Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!uflorida!haven!aplcen!osiris!consult From: consult@osiris.UUCP (Unix Consultation Mailbox ) Newsgroups: comp.sources.bugs Subject: Re: Lacking (Was: Re: PAX won't make on my UNIX PC) Message-ID: <2839@osiris.UUCP> Date: 14 Feb 89 18:52:34 GMT References: <550@kosman.UUCP> <523@jhereg.Jhereg.MN.ORG> <17934@vax5.CIT.CORNELL.EDU> <542@jhereg.Jhereg.MN.ORG> Reply-To: consult@osiris.UUCP (Unix Consultation Mailbox (Phil)) Organization: The Johns Hopkins Hospital, Information Systems Lines: 26 In article <542@jhereg.Jhereg.MN.ORG> mark@jhereg.MN.ORG (Mark H. Colburn) writes: >In article <17934@vax5.CIT.CORNELL.EDU> jhc@vax5.cit.cornell.edu (James H. Cloos, Jr.) writes: >>It appears that the NeXTs are also lacking . And that with gcc as >>the standard compiler. > >I beleive it is not included because some of the values are somewhat >hardware dependant. Well, yeah, is (more than somewhat) hardware dependent... which is certainly no reason *not* to include it in the distribution, but in fact the exact opposite - it's not the kind of file you *can* port from another random system without problems, so it *should* be distributed. Maybe what Mark Colburn said was based on misinterpreting James Cloos' statement to mean that *gcc* was not distributed with limits.h (reasonable), when in fact James meant that the NeXT was not shipped with as part of its standard software (NOT reasonable). >I have a limits.h which is the minimum required by >ANSI and POSIX, you may use it if you like... I looked at this and while it looks pretty close, I'll bet it doesn't work on a NeXT (the 16-bit int and default unsigned char make me suspicious) - what kind of system was this limits.h tweaked for? phil