Using [uw]tmp
Using [uw]tmp
- Subject: Using [uw]tmp
- From: Graham J Lee <email@hidden>
- Date: Tue, 13 Dec 2005 15:38:36 +0000
Hi,
I want to get login information for users (basically my goal is a
Bonjour-enabled equivalent to the old GNU or ICSI distributed finger
server[*], where any systems on the local network register themselves
with the server so I can find out if a user has logged in anywhere on
the local network) and as neither of the distributed fingers come
anywhere near close to compiling on Darwin (8.3.0) have tried to
start from scratch. Having observed that /var/run/utmpx is empty,
and that there's no getutent on Darwin, I'm doing it the same way
that programs like w(1) do and reading struct utmps from the
appropriate file. However, what I get back doesn't seem consistent
with the idea that utmp(5) [or indeed wtmp(5)] is just a bunch of
struct utmps stored back to back. Here's my code:
#import <Foundation/Foundation.h>
#import <utmp.h>
#import <stdlib.h>
#import <stdio.h>
extern int errno;
int main (int argc, const char * argv[]) {
NSAutoreleasePool *pool=[[NSAutoreleasePool alloc] init];
FILE *utmp_file;
struct utmp *utmp_store;
errno=0;
if((utmp_file=fopen(_PATH_UTMP,"rb"))==NULL)
{
perror("csalutd: couldn't open utmp");
exit(EXIT_FAILURE);
}
if((utmp_store=(struct utmp *)malloc(sizeof(utmp_store)))==NULL)
{
perror("csalutd: couldn't malloc struct");
exit(EXIT_FAILURE);
}
while(fread(utmp_store,sizeof(utmp_store),1,utmp_file)==1)
{
printf("That fool %s, over %s way, was responsible.\nIt was at %d
on %s.\n",utmp_store->ut_name,utmp_store->ut_host,utmp_store-
>ut_time,utmp_store->ut_line);
}
fclose(utmp_file);
[pool release];
return 0;
}
The first few lines of output:
[Session started at 2005-12-13 15:32:15 +0000.]
That fool , over way, was responsible.
It was at 0 on .
That fool , over way, was responsible.
It was at 0 on .
That fool , over way, was responsible.
It was at 0 on .
That fool , over way, was responsible.
It was at 0 on .
That fool , over way, was responsible.
It was at 0 on .
That fool , over way, was responsible.
It was at 0 on .
That fool , over way, was responsible.
It was at 0 on .
That fool , over way, was responsible.
It was at 0 on .
That fool , over way, was responsible.
It was at 0 on .
That fool , over way, was responsible.
It was at 0 on cons.
That fool , over way, was responsible.
It was at 0 on ole.
That fool , over way, was responsible.
It was at 0 on leeg.
That fool , over way, was responsible.
It was at 0 on .
That fool , over way, was responsible.
It was at 0 on .
That fool , over way, was responsible.
It was at 0 on \242\213\305X.
So there's a lot of 0 going on, and what look like valid data (e.g.
leeg is currently logged in on console). I get similar output if I
use _PATH_WTMP, by the way. But the data don't seem aligned in the
same way that <utmp.h> suggests they would be, especially as there
seem to be so many \0 characters all over the place.
How should I read struct utmp records from the utmp file?
[*]It's nice of you to be concerned about the security implications
of a distributed finger service, I've taken those into consideration
and still need such a service kthx.
Thanks,
Graham.
--
Graham Lee GPG Key ID: 01D5B9D8
UNIX Systems Manager,
Oxford Physics Practical Course
http://users.ox.ac.uk/~wadh1342 01865 273450
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Darwin-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden