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

Re: gtkwave update for {bookworm,bullseye,buster}-security



On 29/03/2024 00:06, Adrian Bunk wrote:
Hi,

attached are proposed debdiffs for updating gtkwave to 3.3.118 in
{bookworm,bullseye,buster}-security for review for a DSA
(and as preview for buster).

General notes:

As suggested by the security team in #1060407, this is a backport of a
new upstream version to fix the 82 CVEs.

I checked a handful CVEs, and they were also present in buster.
If anyone insists that I check for every single CVE whether it is also
in buster I can do that, but that would be a lot of work.

As already mentioned in #1060407, the ghwdump tool (and manpage) was
dropped in 3.3.110 from the upstream sources, and is now in ghdl-tools.
For bullseye and buster it is therefore readded.

As mentioned in #1060407 there are different tarballs for GTK 2 and GTK 3.
Looking closer I realized that this is actually one tarball that
supports GTK 1+2, and one tarball that supports GTK 2+3.
I did stay at the GTK 1+2 tarball that was already used before
for bullseye and buster since there was anyway a different upstream
tarball required for the +really version that is required to avoid
creating file conflicts with ghwdump when upgrading to bookworm.

What does the security team consider the best versioning for bullseye?
In #1060407 I suggested 3.3.104+really3.3.118-0.1, but now I ended up
preferring 3.3.104+really3.3.118-0+deb11u1

I saw this earlier but I couldn't think of a better versioning scheme, though this looked awkward. Now I have thought of a (possibly) better one, so I'm stating it here in case we find ourselves in a similar situation in the future and someone remembers this thread.

I would have gone with

  3.3.118-0.1~deb12u1
  3.3.118+gtk2-0+deb11u1
  3.3.118+gtk2-0+deb10u1

Similar to how we do +dfsg or +repack. The +really is usually used for going back without adding an epoch, but here we're going forward, so perhaps such a naming would have made more sense. It also makes it clearer why there's a different tarball.

Cheers,
Emilio


Reply to: