versus GCC 4 will be moot if the changes > needed actually FIX the WebKit code like the above :) > > Ryan > >