On Thu, Jan 08, 2004 at 04:24:22PM +0100, Joerg Schilling wrote: > It _is_ wrong to assume that a random program compiled for OS revision A > will run correctly on OS revision B
Definetly NOT. e.g. "grep". grep only uses libc-interface. As long as the program <-> libc interface is stable it will have no problem with the libc <-> <whatever> site. It is excatly THE job of libc to abstract away the "right" side. (Or the left when you assume hardware/kernel is leftmost) Only "system dependend"(hardware, kernel interfaces, ..) software (e.g. cdrecord, star, ps, lspci, iptables) have this type of problem. And btw. There is also a "binary" and "compile" compatiblity factor in this. e.g. libc6 aka glibc2 broke compatiblity with libc5 so A FEW programms needed patches to be COMPILABLE on new systems whereas (when the needed shared libraries where installed or the programm was static compiled) the libc5 binaries where still runable. Or in other words for >95% of all programms your statement is false! There is, always was and will ever be a small fraction of programs with this type of problem. The majority of software doesn't have this type of problem(s)! Or in other other words: It is wrong to assume that a random system dependend program compiled for OS revision A will run correctly on OS revision B for system. I have more other words: If Linux (for 2.8/3.0 whatever) would get incompatibel to every other Unix(type) OS AND to POSIX, BSD (and would need to drop glibc2). Then you can say what you have said. Bis denn -- Real Programmers consider "what you see is what you get" to be just as bad a concept in Text Editors as it is in women. No, the Real Programmer wants a "you asked for it, you got it" text editor -- complicated, cryptic, powerful, unforgiving, dangerous.
