[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: lynx-dev Re: A vulnerability in Lynx (all versions) <bug 004352> [BS
From: |
T.E.Dickey |
Subject: |
Re: lynx-dev Re: A vulnerability in Lynx (all versions) <bug 004352> [BSDI-Support-Request #41289] |
Date: |
Wed, 1 Jul 1998 09:33:36 -0400 (EDT) |
>
> 980701 Thomas Dickey wrote: 980630 Philip Webb wrote:
> >> AFAIK this was corrected in 2-7-2 & 2-8
> > not exactly - 2.7.2 and 2.8 have a not-very-good fix.
>
> like, a how-far-not-very-good fix?
> do you mean a fix which reliably avoids the race condition,
> but at the expense of clumsy code & poor maintainability,
> or a fix which works most of the time as far as anyone knows for now ... ?
No. It doesn't avoid the race condition at all. It simply follows the
code that opens the file with a call to make it unreadable.
> > I have a generic fix in the development version, which can be improved
> > (unless you're logged in as root, the generic fix works just fine,
> > but there's the special cases as usual).
>
> why-on-earth would root have to worry? s/he can do anything anyway ...
The code that I used would behave differently if root is running (so that
opens up a hole). Someone told me about it; will fix when I get there.
I have also to check/see whether Lynx should cache temporary filenames
for reuse (doesn't seem needed, but would take work to disentangle).
> > The issue of a 'race condition' refers to the fact
> > that one could easily devise a program that predicts
> > the next temporary-filename that 2.7.2 would use (2.8 has the same code)
> > and create a spoof filename that's linked to another location.
>
> so it is the same much-debated problem,
> to which -- among other people -- FM offered 2-7-2 as a solution,
> which was incorporated in 2-8.
>
> should users be concerned as of 980701 ?
not ordinary users - the dev.17 (or whatever) version works for them.
--
Thomas E. Dickey
address@hidden
http://www.clark.net/pub/dickey