Re: Bug#773834: Preparing a release for stable and lts
On Tue, May 12, 2015 at 1:55 PM, Holger Levsen <firstname.lastname@example.org> wrote:
> On Dienstag, 12. Mai 2015, Vincent Fourmond wrote:
>> Looks like I've started enjoying my vacation before attaching the
>> file... Sorry about that.
> I hope you had a good one! :) And thanks for coming back to this issue...
>> Bastien has worked more on the patches. I'm attaching the debdiff
>> between 8:184.108.40.206-3+squeeze5 and the proposed new version. Please note
>> that, for internal (gitpkg-related) reasons, some patches have changed
>> names between the squeeze5 and squeeze6 version, which is not easy to
>> detect because of the huge number of new patches in the squeeze6
> sigh. i'm not sure this is a the right thing to do...
I can understand that. The way we use gitpkg is great, but not too
NMU-friendly. Anyway, I don't know if you had the chance to take a
closer look at the code changes, but it's a huge patch queue in the
first place. I have generated the diffs between compiled code (ie
diffs between dpkg-source -x codes, there are no further patches
applied at build time), if you like, I can comment them, and show that
the all the diffs are still there in the new version.
>> However, I've checked manually that the patches are still
>> applied to the final code (one of them has further changed), so I
>> don't expect any regression on that point.
>> I haven't tried to build yet, but that's what I'm going to do right
>> now, if I can get my squeeze chroot up again.
>> If I get it to build, would it be OK to upload as is ?
> once you have build it, you should *test* that package. once that has been
> done, we can think about uploading ;-)
Argh, unfortunately, I'm nowhere near that yet. Looks like some of
the patches are harder to backport than anticipated, as it FTBS for
In any case, the build comprises a heavy test suite that should
catch most of the problems, but probably not the security-related