Hi Noah,

On Fri, Sep 26, 2025 at 11:24:13PM +0200, Salvatore Bonaccorso wrote:
> Hi Noah,
> 
> On Wed, Sep 24, 2025 at 09:29:27AM -0400, Noah Meyerhans wrote:
> > On Tue, Sep 23, 2025 at 10:53:08PM +0200, Salvatore Bonaccorso wrote:
> > > > > > I've published a trixie build based on the just uploaded
> > > > > > 1:2.4.1+dfsg1-7.  You can install it from my people.debian.org
> > > > > > repository.  See https://people.debian.org/~noahm/repo/ for 
> > > > > > details, and
> > > > > > use the following sources file:
> > > > > >
> > > > > > Types: deb deb-src
> > > > > > URIs: https://people.debian.org/~noahm/repo
> > > > > > Suites: trixie-backports
> > > > > > Components: main
> > > > > > Signed-By: /etc/apt/noahm.gpg
> > > > > >
> > > > > > Let me know if this resolves the issue.  Similar packages will 
> > > > > > likely
> > > > > > ship in a forthcoming trixie point release.
> > > > > 
> > > > > Shouldn't these be shipped through stable-security?
> > > > > 
> > > > 
> > > > Possibly.  Let's see what the security team thinks.  Multiple people
> > > > have encountered this issue since the trixie release, and the impact is
> > > > a significant breach of privacy.  It doesn't impact the default
> > > > configuration, but it only takes uncommenting and adjusting one line to
> > > > trigger it.
> > > > 
> > > > Since we just released 13.1, there won't be another trixie point release
> > > > for a few months, which argues in favor of a DSA IMO.
> > > 
> > > As the next point release is on 15 November only and given the impact,
> > > yes tend to agree to release a DSA for this issue. Can you prepare the
> > > trixie-security debdiff?
> > 
> > See attached.  The diffstat is
> >  changelog                                                           |    8 
> > ++
> >  patches/auth__Use_AUTH_CACHE_KEY_USER_instead_of_per-database.patch |  124 
> > ++++++++++++++++++++++++++++++++++++++++++
> >  patches/series                                                      |    1 
> >  3 files changed, 133 insertions(+)
> > 
> > Note that there's no CVE referenced in the changelog, as we don't seem
> > to have one for this issue yet.
> 
> I will try to have a look at this over the weekend and come back to
> you.

Looks, good please upload to security-master (needs to be built with
-sa).

Regards,
Salvatore

Reply via email to