[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [PATCH 0/1] Gst-plugins-good security update
From: |
Leo Famulari |
Subject: |
Re: [PATCH 0/1] Gst-plugins-good security update |
Date: |
Sat, 26 Nov 2016 14:38:42 -0500 |
User-agent: |
Mutt/1.7.1 (2016-10-04) |
On Sat, Nov 26, 2016 at 09:51:30AM +0100, Marius Bakke wrote:
> Leo Famulari <address@hidden> writes:
>
> > This patch should fix the bugs named here:
> >
> > http://seclists.org/oss-sec/2016/q4/517
> >
> > I copied Debian's approach, which is to take all the recent patches for
> > the vulnerable component (the FLIC decoder).
> >
> > My understanding is that the first two patches fix the CVEs, the 3rd
> > fixes an unrelated bug, and the 4th is a total rewrite of the component,
> > because "code is terrible, it should be entirely re-written" [0].
> >
> > The CVE bug fixes are not split into discrete patches, so it doesn't
> > work to make patches for each CVE ID, like we normally do.
> >
> > Is this approach (concatenating the patches) okay?
>
> I prefer having them separately, so the upstream commit can be clearly
> referenced in the patch header; and they can be reviewed and modified
> independently.
>
> In this instance it's okay, since I just checked out the 1.10 branch and
> concatenated the four commits and ended up with the same patch :-)
>
> That's not to say it should not be allowed. I think this approach is
> fine for long patch series, but at only four patches it's not the best
> precedent.
>
> Anyway, thanks for taking care of this, and LGTM! Please push! :-)
I split them up and (hopefully) annotated them well enough that readers
can follow along.
Pushed as 9e46245b89e0f30397f69391a2219a29caa336a2
signature.asc
Description: PGP signature