Bug#1120709: RFS: super/3.30.3-2.2 [NMU] [RC] -- Execute commands setuid root
Control: tags -1 - moreinfo
Hi Andreas and Tobi,
Thanks for the reviews! I have made the requested changes.
On Sat, Nov 15, 2025 at 01:13:25PM +0100, Andreas Metzler wrote:
> Doing a sponsored NMU for a bug like 1106914 which requires a
> single-word-addition is terribly inefficient, the sponsee's contribution
> just does not save any work,
Yes, it is inefficient for both parties.
I am more than happy for someone to ignore/close my RFS and simply
upload the fix themselves unattributed. I will be satisfied knowing that
I prompted the action and my own sanity checks contributed to risk
reduction.
On Sat, Nov 15, 2025 at 04:13:01PM +0100, Tobias Frost wrote:
[...]
> as Andreas already wrote, these changes are inappropiate for a NMU,
> please consult the developers-reference on this topic for details.
>
> Specifically everything that does not close filed bugs or is just
> cosmetics is not appropiate.
I am not entirely convinced this characterisation fits the changes I
have now dropped:
> > * d/lintian-overrides: fix broken/redundant overrides
The broken lintian overrides are a pure function of the passage of time.
No one is going to contest the mechanical fixes. I had to check the
lintian errors in case there was a regression or something important,
doubled in number due to the incorrect overrides being reported. Fixing
these now reduces the tax on future contributors and reviewers so they
don't need to redo my work reviewing them all over again, each time. It
is a good value fix and in scope for an NMU by my reading of the
developers' reference. I have moved them to a separate merge request for
the maintainer to look at when he has a chance.
(That said, I did miss something in the initial version of this fix, so
something good came out of the rework!)
> > + remove ancient dependency versioning
This is also a function of the passage of time. Specifically 2001 and
2003. Ancient package relationships are flagged in maintscripts and by
package reviewers I have observed. This improves SnR, therefore
improving the possibility of spotting other issues.
> > + wrap-and-sort -ast
FWIW I wasn't sure about this, mostly because the changelog makes it
sound more intrusive and frivolous than it is. The best time to convert
inline lists into multi-line form with trailing commas to support
composability and reviewability of git commits is previously. The second
best time is in a commit preceding new changes to the lists to mitigate
the increased technical debt.
> Please revert this changes and then reupload a new package to
> mentors.
> Remove the moreinfo tag when ready.
All done, thanks!
Reply to: