datalinkcom

<0> The "thing" part is meaningful to the compiler and to the guy reading the source code. <0> i.e. "thing *foo;" can contain an address which represents the start of a "thing", or it could represent the start of a "something_else". <0> The compiler takes it on faith that you will be using "foo" to contain the address of the start of a "thing". Casting works around this :). <0> It's all in the interpretation. <0> so a "char **argv" just says that "argv contains an address which is the start of an(other) address; that 'other' address will contain a char." <0> (theoretically). <0> I say "theoretically" because it's not illegal to have "char *foo" and "int *bar" contain the same address; as I said above, it's all in the interpretation: References to "*foo" will be assumed to refer to a char, while references to "*bar" will be assumed to refer to an int, even though "foo" and "bar" contain the exact same address. <1> anyone done any rpc programming? <2> Greywolf: i've seen unions that do just that <0> reuben: That's kind of the point of a union. <2> dsr: unions are pointless without pointers? <2> s/dsr/Greywolf/ <0> Aha. Ahaha. GIVE me thaaat... <2> a union's point is pointers, that point to a liter of beer, and to those who prefer their unions, here's a pointerless pointless Hear Hear! <0> I've used unions for multiple data types. I always thought that a stat structure should have had a union for rdev/size, since on a device, size is meaningless, and on everything else, rdev is meaningless. <3> Greywolf- I got that part. I was didn't know when to use ** instead of * <3> anyway, time for dinner <3> I'll be back into it tomorrow morning, I think <2> greywolf: a device generally makes a union meaningless