tchrist@convexe.uucp (Tom Christiansen) (01/16/90)
In compiling for patchlevel 8, I found that because we have
(or will have) a <utime.h>, Configure #defined I_UTIME.
However, that file looks like this:
#ifndef _UTIME_H_
#define _UTIME_H_ 1
#ifndef _PROTO
#include <proto.h>
#endif
struct utimbuf {
time_t actime;
time_t modtime;
};
extern int utime _PROTO((char *path,struct utimbuf *times));
#endif
Which meant that doio.c would blow up like this:
gcc -c -fpcc-struct-return -O doio.c
doio.c: In function apply:
doio.c:1999: structure has no member named `acusec'
doio.c:2000: structure has no member named `modusec'
The solution was #undefining I_UTIME in config.h
--tom
Tom Christiansen {uunet,uiucdcs,sun}!convex!tchrist
Convex Computer Corporation tchrist@convex.COM
"EMACS belongs in <sys/errno.h>: Editor too big!"
jand@kuling.UUCP (Jan Dj{rv) (01/18/90)
In article <4610@convex.UUCP> tchrist@convex.Com (Tom Christiansen) writes: >In compiling for patchlevel 8, I found that because we have >(or will have) a <utime.h>, Configure #defined I_UTIME. > [stuff deleted] >The solution was #undefining I_UTIME in config.h > > >--tom This subject has been touched upon before. It seems that many machines don't have the utime.h that Configure takes for granted. We had the same problem you describe on HP-UX, and the solution was the one you suggest. Jan D.