Build error on HP-UX

[Rick Jones]

Ben Lindstrom wrote:
> 
> On Thu, 19 Jun 2003, Rick Jones wrote:
> 
> > Ben Lindstrom wrote:
> > >
> > > How does HP/UX define utimes?
> >
> > Thusly on 11.0:
> >
> >
> >  utimes(2)
> > utimes(2)
> >
> >  NAME
> >       utimes - set file access and modification times
> >
> >  SYNOPSIS
> >       #include <sys/time.h>
> >
> >       int utimes(const char *path, const struct timeval times[2]);
> >
> 
> Interesting enough Linux and OpenBSD at least deinfe it *times instead of
> times[2], but the code itself should be fine.

I wonder what Posix/X/Open say about it, if anything.

> It bothers me that configure.ac generated the wrong check.  What library
> is utimes() part of?  

I'm guessing it is part of libc:

$ nm /usr/lib/libc.sl | grep utimes
_utimes             |    739636|extern|code   |$CODE$
utimes              |    739596|extern|entry  |
utimes              |    739636|extern|code   |$CODE$

> Maybe we need additional check besides:
> 
> AC_CHECK_FUNC(utimes,
>         [], [ AC_CHECK_LIB(c89, utimes, [AC_DEFINE(HAVE_UTIMES)
>                                         LIBS="$LIBS -lc89"]) ]
> )

Likely - if I try to compile my foo.c with a -lc89:

$ cc foo.c -lc89
/usr/ccs/bin/ld: Can't find library: "c89"

which could I suppose make configure think there is no utimes on UX...
perhaps it should try first without the -lc89 and only try it if the
first one fails?

> [..]
> > > bsd-misc.c:int utimes(char *filename, struct timeval *tvp)
> > > bsd-misc.h:int utimes(char *, struct timeval *);
> > >
> > > Try in the bsd-misc.h changing it to be exactly as it's defined in the
> > > bsd-misc.c.  Maybe your compiler is picky.
> >
> > Perhaps the use of const in the UX headers is the trigger for this?
> >
> No it is because the second argument is different  (*tvp vs tvp[2])

Thanks - I confirmed it by changing the timeval to an int and dropping
the sys/time/h include and seeing that it compiled just fine.

> I'd say something is wrong with configure.ac script.  If you go into
> config.h and uncomment "#define HAVE_UTIMES 1" and compile does HP/UX 11
> work correctly?

I'll have to let the original reporter report on that. 

rick
-- 
Wisdom Teeth are impacted, people are affected by the effects of events.
these opinions are mine, all mine; HP might not want them anyway... :)
feel free to post, OR email to raj in cup.hp.com  but NOT BOTH...