Re: Reducing the number of system calls for includes

> That said, we could possibly have a fast mode and use fstat() to get some > device information on the open file. I am worried about functionality > though. I'm not sure it's worth breaking. Yes, fstat() is fast. And perhaps it isn't worth breaking functionality across the board, but in my particular case I don't really have a choice. It is simply too slow to handle the loads that it needs to handle and either I need to get twice as many machines or I need to lose some functionality. Another idea would be to try to streamline the normal case for an absolute pathname. How about doing a readlink() on it. If that tells us that it is a link, then we know there is something funky and we can do a realpath, otherwise we just assume it is canonical assuming there are no . .. or // in the filename in which case we fall back to a realpath() as well. Or, the really straightforward dumbass solution, have some hint we can pass along that tells the parser to treat what follows as a canonical filename. include "\001/home/rasmus/foo/u2.inc"; A bit too hacky I guess. Perhaps something like: include_canonical "/home/rasmus/foo/u2.inc"; And we would simply assume that this is a canonical filename and skip all the checks. This last one is probably what I will need to do, but I will do some tests on the readlink() optimization and see if that buys me anything. -Rasmus