Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Sun, 04 Mar 2012 20:10:36 +0100
From: finkler <>
Subject: Re: utmpx support

On 04.03.2012 19:18, Rich Felker wrote:
> On Sun, Mar 04, 2012 at 06:41:25PM +0100, finkler wrote:
> It's intentional, but if you have a real need for utmp support, I'd be
> willing to hear about it.

Well, I am currently trying to implement new (POSIX) user-space 
utilities for Linux, including my own init/login system.
I thought quite a bit about the purpose of utmp(x) myself I admit, 
however, I came to the conclusion that even though I know no one who 
still uses serial terminals to connect to a machine, connecting through 
LAN/Internet is very popular, and it couldn't hurt that an administrator 
has the means to easily see who is currently logged in (who).

I agree that utmpx has grown disproportionally and has lots of misuses 
which just bloat the code.
But as I said, some method of accessing login information seems sane to 
me. And if used with restrain utmpx is quite well suited for this ... I 

> Perhaps a better approach would be making a separate small static
> libutmp.a that could be linked by people wanting real utmp support as
> opposed to the stubs.

Certainly, for what it's worth, I hacked up an utmpx implementation.
I hope I didn't make any mistakes (code and standard wise), but the code 
size is small, so it should be easily verifiable for another set of eyes 

Regards, and thank you very much for the good work with this lib, it is 
really great.


View attachment "utmpx.c" of type "text/plain" (1775 bytes)

Powered by blists - more mailing lists

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.