[Glibc-bsd-hackers] STATUS page

[Nathan Hawkins]

Robert Millan wrote:

>On Wed, Jun 30, 2004 at 07:38:05AM -0500, Nathan Hawkins wrote:
>  
>
>>- FIXME: ifupdown does not handle the gateway entry correctly, so do not
>>  add it in your interfaces file, just add an up hook, that will do as a
>>  gateway entry:
>>
>>  auto rl0
>>  iface rl0 inet static
>>	address 192.168.1.100
>>	netmask 255.255.255.0
>>	broadcast 192.168.1.255
>>	up route add -net 0.0.0.0 192.168.1.1
>>
>>I believe this is relatively easy to fix. (Or would be if ifupdown didn't 
>>have obfuscated
>>sources.) Just need to fix the gateway command in the template.
>>    
>>
>
>Actualy, we're going to replace /sbin/ifconfig with the one in inetutils, so
>it's better to not bother with ifupdown for now.
>  
>
Hmm. That's not what I would do - it seems like a lot of work to make 
inetutils support everything that BSD ifconfig does.

>>- FIXME: DNS resolving is broken, make sure /etc/hosts has everything you
>>  need. MX resolving won't work so MTA doesn't as well [working on it].
>>
>>This is the major problem keeping me from being able to use this system. I 
>>can work around
>>nearly everything else.
>>    
>>
>
>This has an high priority in my TODO. Right after gcc which I almost finished
>(I have patches for gcc-3.5 that port all the stuff except gnat)
>  
>
This is largely becoming a moot point for me, since my interest in this 
project is rapidly becoming academic. Real life, and other commitments 
are taking up all of my time. :-(

>>- FIXME: df is broken when not specifying any mount point)
>>  (Workaround: Just copy the /bin/df from the FreeBSD chroot)
>>
>>I believe this is broken headers in /usr/include/sys.
>>The GNU df works when built against native libc.
>>    
>>
>
>This is because FreeBSD doesn't define _PATH_MTAB. We do, however, define it
>because we're going to provide it (libpmount already supports mtab handling).
>  
>
Yeah, they do something different, in /var, I think, which is actually a 
better place for it.

>>- FIXME: terminal broken, does not handle Ctrl-C nor Ctrl-Z.
>>  [ It may be a bad tty initialized by init ]
>>
>>This is because the terminal isn't being designated a controlling terminal. 
>>(An arcane area
>>of Unix I knew nothing about and still know virtually nothing about...) I'm 
>>not sure
>>precisely where this is supposed to happen, but I remember that I patched 
>>getty with a
>>single ioctl, and the problem went away.
>>    
>>
>
>Which ioctl? Can you find out? We were discussing this in glibc-bsd-devel a
>while ago, you can check the archives.
>  
>
TIOSCTTY, I believe. It is the same on Linux and BSD, it's just that for 
some reason it doesn't seem to be getting called when sysvinit is 
compiled on FreeBSD.  The precise timing of when this should happen, and 
which process should be calling it is unclear to me.

       ---Nathan