Hello Robie, Steve,

I will address both of your updates sequentially :

Le 14/01/2016 19:13, Robie Basak a écrit :
> Hi Louis,
> 
> Thank you for working on this merge. Good job identifying some changes that
> could be dropped.
> 
> However I think this work could do with some improvement before I am happy to
> upload it.
> 
> Overall it looks like it might be OK, but I'm concerned that the changelog is
> misleading. There are changes that are being carried forward that are not
> mentioned in your Remaining Changes section, and conversely there are also
> changes that are claimed to have been carried forward but are partially
> missing. Details below:
> 
> udev changes in debian/rules. Was this dropped? In your merge the delta still
> seems to be there, and it still looks relevant. So why does the changelog say
> that this is dropped?

Oh dear! I mistakenly thought that the change was coming from upstream as there
is no mention in grab-merge's REPORT file of any conflict on debian/rules. I
will revert the changelog entry accordingly.

I definitively will stop to use this tool for merges. Not only it makes my life
miserable by having to systematically revert to pristine source packages for
comparaison, but it seems to silently changes files without any indication.

So far, grab-merge has been a source of problems for me. I had to restart the
rsyslog merge twice because of mismatch and build problems.

I may need your help to get my head around it, but I will give a second look at
your git workflow that I bookmarked earlier.
> 
> Some but not all changes made in 2.7.1-1ubuntu5 seem to have disappeared. Are
> they still needed? Why is this part of the delta not mentioned in the
> changelog?
>

Indeed, the .ubuntu5 modification that you uploaded got dropped. I was mislead
to believe that it was an upstream change added and not dropped from our part. I
will reintroduce it.

I also added a mention of dropped changes in debian/tests/test-nut.py that are
now superseded upstream but had no mention in the previous changelog.

> What about these parts of the previous delta, for which I don't see any
> information in your merge changelog about whether these were carried forward or
> dropped?
> 
>       * Disable systemd unit; it does not check nut.conf whether nut is
>         configured, and thus fails to start (and the package install). The init.d
>         script works well enough for the time being. (LP: #1313231)
> 
>       * debian/tests/test-nut.py: in the CVE_2012_2944 test, give nut at most 5
>         seconds to shut down, instead of expecting it to be shut down immediately
>         after sending the killall. (LP: #1291378)
> 
>       * debian/patches/0006-ups-conf-reorder.patch: Move maxretry setting
>         above Examples section, closer to the "outside of a driver definition"
>         comment. (LP: #1405822)
> 

My changelog mention the following :

    * Ubuntu bugs fixed upstream:

      LP: #1405822, #1291378, #1313231


Those are the three bugs that you mention above.  I do agree that I should have
carried forward the previous changelog description and state that the bug had
been fixed upstream. I'll carry that change in a new debdiff.

> Please could you address these questions above? In particular I'd like
> the changelog to be accurate in explaining what is being carried
> forward, what is being dropped and any additional changes made (I think
> there are none in this case). Otherwise it's tough both to review now
> and for a future merge.
> 
> Minor: I'd put the extra Ubuntu delta lines for debian/patches/series after all
> the Debian ones, unless there's a good reason to insert them into the middle.
> 

Regarding Steve's update now :

>  - nut-{cgi,server}.postinst: the existing delta with Debian is bad form,
> adduser is designed to be called unconditionally. Not strictly speaking part
> of the merge, but something I spotted along the way that I think it's
> important to fix. Along the way, I fixed up the changelog explanation for why
> this is still needed (since we no longer do upgrades from hardy, if that were
> the only reason we could drop the delta).
>

Fixed in the new merge.

>  - debian/patches/backport-fix-lp753661.patch has been a pointless diff for
> several upstream releases. The patch still applies, but upstream fixed the
> issue differently (using rtrim() ).

Fixed in the new merge.

>  - as Robie noted, you say the udev delta is dropped, but the delta to
> debian/rules still exists. I've dropped the delta, with more explanation in
> the changelog to why it can be dropped.

Fixed in the new merge with your drop of the previous delta.

>  - the changes from 2.7.7-1ubuntu5 are restored, as still required.

Yes, already noted in Robie's comment

Regarding the apport hooks, I will treat them in a separate bug.

I hope that everything is correct this time. I will attatch a new debdiff shortly.

Kind regards,

..Louis


-- 
Louis Bouchard
Software engineer, Cloud & Sustaining eng.
Canonical Ltd
Ubuntu developer                       Debian Maintainer
GPG : 429D 7A3B DD05 B6F8 AF63  B9C4 8B3D 867C 823E 7A61