friedl@vsi.UUCP (Stephen J. Friedl) (07/01/88)
Hi net.folks,
For some unknown reason, I have always felt uncomfortable
doing a return(n) from main: I used exit(n) instead. Can anybody
think of any possible reason why a return is not a good idea? I
have the feeling that my discomfort is entirely unfounded...
Steve
--
Steve Friedl V-Systems, Inc. (714) 545-6442 3B2-kind-of-guy
friedl@vsi.com {backbones}!vsi.com!friedl attmail!vsi!friedl
Nancy Reagan on John DeLorean: "Just say snow"guy@gorodish.Sun.COM (Guy Harris) (07/01/88)
> For some unknown reason, I have always felt uncomfortable > doing a return(n) from main: I used exit(n) instead. Can anybody > think of any possible reason why a return is not a good idea? Some implementations don't properly implement "return(n)" from "main". SunOS's C implementations did not do so from SunOS 2.0 to SunOS 3.5; they implement it correctly in 4.0. I seem to remember somebody claiming some other implementation did not do so. I think every S5 implementation from AT&T handled "return(n)" from "main" correctly, as did 4.[23]BSD; this does not, of course, guarantee that any particular implementation derived from either or both of those will do so correctly. ANSI C mandates it, so everybody's going to have to get with the program eventually.
davidsen@steinmetz.ge.com (William E. Davidsen Jr) (07/01/88)
In article <741@vsi.UUCP> friedl@vsi.UUCP (Stephen J. Friedl) writes: | Hi net.folks, | | For some unknown reason, I have always felt uncomfortable | doing a return(n) from main: I used exit(n) instead. Can anybody Some versions of C don't use the value of the return from main as a call to exit. Instead, the assume that if you don't call exit with a value you terminated normally and should have a zero exit status. Your gut feeling is a good one, it will save you a lot of pain in the long run, I suspect. -- bill davidsen (wedu@ge-crd.arpa) {uunet | philabs | seismo}!steinmetz!crdos1!davidsen "Stupidity, like virtue, is its own reward" -me
gwyn@brl-smoke.ARPA (Doug Gwyn ) (07/02/88)
In article <741@vsi.UUCP> friedl@vsi.UUCP (Stephen J. Friedl) writes: >doing a return(n) from main: I used exit(n) instead. "return n" is technically correct, and is slightly more satisfactory from a "top-down control" viewpoint. I have heard that there are some purported C implementations that ignore the return value from main(), though, always returning a success indicator unless exit() is used. Those aren't Standard-conforming implementations, but you might be stuck with one. P.S. You don't need parentheses around the return expression.
bill@proxftl.UUCP (T. William Wells) (07/04/88)
In article <741@vsi.UUCP>, friedl@vsi.UUCP (Stephen J. Friedl) writes: > For some unknown reason, I have always felt uncomfortable > doing a return(n) from main: I used exit(n) instead. Can anybody > think of any possible reason why a return is not a good idea? I > have the feeling that my discomfort is entirely unfounded... The only reason I can think of is broken compilers. The two should be equivalent. Of course, one of those broken compilers is the one we have ... on a Sun 3 running SunOS 3.4. It acts as if return (n) were a return (0). [I understand that that has been fixed in a later release.]
ckl@uwbln.UUCP (Christoph Kuenkel) (07/15/88)
In article <741@vsi.UUCP>, friedl@vsi.UUCP (Stephen J. Friedl) writes: > For some unknown reason, I have always felt uncomfortable > doing a return(n) from main: I used exit(n) instead. Can anybody > think of any possible reason why a return is not a good idea? There are simply some implementations of _startup() that don't take care of the returnvalue of main() (we have that problem here). You should code: main() { ... exit(retcode); /*NOTREACHED*/ } This should work in any case. christoph -- Christoph Kuenkel/UniWare ck@tub.BITNET Kantstr. 152 uunet!unido!tub!ck 1000 Berlin 12 {unido,tmpmbx,uwnue,nixbln,tub}!uwbln!ckl West Germany