[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: sshd no longer accepts connections

On Wed, Feb 14, 2001 at 01:01:59PM +0100, Paul Slootman wrote:
> On Tue 13 Feb 2001, Jon Leonard wrote:
> > There are 3 packages that are currently giving me trouble:
> > w3m, man-db, and vim-gtk.
[dselect stuff snipped]
> > w3m depends on libc6.1 (>= 2.1.97)
> Ah, this is potato!

I should have made that more clear, yeah.

[vi behavior]
> Yes, I can reproduce this on a potato system.
> I'm not sure how to cope with this; I guess I could upload to
> "stable", and hope it gets into proposed-updates; however, I
> recall one of the ftp-masters writing that proposed-updates
> should go away...
> The point is that potato is "stable", meaning nothing really gets
> changed there.
> Aargh, man-db got installed as a security-fix a couple of days
> ago. Whoever Somehow the alpha package must have been botched
> (the i386 package is OK).  I'll have a look at fixing this for
> potato.

Maybe there needs to be a policy discussion on what to do with obviously
broken security updates?  It seems to me that a system where the admin
can't use man may be some sort of security risk all by itself.

> > It only generates the zsoelim message once for each man page looked at, though
> > installing a man-db package resets that.  My best guess would be that man-db
> It only generates a manpage if there isn't one in the cache.
> Apparently even with the error a file (possibly empty) is created.

I figured as much.  I mentioned it because it affected the reproducability
of the problem.

> > I suspect that the vim-gtk was built on a system that didn't have gtk
> > installed at the time, so the configure step built a non-graphical vim.
> Build-Depends should help against that in the future...

Cool.  I take it there's basically no chance of getting a fixed (rebuilt)
version into stable or proposed-updates, though?

> > user unaligned acc      : 39158409 (pc=12004e170,va=1203801d2)
> Ouch, that's a lot...

Well, the system has been up for a while, but yeah.

> Anyway, it might be recommendable to upgrade to "testing",
> which is quite usable IMHO. At least there things can be
> fixed easily.

That sort of defeats the purpose of stable, doesn't it?  I'm running Debian
stable because I want a system that won't break on me, but still gets timely
security fixes.  (And for the most part, Debian stable does this admirably)
I recommend Debian to my less computer-adept friends for similar reasons.

Maybe we should call testing stable instead, if it's less broken.

Anyway, thanks for your help.

Jon Leonard

Reply to: